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 3.0.4 and 3.0.5 [xx XXX xxxx]
33 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
34 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
35 This issue makes the RSA implementation with 2048 bit private keys
36 incorrect on such machines and memory corruption will happen during
37 the computation. As a consequence of the memory corruption an attacker
38 may be able to trigger a remote code execution on the machine performing
41 SSL/TLS servers or other servers using 2048 bit RSA private keys running
42 on machines supporting AVX512IFMA instructions of the X86_64 architecture
43 are affected by this issue.
48 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
49 implementation would not encrypt the entirety of the data under some
50 circumstances. This could reveal sixteen bytes of data that was
51 preexisting in the memory that wasn't written. In the special case of
52 "in place" encryption, sixteen bytes of the plaintext would be revealed.
54 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
55 they are both unaffected.
58 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
60 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
62 * In addition to the c_rehash shell command injection identified in
63 CVE-2022-1292, further bugs where the c_rehash script does not
64 properly sanitise shell metacharacters to prevent command injection have been
67 When the CVE-2022-1292 was fixed it was not discovered that there
68 are other places in the script where the file names of certificates
69 being hashed were possibly passed to a command executed through the shell.
71 This script is distributed by some operating systems in a manner where
72 it is automatically executed. On such operating systems, an attacker
73 could execute arbitrary commands with the privileges of the script.
75 Use of the c_rehash script is considered obsolete and should be replaced
76 by the OpenSSL rehash command line tool.
79 *Daniel Fiala, Tomáš Mráz*
81 * Case insensitive string comparison no longer uses locales. It has instead
82 been directly implemented.
86 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
88 * Case insensitive string comparison is reimplemented via new locale-agnostic
89 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
90 comparison. The previous implementation had problems when the Turkish locale
95 * Fixed a bug in the c_rehash script which was not properly sanitising shell
96 metacharacters to prevent command injection. This script is distributed by
97 some operating systems in a manner where it is automatically executed. On
98 such operating systems, an attacker could execute arbitrary commands with the
99 privileges of the script.
101 Use of the c_rehash script is considered obsolete and should be replaced
102 by the OpenSSL rehash command line tool.
107 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
108 certificate on an OCSP response. The bug caused the function in the case
109 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
110 response (meaning a successful verification) even in the case where the
111 response signing certificate fails to verify.
113 It is anticipated that most users of `OCSP_basic_verify` will not use the
114 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
115 a negative value (indicating a fatal error) in the case of a certificate
116 verification failure. The normal expected return value in this case would be
119 This issue also impacts the command line OpenSSL "ocsp" application. When
120 verifying an ocsp response with the "-no_cert_checks" option the command line
121 application will report that the verification is successful even though it
122 has in fact failed. In this case the incorrect successful response will also
123 be accompanied by error messages showing the failure and contradicting the
124 apparently successful result.
129 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
130 AAD data as the MAC key. This made the MAC key trivially predictable.
132 An attacker could exploit this issue by performing a man-in-the-middle attack
133 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
134 that the modified data would still pass the MAC integrity check.
136 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
137 endpoint will always be rejected by the recipient and the connection will
138 fail at that point. Many application protocols require data to be sent from
139 the client to the server first. Therefore, in such a case, only an OpenSSL
140 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
142 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
143 sent in both directions. In this case both clients and servers could be
144 affected, regardless of the application protocol.
146 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
147 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
148 the handshake when using this ciphersuite.
150 The confidentiality of data is not impacted by this issue, i.e. an attacker
151 cannot decrypt data that has been encrypted using this ciphersuite - they can
154 In order for this attack to work both endpoints must legitimately negotiate
155 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
156 OpenSSL 3.0, and is not available within the default provider or the default
157 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
158 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
159 following must have occurred:
161 1) OpenSSL must have been compiled with the (non-default) compile time option
162 enable-weak-ssl-ciphers
164 2) OpenSSL must have had the legacy provider explicitly loaded (either
165 through application code or via configuration)
167 3) The ciphersuite must have been explicitly added to the ciphersuite list
169 4) The libssl security level must have been set to 0 (default is 1)
171 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
173 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
174 others that both endpoints have in common
179 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
180 occuppied by the removed hash table entries.
182 This function is used when decoding certificates or keys. If a long lived
183 process periodically decodes certificates or keys its memory usage will
184 expand without bounds and the process might be terminated by the operating
185 system causing a denial of service. Also traversing the empty hash table
186 entries will take increasingly more time.
188 Typically such long lived processes might be TLS clients or TLS servers
189 configured to accept client certificate authentication.
192 *Hugo Landau, Aliaksei Levin*
194 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
195 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
196 statistics are no longer supported. For compatibility, these statistics are
197 still listed in the output but are now always reported as zero.
201 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
203 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
204 for non-prime moduli.
206 Internally this function is used when parsing certificates that contain
207 elliptic curve public keys in compressed form or explicit elliptic curve
208 parameters with a base point encoded in compressed form.
210 It is possible to trigger the infinite loop by crafting a certificate that
211 has invalid explicit curve parameters.
213 Since certificate parsing happens prior to verification of the certificate
214 signature, any process that parses an externally supplied certificate may thus
215 be subject to a denial of service attack. The infinite loop can also be
216 reached when parsing crafted private keys as they can contain explicit
217 elliptic curve parameters.
219 Thus vulnerable situations include:
221 - TLS clients consuming server certificates
222 - TLS servers consuming client certificates
223 - Hosting providers taking certificates or private keys from customers
224 - Certificate authorities parsing certification requests from subscribers
225 - Anything else which parses ASN.1 elliptic curve parameters
227 Also any other applications that use the BN_mod_sqrt() where the attacker
228 can control the parameter values are vulnerable to this DoS issue.
233 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
234 to the list of ciphersuites providing Perfect Forward Secrecy as
235 required by SECLEVEL >= 3.
237 *Dmitry Belyavskiy, Nicola Tuveri*
239 * Made the AES constant time code for no-asm configurations
240 optional due to the resulting 95% performance degradation.
241 The AES constant time code can be enabled, for no assembly
242 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
246 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
251 * The negative return value handling of the certificate verification callback
252 was reverted. The replacement is to set the verification retry state with
253 the SSL_set_retry_verify() function.
257 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
259 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
260 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
261 verify a certificate supplied by a server. That function may return a
262 negative return value to indicate an internal error (for example out of
263 memory). Such a negative return value is mishandled by OpenSSL and will cause
264 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
265 success and a subsequent call to SSL_get_error() to return the value
266 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
267 returned by OpenSSL if the application has previously called
268 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
269 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
270 totally unexpected and applications may not behave correctly as a result. The
271 exact behaviour will depend on the application but it could result in
272 crashes, infinite loops or other similar incorrect responses.
274 This issue is made more serious in combination with a separate bug in OpenSSL
275 3.0 that will cause X509_verify_cert() to indicate an internal error when
276 processing a certificate chain. This will occur where a certificate does not
277 include the Subject Alternative Name extension but where a Certificate
278 Authority has enforced name constraints. This issue can occur even with valid
284 * Corrected a few file name and file reference bugs in the build,
285 installation and setup scripts, which lead to installation verification
286 failures. Slightly enhanced the installation verification script.
290 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
295 * Fixed PVK encoder to properly query for the passphrase.
299 * Multiple fixes in the OSSL_HTTP API functions.
303 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
304 OSSL_PARAM_INTEGER data type and return error on negative numbers
305 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
306 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
310 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
314 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
318 * Multiple threading fixes.
322 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
326 * Allow fetching an operation from the provider that owns an unexportable key
327 as a fallback if that is still allowed by the property query.
331 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
333 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
338 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
339 S390X capability vector to zero. This simplifies testing of different code
340 paths on S390X architecture.
344 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
345 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
346 SP 800-38D". The communication will fail at this point.
350 * The EC_GROUP_clear_free() function is deprecated as there is nothing
351 confidential in EC_GROUP data.
355 * The byte order mark (BOM) character is ignored if encountered at the
356 beginning of a PEM-formatted file.
360 * Added CMS support for the Russian GOST algorithms.
364 * Due to move of the implementation of cryptographic operations
365 to the providers, validation of various operation parameters can
366 be postponed until the actual operation is executed where previously
367 it happened immediately when an operation parameter was set.
369 For example when setting an unsupported curve with
370 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
371 fail but later keygen operations with the EVP_PKEY_CTX will fail.
373 *OpenSSL team members and many third party contributors*
375 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
376 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
377 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
378 instead to retrieve these algorithms from a provider.
382 * On build targets where the multilib postfix is set in the build
383 configuration the libdir directory was changing based on whether
384 the lib directory with the multilib postfix exists on the system
385 or not. This unpredictable behavior was removed and eventual
386 multilib postfix is now always added to the default libdir. Use
387 `--libdir=lib` to override the libdir if adding the postfix is
392 * The triple DES key wrap functionality now conforms to RFC 3217 but is
393 no longer interoperable with OpenSSL 1.1.1.
397 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
398 function codes, this function can only cause problems for calling
403 * Add a configurable flag to output date formats as ISO 8601. Does not
404 change the default date format.
408 * Version of MSVC earlier than 1300 could get link warnings, which could
409 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
410 Support for this flag has been removed.
414 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
415 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
416 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
417 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
418 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
422 * The signatures of the functions to get and set options on SSL and
423 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
424 Some source code changes may be required.
428 * The public definitions of conf_method_st and conf_st have been
429 deprecated. They will be made opaque in a future release.
431 *Rich Salz and Tomáš Mráz*
433 * Client-initiated renegotiation is disabled by default. To allow it, use
434 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
435 flag, or the "ClientRenegotiation" config parameter as appropriate.
439 * Add "abspath" and "includedir" pragma's to config files, to prevent,
440 or modify relative pathname inclusion.
444 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
445 validated. Please consult the README-FIPS and
446 README-PROVIDERS files, as well as the migration guide.
448 *OpenSSL team members and many third party contributors*
450 * For the key types DH and DHX the allowed settable parameters are now different.
454 * The openssl commands that read keys, certificates, and CRLs now
455 automatically detect the PEM or DER format of the input files.
457 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
459 * Added enhanced PKCS#12 APIs which accept a library context.
463 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
467 * Added support for Kernel TLS (KTLS).
469 *Boris Pismenny, John Baldwin and Andrew Gallatin*
471 * Support for RFC 5746 secure renegotiation is now required by default for
472 SSL or TLS connections to succeed.
476 * The signature of the `copy` functional parameter of the
477 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
478 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
479 the signature of the `pub_decode` functional parameter of the
480 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
481 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
485 * The error return values from some control calls (ctrl) have changed.
489 * A public key check is now performed during EVP_PKEY_derive_set_peer().
493 * Many functions in the EVP_ namespace that are getters of values from
494 implementations or contexts were renamed to include get or get0 in their
495 names. Old names are provided as macro aliases for compatibility and
500 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
501 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
502 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
507 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
510 * The output from the command line applications may have minor
515 * The output from numerous "printing" may have minor changes.
519 * Windows thread synchronization uses read/write primitives (SRWLock) when
520 supported by the OS, otherwise CriticalSection continues to be used.
524 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
525 work on read only BIO source/sinks that do not support these functions.
526 This allows piping or redirection of a file BIO using stdin to be buffered
527 into memory. This is used internally in OSSL_DECODER_from_bio().
531 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
532 this function would return one of the values OSSL_STORE_INFO_NAME,
533 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
534 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
535 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
536 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
537 using this function should be amended to handle the changed return value.
541 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
542 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
543 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
544 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
545 contains more than one certificate identifier: This means that all
546 certificates referenced there MUST be part of the validation chain.
550 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
551 RC5, DESX and DES have been moved to the legacy provider.
555 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
556 RIPEMD-160 have been moved to the legacy provider.
560 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
565 * The deprecated functions EVP_PKEY_get0_RSA(),
566 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
567 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
568 well as the similarly named "get1" functions behave differently in
573 * A number of functions handling low-level keys or engines were deprecated
574 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
575 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
576 EVP_PKEY_get0_siphash().
580 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
581 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
582 will need to load the legacy crypto provider. This includes these PBE
583 algorithms which use this KDF:
584 - NID_pbeWithMD2AndDES_CBC
585 - NID_pbeWithMD5AndDES_CBC
586 - NID_pbeWithSHA1AndRC2_CBC
587 - NID_pbeWithMD2AndRC2_CBC
588 - NID_pbeWithMD5AndRC2_CBC
589 - NID_pbeWithSHA1AndDES_CBC
593 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
594 BIO_debug_callback() functions.
598 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
599 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
603 * The RAND_METHOD APIs have been deprecated.
607 * The SRP APIs have been deprecated.
611 * Add a compile time option to prevent the caching of provider fetched
612 algorithms. This is enabled by including the no-cached-fetch option
613 at configuration time.
617 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
618 count of PKCS12_DEFAULT_ITER.
620 *Tomáš Mráz and Sahana Prasad*
622 * The openssl speed command does not use low-level API calls anymore.
626 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
629 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
631 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
635 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
636 providers may supply their own group implementations (using either the "key
637 exchange" or the "key encapsulation" methods) which will automatically be
638 detected and used by libssl.
640 *Matt Caswell, Nicola Tuveri*
642 * The undocumented function X509_certificate_type() has been deprecated;
646 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
650 * Removed RSA padding mode for SSLv23 (which was only used for
651 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
652 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
657 * Deprecated the obsolete X9.31 RSA key generation related functions.
659 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
660 is not allowed to return a value > 1, this is no more taken as failure.
662 *Viktor Dukhovni and David von Oheimb*
664 * Deprecated the obsolete X9.31 RSA key generation related functions
665 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
666 BN_X931_generate_prime_ex().
670 * The default key generation method for the regular 2-prime RSA keys was
671 changed to the FIPS 186-4 B.3.6 method.
675 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
679 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
683 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
684 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
686 *Rich Salz, Richard Levitte, and David von Oheimb*
688 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
692 * Deprecated `OCSP_parse_url()`.
696 * Validation of SM2 keys has been separated from the validation of regular EC
701 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
702 switches: a validation failure triggers an early exit, returning a failure
703 exit status to the parent process.
707 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
708 to ignore unknown ciphers.
712 * The `-cipher-commands` and `-digest-commands` options
713 of the command line utility `list` have been deprecated.
714 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
718 * Added convenience functions for generating asymmetric key pairs:
719 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
720 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
724 * All of the low level EC_KEY functions have been deprecated.
726 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
728 * Deprecated all the libcrypto and libssl error string loading
733 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
734 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
739 * The `-crypt` option to the `passwd` command line tool has been removed.
743 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
748 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
752 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
753 EVP_PKEY_get1_tls_encodedpoint().
757 * The security callback, which can be customised by application code, supports
758 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
759 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
763 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
764 interface. Their functionality remains unchanged.
768 * Added new option for 'openssl list', '-providers', which will display the
769 list of loaded providers, their names, version and status. It optionally
770 displays their gettable parameters.
774 * Removed EVP_PKEY_set_alias_type().
778 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
779 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
783 * Changed all "STACK" functions to be macros instead of inline functions. Macro
784 parameters are still checked for type safety at compile time via helper
789 * Remove the RAND_DRBG API
791 *Paul Dale and Matthias St. Pierre*
793 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
794 as well as actual hostnames.
798 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
799 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
800 conversely, silently ignore DTLS protocol version bounds when configuring
801 TLS-based contexts. The commands can be repeated to set bounds of both
802 types. The same applies with the corresponding "min_protocol" and
803 "max_protocol" command-line switches, in case some application uses both TLS
806 SSL_CTX instances that are created for a fixed protocol version (e.g.
807 `TLSv1_server_method()`) also silently ignore version bounds. Previously
808 attempts to apply bounds to these protocol versions would result in an
809 error. Now only the "version-flexible" SSL_CTX instances are subject to
810 limits in configuration files in command-line options.
814 * Deprecated the `ENGINE` API. Engines should be replaced with providers
819 * Reworked the recorded ERR codes to make better space for system errors.
820 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
821 given code is a system error (true) or an OpenSSL error (false).
825 * Reworked the test perl framework to better allow parallel testing.
827 *Nicola Tuveri and David von Oheimb*
829 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
830 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
834 * 'Configure' has been changed to figure out the configuration target if
835 none is given on the command line. Consequently, the 'config' script is
836 now only a mere wrapper. All documentation is changed to only mention
839 *Rich Salz and Richard Levitte*
841 * Added a library context `OSSL_LIB_CTX` that applications as well as
842 other libraries can use to form a separate context within which
843 libcrypto operations are performed.
847 * Added various `_ex` functions to the OpenSSL API that support using
848 a non-default `OSSL_LIB_CTX`.
852 * Handshake now fails if Extended Master Secret extension is dropped
857 * Dropped interactive mode from the `openssl` program.
861 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
863 *David von Oheimb and Shane Lontis*
865 * Deprecated `EC_METHOD_get_field_type()`.
869 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
870 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
871 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
875 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
879 * Add CAdES-BES signature verification support, mostly derived
880 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
882 *Filipe Raimundo da Silva*
884 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
888 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
889 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
893 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
897 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
898 EC_KEY_precompute_mult().
902 * Deprecated EC_POINTs_mul().
906 * Removed FIPS_mode() and FIPS_mode_set().
910 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
914 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
915 EC_POINT_get_Jprojective_coordinates_GFp().
919 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
920 arrays to be more easily constructed via a series of utility functions.
921 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
922 the various push functions and finally convert to a passable OSSL_PARAM
923 array using OSSL_PARAM_BLD_to_param().
927 * The security strength of SHA1 and MD5 based signatures in TLS has been
932 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
933 contain a provider side internal key.
937 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
941 * Project text documents not yet having a proper file name extension
942 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
943 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
947 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
948 have been converted to Markdown with the goal to produce documents
949 which not only look pretty when viewed online in the browser, but
950 remain well readable inside a plain text editor.
952 To achieve this goal, a 'minimalistic' Markdown style has been applied
953 which avoids formatting elements that interfere too much with the
954 reading flow in the text file. For example, it
956 * avoids [ATX headings][] and uses [setext headings][] instead
957 (which works for `<h1>` and `<h2>` headings only).
958 * avoids [inline links][] and uses [reference links][] instead.
959 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
961 [ATX headings]: https://github.github.com/gfm/#atx-headings
962 [setext headings]: https://github.github.com/gfm/#setext-headings
963 [inline links]: https://github.github.com/gfm/#inline-link
964 [reference links]: https://github.github.com/gfm/#reference-link
965 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
966 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
968 *Matthias St. Pierre*
970 * The test suite is changed to preserve results of each test recipe.
971 A new directory test-runs/ with subdirectories named like the
972 test recipes are created in the build tree for this purpose.
976 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
977 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
978 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
980 *David von Oheimb, Martin Peylo*
982 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
983 It supports arbitrary request and response content types, GET redirection,
984 TLS, connections via HTTP(S) proxies, connections and exchange via
985 user-defined BIOs (allowing implicit connections), persistent connections,
986 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
987 The legacy OCSP-focused (and only partly documented) API
988 is retained for backward compatibility, while most of it is deprecated.
992 * Added `util/check-format.pl`, a tool for checking adherence to the
993 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
994 The checks performed are incomplete and yield some false positives.
995 Still the tool should be useful for detecting most typical glitches.
999 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1000 If domain name resolution yields multiple IP addresses all of them are tried
1001 after `connect()` failures.
1005 * All of the low level RSA functions have been deprecated.
1009 * X509 certificates signed using SHA1 are no longer allowed at security
1014 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1015 modified to use PKEY APIs. These commands are now in maintenance mode
1016 and no new features will be added to them.
1020 * The command line utility rsautl has been deprecated.
1024 * The command line utilities genrsa and rsa have been modified to use PKEY
1025 APIs. They now write PKCS#8 keys by default. These commands are now in
1026 maintenance mode and no new features will be added to them.
1030 * All of the low level DH functions have been deprecated.
1032 *Paul Dale and Matt Caswell*
1034 * All of the low level DSA functions have been deprecated.
1038 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1039 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1043 * Deprecated low level ECDH and ECDSA functions.
1047 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1051 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1052 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1053 a new formulation to include all the things it can be used for,
1054 as well as words of caution.
1058 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1062 * All of the low level HMAC functions have been deprecated.
1064 *Paul Dale and David von Oheimb*
1066 * Over two thousand fixes were made to the documentation, including:
1067 - Common options (such as -rand/-writerand, TLS version control, etc)
1068 were refactored and point to newly-enhanced descriptions in openssl.pod.
1069 - Added style conformance for all options (with help from Richard Levitte),
1070 documented all reported missing options, added a CI build to check
1071 that all options are documented and that no unimplemented options
1073 - Documented some internals, such as all use of environment variables.
1074 - Addressed all internal broken L<> references.
1078 * All of the low level CMAC functions have been deprecated.
1082 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1083 functions have been deprecated.
1085 *Paul Dale and David von Oheimb*
1087 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1088 set of functions. The documentation mentioned negative values for some
1089 errors, but this was never the case, so the mention of negative values
1092 Code that followed the documentation and thereby check with something
1093 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1097 * All of the low level cipher functions have been deprecated.
1099 *Matt Caswell and Paul Dale*
1101 * Removed include/openssl/opensslconf.h.in and replaced it with
1102 include/openssl/configuration.h.in, which differs in not including
1103 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1104 was added to include both.
1106 This allows internal hacks where one might need to modify the set
1107 of configured macros, for example this if deprecated symbols are
1108 still supposed to be available internally:
1110 #include <openssl/configuration.h>
1112 #undef OPENSSL_NO_DEPRECATED
1113 #define OPENSSL_SUPPRESS_DEPRECATED
1115 #include <openssl/macros.h>
1117 This should not be used by applications that use the exported
1118 symbols, as that will lead to linking errors.
1122 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1123 used in exponentiation with 512-bit moduli. No EC algorithms are
1124 affected. Analysis suggests that attacks against 2-prime RSA1024,
1125 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1126 difficult to perform and are not believed likely. Attacks against DH512
1127 are considered just feasible. However, for an attack the target would
1128 have to re-use the DH512 private key, which is not recommended anyway.
1129 Also applications directly using the low-level API BN_mod_exp may be
1130 affected if they use BN_FLG_CONSTTIME.
1135 * Most memory-debug features have been deprecated, and the functionality
1136 replaced with no-ops.
1140 * Added documentation for the STACK API.
1144 * Introduced a new method type and API, OSSL_ENCODER, to represent
1145 generic encoders. These do the same sort of job that PEM writers
1146 and d2i functions do, but with support for methods supplied by
1147 providers, and the possibility for providers to support other
1152 * Introduced a new method type and API, OSSL_DECODER, to represent
1153 generic decoders. These do the same sort of job that PEM readers
1154 and i2d functions do, but with support for methods supplied by
1155 providers, and the possibility for providers to support other
1160 * Added a .pragma directive to the syntax of configuration files, to
1161 allow varying behavior in a supported and predictable manner.
1162 Currently added pragma:
1166 This allows dollar signs to be a keyword character unless it's
1167 followed by a opening brace or parenthesis. This is useful for
1168 platforms where dollar signs are commonly used in names, such as
1169 volume names and system directory names on VMS.
1173 * Added functionality to create an EVP_PKEY from user data.
1177 * Change the interpretation of the '--api' configuration option to
1178 mean that this is a desired API compatibility level with no
1179 further meaning. The previous interpretation, that this would
1180 also mean to remove all deprecated symbols up to and including
1181 the given version, no requires that 'no-deprecated' is also used
1182 in the configuration.
1184 When building applications, the desired API compatibility level
1185 can be set with the OPENSSL_API_COMPAT macro like before. For
1186 API compatibility version below 3.0, the old style numerical
1187 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1188 For version 3.0 and on, the value is expected to be the decimal
1189 value calculated from the major and minor version like this:
1191 MAJOR * 10000 + MINOR * 100
1195 -DOPENSSL_API_COMPAT=30000 For 3.0
1196 -DOPENSSL_API_COMPAT=30200 For 3.2
1198 To hide declarations that are deprecated up to and including the
1199 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1200 given when building the application as well.
1204 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1205 access to certificate and CRL stores via URIs and OSSL_STORE
1208 This adds the following functions:
1210 - X509_LOOKUP_store()
1211 - X509_STORE_load_file()
1212 - X509_STORE_load_path()
1213 - X509_STORE_load_store()
1214 - SSL_add_store_cert_subjects_to_stack()
1215 - SSL_CTX_set_default_verify_store()
1216 - SSL_CTX_load_verify_file()
1217 - SSL_CTX_load_verify_dir()
1218 - SSL_CTX_load_verify_store()
1222 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1223 The presence of this system service is determined at run-time.
1227 * Added functionality to create an EVP_PKEY context based on data
1228 for methods from providers. This takes an algorithm name and a
1229 property query string and simply stores them, with the intent
1230 that any operation that uses this context will use those strings
1231 to fetch the needed methods implicitly, thereby making the port
1232 of application written for pre-3.0 OpenSSL easier.
1236 * The undocumented function NCONF_WIN32() has been deprecated; for
1237 conversion details see the HISTORY section of doc/man5/config.pod
1241 * Introduced the new functions EVP_DigestSignInit_ex() and
1242 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1243 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1244 pages for further details.
1248 * Over two thousand fixes were made to the documentation, including:
1249 adding missing command flags, better style conformance, documentation
1252 *Rich Salz, Richard Levitte*
1254 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1255 X25519, X448, Ed25519 and Ed448.
1259 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1264 * Deprecated the public definition of `ERR_STATE` as well as the function
1265 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1270 * Added ERR functionality to give callers access to the stored function
1271 names that have replaced the older function code based functions.
1273 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1274 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1275 ERR_peek_error_all() and ERR_peek_last_error_all().
1277 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1278 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1279 ERR_func_error_string().
1283 * Extended testing to be verbose for failing tests only. The make variables
1284 VERBOSE_FAILURE or VF can be used to enable this:
1286 $ make VF=1 test # Unix
1287 $ mms /macro=(VF=1) test ! OpenVMS
1288 $ nmake VF=1 test # Windows
1292 * Added the `-copy_extensions` option to the `x509` command for use with
1293 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1294 all extensions in the request are copied to the certificate or vice versa.
1296 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1298 * Added the `-copy_extensions` option to the `req` command for use with
1299 `-x509`. When given with the `copy` or `copyall` argument,
1300 all extensions in the certification request are copied to the certificate.
1304 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1305 they generate are by default RFC 5280 compliant in the following sense:
1306 There is a subjectKeyIdentifier extension with a hash value of the public key
1307 and for not self-signed certs there is an authorityKeyIdentifier extension
1308 with a keyIdentifier field or issuer information identifying the signing key.
1309 This is done unless some configuration overrides the new default behavior,
1310 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1314 * Added several checks to `X509_verify_cert()` according to requirements in
1315 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1316 (which may be done by using the CLI option `-x509_strict`):
1317 * The basicConstraints of CA certificates must be marked critical.
1318 * CA certificates must explicitly include the keyUsage extension.
1319 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1320 * The issuer name of any certificate must not be empty.
1321 * The subject name of CA certs, certs with keyUsage crlSign,
1322 and certs without subjectAlternativeName must not be empty.
1323 * If a subjectAlternativeName extension is given it must not be empty.
1324 * The signatureAlgorithm field and the cert signature must be consistent.
1325 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1326 must not be marked critical.
1327 * The authorityKeyIdentifier must be given for X.509v3 certs
1328 unless they are self-signed.
1329 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1333 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1334 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1338 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1339 used even when parsing explicit parameters, when loading a encoded key
1340 or calling `EC_GROUP_new_from_ecpkparameters()`/
1341 `EC_GROUP_new_from_ecparameters()`.
1342 This prevents bypass of security hardening and performance gains,
1343 especially for curves with specialized EC_METHODs.
1344 By default, if a key encoded with explicit parameters is loaded and later
1345 encoded, the output is still encoded with explicit parameters, even if
1346 internally a "named" EC_GROUP is used for computation.
1350 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1351 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1352 NULL. After this change, only the cofactor parameter can be NULL. It also
1353 does some minimal sanity checks on the passed order.
1358 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1359 An attack is simple, if the first CMS_recipientInfo is valid but the
1360 second CMS_recipientInfo is chosen ciphertext. If the second
1361 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1362 encryption key will be replaced by garbage, and the message cannot be
1363 decoded, but if the RSA decryption fails, the correct encryption key is
1364 used and the recipient will not notice the attack.
1365 As a work around for this potential attack the length of the decrypted
1366 key must be equal to the cipher default key length, in case the
1367 certifiate is not given and all recipientInfo are tried out.
1368 The old behaviour can be re-enabled in the CMS code by setting the
1369 CMS_DEBUG_DECRYPT flag.
1373 * Early start up entropy quality from the DEVRANDOM seed source has been
1374 improved for older Linux systems. The RAND subsystem will wait for
1375 /dev/random to be producing output before seeding from /dev/urandom.
1376 The seeded state is stored for future library initialisations using
1377 a system global shared memory segment. The shared memory identifier
1378 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1379 the desired value. The default identifier is 114.
1383 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1384 when primes for RSA keys are computed.
1385 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1386 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1387 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1388 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1389 This avoids possible fingerprinting of newly generated RSA modules.
1393 * Correct the extended master secret constant on EBCDIC systems. Without this
1394 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1395 negotiate EMS will fail. Unfortunately this also means that TLS connections
1396 between EBCDIC systems with this fix, and EBCDIC systems without this
1397 fix will fail if they negotiate EMS.
1401 * Changed the library initialisation so that the config file is now loaded
1402 by default. This was already the case for libssl. It now occurs for both
1403 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1404 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1408 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1409 where the former acts as a replacement for `ERR_put_error()`, and the
1410 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1411 `ERR_raise_data()` adds more flexibility by taking a format string and
1412 an arbitrary number of arguments following it, to be processed with
1417 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1418 to check if a named provider is loaded and available. When called, it
1419 will also activate all fallback providers if such are still present.
1423 * Enforce a minimum DH modulus size of 512 bits.
1427 * Changed DH parameters to generate the order q subgroup instead of 2q.
1428 Previously generated DH parameters are still accepted by DH_check
1429 but DH_generate_key works around that by clearing bit 0 of the
1430 private key for those. This avoids leaking bit 0 of the private key.
1434 * Significantly reduce secure memory usage by the randomness pools.
1438 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1443 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1444 algorithms. An implementation of a key exchange algorithm can be obtained
1445 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1446 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1447 the older EVP_PKEY_derive_init() function. See the man pages for the new
1448 functions for further details.
1452 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1456 * Removed the function names from error messages and deprecated the
1461 * Removed NextStep support and the macro OPENSSL_UNISTD
1465 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1466 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1467 Also removed "export var as function" capability; we do not export
1468 variables, only functions.
1472 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1473 an error and 1 indicating success. In previous versions of OpenSSL this
1474 was a void type. If a key was set longer than the maximum possible this
1479 * Support SM2 signing and verification schemes with X509 certificate.
1483 * Use SHA256 as the default digest for TS query in the `ts` app.
1487 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1491 * Default cipher lists/suites are now available via a function, the
1492 #defines are deprecated.
1496 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1497 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1498 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1502 * Join the directories crypto/x509 and crypto/x509v3
1506 * Added command 'openssl kdf' that uses the EVP_KDF API.
1510 * Added command 'openssl mac' that uses the EVP_MAC API.
1514 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1515 as default directories. Also added the command 'openssl info'
1516 for scripting purposes.
1520 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1525 * Add prediction resistance to the DRBG reseeding process.
1529 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1530 mandated by IEEE Std 1619-2018.
1534 * Added newline escaping functionality to a filename when using openssl dgst.
1535 This output format is to replicate the output format found in the `*sum`
1536 checksum programs. This aims to preserve backward compatibility.
1538 *Matt Eaton, Richard Levitte, and Paul Dale*
1540 * Removed the heartbeat message in DTLS feature, as it has very
1541 little usage and doesn't seem to fulfill a valuable purpose.
1542 The configuration option is now deprecated.
1546 * Changed the output of 'openssl {digestname} < file' to display the
1547 digest name in its output.
1551 * Added a new generic trace API which provides support for enabling
1552 instrumentation through trace output.
1554 *Richard Levitte & Matthias St. Pierre*
1556 * Added build tests for C++. These are generated files that only do one
1557 thing, to include one public OpenSSL head file each. This tests that
1558 the public header files can be usefully included in a C++ application.
1560 This test isn't enabled by default. It can be enabled with the option
1561 'enable-buildtest-c++'.
1565 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1569 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1573 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1577 * Added KMAC to EVP_MAC.
1581 * Added property based algorithm implementation selection framework to
1586 * Added SCA hardening for modular field inversion in EC_GROUP through
1587 a new dedicated field_inv() pointer in EC_METHOD.
1588 This also addresses a leakage affecting conversions from projective
1589 to affine coordinates.
1591 *Billy Bob Brumley, Nicola Tuveri*
1593 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1594 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1595 those algorithms that were already supported through the EVP_PKEY API
1596 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1597 and scrypt are now wrappers that call EVP_KDF.
1601 * Build devcrypto engine as a dynamic engine.
1603 *Eneas U de Queiroz*
1605 * Add keyed BLAKE2 to EVP_MAC.
1609 * Fix a bug in the computation of the endpoint-pair shared secret used
1610 by DTLS over SCTP. This breaks interoperability with older versions
1611 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1612 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1613 interoperability with such broken implementations. However, enabling
1614 this switch breaks interoperability with correct implementations.
1616 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1617 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1621 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1625 * Changed the license to the Apache License v2.0.
1629 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1631 - Major releases (indicated by incrementing the MAJOR release number)
1632 may introduce incompatible API/ABI changes.
1633 - Minor releases (indicated by incrementing the MINOR release number)
1634 may introduce new features but retain API/ABI compatibility.
1635 - Patch releases (indicated by incrementing the PATCH number)
1636 are intended for bug fixes and other improvements of existing
1637 features only (like improving performance or adding documentation)
1638 and retain API/ABI compatibility.
1642 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1646 * Remove the 'dist' target and add a tarball building script. The
1647 'dist' target has fallen out of use, and it shouldn't be
1648 necessary to configure just to create a source distribution.
1652 * Recreate the OS390-Unix config target. It no longer relies on a
1653 special script like it did for OpenSSL pre-1.1.0.
1657 * Instead of having the source directories listed in Configure, add
1658 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1663 * Add GMAC to EVP_MAC.
1667 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1671 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1672 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1673 to facilitate the continued use of MACs through raw private keys in
1674 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1678 * Deprecate ECDH_KDF_X9_62().
1682 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1683 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1684 are retained for backwards compatibility.
1688 * AES-XTS mode now enforces that its two keys are different to mitigate
1689 the attacked described in "Efficient Instantiations of Tweakable
1690 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1691 Details of this attack can be obtained from:
1692 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1696 * Rename the object files, i.e. give them other names than in previous
1697 versions. Their names now include the name of the final product, as
1698 well as its type mnemonic (bin, lib, shlib).
1702 * Added new option for 'openssl list', '-objects', which will display the
1703 list of built in objects, i.e. OIDs with names.
1707 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1708 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1713 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1714 improves application performance by removing data copies and providing
1715 applications with zero-copy system calls such as sendfile and splice.
1719 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1723 * `PKCS12_parse` now maintains the order of the parsed certificates
1724 when outputting them via `*ca` (rather than reversing it).
1728 * Deprecated pthread fork support methods.
1732 * Added support for FFDHE key exchange in TLS 1.3.
1736 * Added a new concept for OpenSSL plugability: providers. This
1737 functionality is designed to replace the ENGINE API and ENGINE
1738 implementations, and to be much more dynamic, allowing provider
1739 authors to introduce new algorithms among other things, as long as
1740 there's an API that supports the algorithm type.
1742 With this concept comes a new core API for interaction between
1743 libcrypto and provider implementations. Public libcrypto functions
1744 that want to use providers do so through this core API.
1746 The main documentation for this core API is found in
1747 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1748 refer to other manuals describing the API specific for supported
1749 algorithm types (also called operations).
1756 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1758 * Avoid loading of a dynamic engine twice.
1762 * Prioritise DANE TLSA issuer certs over peer certs
1766 * Fixed random API for MacOS prior to 10.12
1768 These MacOS versions don't support the CommonCrypto APIs
1772 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1774 * Fixed an SM2 Decryption Buffer Overflow.
1776 In order to decrypt SM2 encrypted data an application is expected to
1777 call the API function EVP_PKEY_decrypt(). Typically an application will
1778 call this function twice. The first time, on entry, the "out" parameter
1779 can be NULL and, on exit, the "outlen" parameter is populated with the
1780 buffer size required to hold the decrypted plaintext. The application
1781 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1782 again, but this time passing a non-NULL value for the "out" parameter.
1784 A bug in the implementation of the SM2 decryption code means that the
1785 calculation of the buffer size required to hold the plaintext returned
1786 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1787 size required by the second call. This can lead to a buffer overflow
1788 when EVP_PKEY_decrypt() is called by the application a second time with
1789 a buffer that is too small.
1791 A malicious attacker who is able present SM2 content for decryption to
1792 an application could cause attacker chosen data to overflow the buffer
1793 by up to a maximum of 62 bytes altering the contents of other data held
1794 after the buffer, possibly changing application behaviour or causing
1795 the application to crash. The location of the buffer is application
1796 dependent but is typically heap allocated.
1801 * Fixed various read buffer overruns processing ASN.1 strings
1803 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1804 structure which contains a buffer holding the string data and a field
1805 holding the buffer length. This contrasts with normal C strings which
1806 are repesented as a buffer for the string data which is terminated
1807 with a NUL (0) byte.
1809 Although not a strict requirement, ASN.1 strings that are parsed using
1810 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1811 well as any string whose value has been set with the ASN1_STRING_set()
1812 function will additionally NUL terminate the byte array in the
1813 ASN1_STRING structure.
1815 However, it is possible for applications to directly construct valid
1816 ASN1_STRING structures which do not NUL terminate the byte array by
1817 directly setting the "data" and "length" fields in the ASN1_STRING
1818 array. This can also happen by using the ASN1_STRING_set0() function.
1820 Numerous OpenSSL functions that print ASN.1 data have been found to
1821 assume that the ASN1_STRING byte array will be NUL terminated, even
1822 though this is not guaranteed for strings that have been directly
1823 constructed. Where an application requests an ASN.1 structure to be
1824 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1825 been directly constructed by the application without NUL terminating
1826 the "data" field, then a read buffer overrun can occur.
1828 The same thing can also occur during name constraints processing
1829 of certificates (for example if a certificate has been directly
1830 constructed by the application instead of loading it via the OpenSSL
1831 parsing functions, and the certificate contains non NUL terminated
1832 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1833 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1835 If a malicious actor can cause an application to directly construct an
1836 ASN1_STRING and then process it through one of the affected OpenSSL
1837 functions then this issue could be hit. This might result in a crash
1838 (causing a Denial of Service attack). It could also result in the
1839 disclosure of private memory contents (such as private keys, or
1840 sensitive plaintext).
1845 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1847 * Fixed a problem with verifying a certificate chain when using the
1848 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1849 the certificates present in a certificate chain. It is not set by default.
1851 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1852 the chain that have explicitly encoded elliptic curve parameters was added
1853 as an additional strict check.
1855 An error in the implementation of this check meant that the result of a
1856 previous check to confirm that certificates in the chain are valid CA
1857 certificates was overwritten. This effectively bypasses the check
1858 that non-CA certificates must not be able to issue other certificates.
1860 If a "purpose" has been configured then there is a subsequent opportunity
1861 for checks that the certificate is a valid CA. All of the named "purpose"
1862 values implemented in libcrypto perform this check. Therefore, where
1863 a purpose is set the certificate chain will still be rejected even when the
1864 strict flag has been used. A purpose is set by default in libssl client and
1865 server certificate verification routines, but it can be overridden or
1866 removed by an application.
1868 In order to be affected, an application must explicitly set the
1869 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1870 for the certificate verification or, in the case of TLS client or server
1871 applications, override the default purpose.
1876 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1877 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1878 renegotiation ClientHello omits the signature_algorithms extension (where it
1879 was present in the initial ClientHello), but includes a
1880 signature_algorithms_cert extension then a NULL pointer dereference will
1881 result, leading to a crash and a denial of service attack.
1883 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1884 (which is the default configuration). OpenSSL TLS clients are not impacted by
1888 *Peter Kästle and Samuel Sapalski*
1890 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1892 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1893 create a unique hash value based on the issuer and serial number data
1894 contained within an X509 certificate. However it was failing to correctly
1895 handle any errors that may occur while parsing the issuer field (which might
1896 occur if the issuer field is maliciously constructed). This may subsequently
1897 result in a NULL pointer deref and a crash leading to a potential denial of
1903 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1904 padding mode to correctly check for rollback attacks. This is considered a
1905 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1910 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1911 functions. Previously they could overflow the output length argument in some
1912 cases where the input length is close to the maximum permissable length for
1913 an integer on the platform. In such cases the return value from the function
1914 call would be 1 (indicating success), but the output length value would be
1915 negative. This could cause applications to behave incorrectly or crash.
1920 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1921 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1922 could be exploited in a side channel attack to recover the password. Since
1923 the attack is local host only this is outside of the current OpenSSL
1924 threat model and therefore no CVE is assigned.
1926 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1931 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1933 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1934 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1935 If an attacker can control both items being compared then this could lead
1936 to a possible denial of service attack. OpenSSL itself uses the
1937 GENERAL_NAME_cmp function for two purposes:
1938 1) Comparing CRL distribution point names between an available CRL and a
1939 CRL distribution point embedded in an X509 certificate
1940 2) When verifying that a timestamp response token signer matches the
1941 timestamp authority name (exposed via the API functions
1942 TS_RESP_verify_response and TS_RESP_verify_token)
1947 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1949 * Certificates with explicit curve parameters are now disallowed in
1950 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1954 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1955 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1956 conversely, silently ignore DTLS protocol version bounds when configuring
1957 TLS-based contexts. The commands can be repeated to set bounds of both
1958 types. The same applies with the corresponding "min_protocol" and
1959 "max_protocol" command-line switches, in case some application uses both TLS
1962 SSL_CTX instances that are created for a fixed protocol version (e.g.
1963 TLSv1_server_method()) also silently ignore version bounds. Previously
1964 attempts to apply bounds to these protocol versions would result in an
1965 error. Now only the "version-flexible" SSL_CTX instances are subject to
1966 limits in configuration files in command-line options.
1970 * Handshake now fails if Extended Master Secret extension is dropped
1975 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1977 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1979 * Fixed segmentation fault in SSL_check_chain()
1980 Server or client applications that call the SSL_check_chain() function
1981 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1982 dereference as a result of incorrect handling of the
1983 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1984 or unrecognised signature algorithm is received from the peer. This could
1985 be exploited by a malicious peer in a Denial of Service attack.
1990 * Added AES consttime code for no-asm configurations
1991 an optional constant time support for AES was added
1992 when building openssl for no-asm.
1993 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1994 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1995 At this time this feature is by default disabled.
1996 It will be enabled by default in 3.0.
2000 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2002 * Revert the change of EOF detection while reading in libssl to avoid
2003 regressions in applications depending on the current way of reporting
2004 the EOF. As the existing method is not fully accurate the change to
2005 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2006 branch and will be present in the 3.0 release.
2010 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2011 when primes for RSA keys are computed.
2012 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2013 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2014 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2015 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2016 This avoids possible fingerprinting of newly generated RSA modules.
2020 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2022 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2023 while reading in libssl then we would report an error back to the
2024 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2025 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2026 therefore give a hint as to what went wrong.
2030 * Check that ed25519 and ed448 are allowed by the security level. Previously
2031 signature algorithms not using an MD were not being checked that they were
2032 allowed by the security level.
2036 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2037 was not quite right. The behaviour was not consistent between resumption
2038 and normal handshakes, and also not quite consistent with historical
2039 behaviour. The behaviour in various scenarios has been clarified and
2040 it has been updated to make it match historical behaviour as closely as
2045 * *[VMS only]* The header files that the VMS compilers include automatically,
2046 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2047 that the C++ compiler doesn't understand. This is a shortcoming in the
2048 compiler, but can be worked around with `__cplusplus` guards.
2050 C++ applications that use OpenSSL libraries must be compiled using the
2051 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2052 functions. Otherwise, only functions with symbols of less than 31
2053 characters can be used, as the linker will not be able to successfully
2054 resolve symbols with longer names.
2058 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2059 The presence of this system service is determined at run-time.
2063 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2068 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2070 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2071 number generator (RNG). This was intended to include protection in the
2072 event of a fork() system call in order to ensure that the parent and child
2073 processes did not share the same RNG state. However this protection was not
2074 being used in the default case.
2076 A partial mitigation for this issue is that the output from a high
2077 precision timer is mixed into the RNG state so the likelihood of a parent
2078 and child process sharing state is significantly reduced.
2080 If an application already calls OPENSSL_init_crypto() explicitly using
2081 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2084 *Matthias St. Pierre*
2086 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2087 used even when parsing explicit parameters, when loading a encoded key
2088 or calling `EC_GROUP_new_from_ecpkparameters()`/
2089 `EC_GROUP_new_from_ecparameters()`.
2090 This prevents bypass of security hardening and performance gains,
2091 especially for curves with specialized EC_METHODs.
2092 By default, if a key encoded with explicit parameters is loaded and later
2093 encoded, the output is still encoded with explicit parameters, even if
2094 internally a "named" EC_GROUP is used for computation.
2098 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2099 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2100 NULL. After this change, only the cofactor parameter can be NULL. It also
2101 does some minimal sanity checks on the passed order.
2106 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2107 An attack is simple, if the first CMS_recipientInfo is valid but the
2108 second CMS_recipientInfo is chosen ciphertext. If the second
2109 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2110 encryption key will be replaced by garbage, and the message cannot be
2111 decoded, but if the RSA decryption fails, the correct encryption key is
2112 used and the recipient will not notice the attack.
2113 As a work around for this potential attack the length of the decrypted
2114 key must be equal to the cipher default key length, in case the
2115 certifiate is not given and all recipientInfo are tried out.
2116 The old behaviour can be re-enabled in the CMS code by setting the
2117 CMS_DEBUG_DECRYPT flag.
2122 * Early start up entropy quality from the DEVRANDOM seed source has been
2123 improved for older Linux systems. The RAND subsystem will wait for
2124 /dev/random to be producing output before seeding from /dev/urandom.
2125 The seeded state is stored for future library initialisations using
2126 a system global shared memory segment. The shared memory identifier
2127 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2128 the desired value. The default identifier is 114.
2132 * Correct the extended master secret constant on EBCDIC systems. Without this
2133 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2134 negotiate EMS will fail. Unfortunately this also means that TLS connections
2135 between EBCDIC systems with this fix, and EBCDIC systems without this
2136 fix will fail if they negotiate EMS.
2140 * Use Windows installation paths in the mingw builds
2142 Mingw isn't a POSIX environment per se, which means that Windows
2143 paths should be used for installation.
2148 * Changed DH_check to accept parameters with order q and 2q subgroups.
2149 With order 2q subgroups the bit 0 of the private key is not secret
2150 but DH_generate_key works around that by clearing bit 0 of the
2151 private key for those. This avoids leaking bit 0 of the private key.
2155 * Significantly reduce secure memory usage by the randomness pools.
2159 * Revert the DEVRANDOM_WAIT feature for Linux systems
2161 The DEVRANDOM_WAIT feature added a select() call to wait for the
2162 /dev/random device to become readable before reading from the
2163 /dev/urandom device.
2165 It turned out that this change had negative side effects on
2166 performance which were not acceptable. After some discussion it
2167 was decided to revert this feature and leave it up to the OS
2168 resp. the platform maintainer to ensure a proper initialization
2169 during early boot time.
2171 *Matthias St. Pierre*
2173 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2175 * Add build tests for C++. These are generated files that only do one
2176 thing, to include one public OpenSSL head file each. This tests that
2177 the public header files can be usefully included in a C++ application.
2179 This test isn't enabled by default. It can be enabled with the option
2180 'enable-buildtest-c++'.
2184 * Enable SHA3 pre-hashing for ECDSA and DSA.
2188 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2189 This changes the size when using the `genpkey` command when no size is given.
2190 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2191 generation commands to use 2048 bits by default.
2195 * Reorganize the manual pages to consistently have RETURN VALUES,
2196 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2197 util/fix-doc-nits accordingly.
2199 *Paul Yang, Joshua Lock*
2201 * Add the missing accessor EVP_PKEY_get0_engine()
2205 * Have commands like `s_client` and `s_server` output the signature scheme
2206 along with other cipher suite parameters when debugging.
2210 * Make OPENSSL_config() error agnostic again.
2214 * Do the error handling in RSA decryption constant time.
2218 * Prevent over long nonces in ChaCha20-Poly1305.
2220 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2221 for every encryption operation. RFC 7539 specifies that the nonce value
2222 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2223 and front pads the nonce with 0 bytes if it is less than 12
2224 bytes. However it also incorrectly allows a nonce to be set of up to 16
2225 bytes. In this case only the last 12 bytes are significant and any
2226 additional leading bytes are ignored.
2228 It is a requirement of using this cipher that nonce values are
2229 unique. Messages encrypted using a reused nonce value are susceptible to
2230 serious confidentiality and integrity attacks. If an application changes
2231 the default nonce length to be longer than 12 bytes and then makes a
2232 change to the leading bytes of the nonce expecting the new value to be a
2233 new unique nonce then such an application could inadvertently encrypt
2234 messages with a reused nonce.
2236 Additionally the ignored bytes in a long nonce are not covered by the
2237 integrity guarantee of this cipher. Any application that relies on the
2238 integrity of these ignored leading bytes of a long nonce may be further
2239 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2240 is safe because no such use sets such a long nonce value. However user
2241 applications that use this cipher directly and set a non-default nonce
2242 length to be longer than 12 bytes may be vulnerable.
2244 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2250 * Add DEVRANDOM_WAIT feature for Linux systems
2252 On older Linux systems where the getrandom() system call is not available,
2253 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2254 Contrary to getrandom(), the /dev/urandom device will not block during
2255 early boot when the kernel CSPRNG has not been seeded yet.
2257 To mitigate this known weakness, use select() to wait for /dev/random to
2258 become readable before reading from /dev/urandom.
2260 * Ensure that SM2 only uses SM3 as digest algorithm
2264 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2266 * Change the info callback signals for the start and end of a post-handshake
2267 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2268 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2269 confused by this and assume that a TLSv1.2 renegotiation has started. This
2270 can break KeyUpdate handling. Instead we no longer signal the start and end
2271 of a post handshake message exchange (although the messages themselves are
2272 still signalled). This could break some applications that were expecting
2273 the old signals. However without this KeyUpdate is not usable for many
2278 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2280 * Timing vulnerability in DSA signature generation
2282 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2283 timing side channel attack. An attacker could use variations in the signing
2284 algorithm to recover the private key.
2286 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2291 * Timing vulnerability in ECDSA signature generation
2293 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2294 timing side channel attack. An attacker could use variations in the signing
2295 algorithm to recover the private key.
2297 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2302 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2303 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2304 of two gigabytes and the error handling improved.
2306 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2307 categorized as a normal bug, not a security issue, because the DRBG reseeds
2308 automatically and is fully functional even without additional randomness
2309 provided by the application.
2311 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2313 * Add a new ClientHello callback. Provides a callback interface that gives
2314 the application the ability to adjust the nascent SSL object at the
2315 earliest stage of ClientHello processing, immediately after extensions have
2316 been collected but before they have been processed. In particular, this
2317 callback can adjust the supported TLS versions in response to the contents
2322 * Add SM2 base algorithm support.
2326 * s390x assembly pack: add (improved) hardware-support for the following
2327 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2328 aes-cfb/cfb8, aes-ecb.
2332 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2333 parameter is no longer accepted, as it leads to a corrupt table. NULL
2334 pem_str is reserved for alias entries only.
2338 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2339 step for prime curves. The new implementation is based on formulae from
2340 differential addition-and-doubling in homogeneous projective coordinates
2341 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2342 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2343 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2344 to work in projective coordinates.
2346 *Billy Bob Brumley, Nicola Tuveri*
2348 * Change generating and checking of primes so that the error rate of not
2349 being prime depends on the intended use based on the size of the input.
2350 For larger primes this will result in more rounds of Miller-Rabin.
2351 The maximal error rate for primes with more than 1080 bits is lowered
2354 *Kurt Roeckx, Annie Yousar*
2356 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2360 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2361 moving between systems, and to avoid confusion when a Windows build is
2362 done with mingw vs with MSVC. For POSIX installs, there's still a
2363 symlink or copy named 'tsget' to avoid that confusion as well.
2367 * Revert blinding in ECDSA sign and instead make problematic addition
2368 length-invariant. Switch even to fixed-length Montgomery multiplication.
2372 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2373 step for binary curves. The new implementation is based on formulae from
2374 differential addition-and-doubling in mixed Lopez-Dahab projective
2375 coordinates, modified to independently blind the operands.
2377 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2379 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2380 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2381 EC_METHODs to implement their own specialized "ladder step", to take
2382 advantage of more favorable coordinate systems or more efficient
2383 differential addition-and-doubling algorithms.
2385 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2387 * Modified the random device based seed sources to keep the relevant
2388 file descriptors open rather than reopening them on each access.
2389 This allows such sources to operate in a chroot() jail without
2390 the associated device nodes being available. This behaviour can be
2391 controlled using RAND_keep_random_devices_open().
2395 * Numerous side-channel attack mitigations have been applied. This may have
2396 performance impacts for some algorithms for the benefit of improved
2397 security. Specific changes are noted in this change log by their respective
2402 * AIX shared library support overhaul. Switch to AIX "natural" way of
2403 handling shared libraries, which means collecting shared objects of
2404 different versions and bitnesses in one common archive. This allows to
2405 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2406 doesn't affect the way 3rd party applications are linked, only how
2407 multi-version installation is managed.
2411 * Make ec_group_do_inverse_ord() more robust and available to other
2412 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2413 mitigations are applied to the fallback BN_mod_inverse().
2414 When using this function rather than BN_mod_inverse() directly, new
2415 EC cryptosystem implementations are then safer-by-default.
2419 * Add coordinate blinding for EC_POINT and implement projective
2420 coordinate blinding for generic prime curves as a countermeasure to
2421 chosen point SCA attacks.
2423 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2425 * Add blinding to ECDSA and DSA signatures to protect against side channel
2426 attacks discovered by Keegan Ryan (NCC Group).
2430 * Enforce checking in the `pkeyutl` command to ensure that the input
2431 length does not exceed the maximum supported digest length when performing
2432 a sign, verify or verifyrecover operation.
2436 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2437 I/O in combination with something like select() or poll() will hang. This
2438 can be turned off again using SSL_CTX_clear_mode().
2439 Many applications do not properly handle non-application data records, and
2440 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2441 around the problems in those applications, but can also break some.
2442 It's recommended to read the manpages about SSL_read(), SSL_write(),
2443 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2444 SSL_CTX_set_read_ahead() again.
2448 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2449 now allow empty (zero character) pass phrases.
2453 * Apply blinding to binary field modular inversion and remove patent
2454 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2458 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2459 binary and prime elliptic curves.
2463 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2464 constant time fixed point multiplication.
2468 * Revise elliptic curve scalar multiplication with timing attack
2469 defenses: ec_wNAF_mul redirects to a constant time implementation
2470 when computing fixed point and variable point multiplication (which
2471 in OpenSSL are mostly used with secret scalars in keygen, sign,
2472 ECDH derive operations).
2473 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2476 * Updated CONTRIBUTING
2480 * Updated DRBG / RAND to request nonce and additional low entropy
2481 randomness from the system.
2483 *Matthias St. Pierre*
2485 * Updated 'openssl rehash' to use OpenSSL consistent default.
2489 * Moved the load of the ssl_conf module to libcrypto, which helps
2490 loading engines that libssl uses before libssl is initialised.
2494 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2498 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2500 *Ingo Schwarze, Rich Salz*
2502 * Added output of accepting IP address and port for 'openssl s_server'
2506 * Added a new API for TLSv1.3 ciphersuites:
2507 SSL_CTX_set_ciphersuites()
2508 SSL_set_ciphersuites()
2512 * Memory allocation failures consistently add an error to the error
2517 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2518 in libcrypto when run as setuid/setgid.
2522 * Load any config file by default when libssl is used.
2526 * Added new public header file <openssl/rand_drbg.h> and documentation
2527 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2529 *Matthias St. Pierre*
2531 * QNX support removed (cannot find contributors to get their approval
2532 for the license change).
2536 * TLSv1.3 replay protection for early data has been implemented. See the
2537 SSL_read_early_data() man page for further details.
2541 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2542 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2543 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2544 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2545 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2546 configuration has been separated out. See the ciphers man page or the
2547 SSL_CTX_set_ciphersuites() man page for more information.
2551 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2552 in responder mode now supports the new "-multi" option, which
2553 spawns the specified number of child processes to handle OCSP
2554 requests. The "-timeout" option now also limits the OCSP
2555 responder's patience to wait to receive the full client request
2556 on a newly accepted connection. Child processes are respawned
2557 as needed, and the CA index file is automatically reloaded
2558 when changed. This makes it possible to run the "ocsp" responder
2559 as a long-running service, making the OpenSSL CA somewhat more
2560 feature-complete. In this mode, most diagnostic messages logged
2561 after entering the event loop are logged via syslog(3) rather than
2566 * Added support for X448 and Ed448. Heavily based on original work by
2571 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2572 objects loaded. This adds the functions OSSL_STORE_expect() and
2573 OSSL_STORE_find() as well as needed tools to construct searches and
2574 get the search data out of them.
2578 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2579 version of OpenSSL should review their configuration settings to ensure
2580 that they are still appropriate for TLSv1.3. For further information see:
2581 <https://wiki.openssl.org/index.php/TLS1.3>
2585 * Grand redesign of the OpenSSL random generator
2587 The default RAND method now utilizes an AES-CTR DRBG according to
2588 NIST standard SP 800-90Ar1. The new random generator is essentially
2589 a port of the default random generator from the OpenSSL FIPS 2.0
2590 object module. It is a hybrid deterministic random bit generator
2591 using an AES-CTR bit stream and which seeds and reseeds itself
2592 automatically using trusted system entropy sources.
2594 Some of its new features are:
2595 - Support for multiple DRBG instances with seed chaining.
2596 - The default RAND method makes use of a DRBG.
2597 - There is a public and private DRBG instance.
2598 - The DRBG instances are fork-safe.
2599 - Keep all global DRBG instances on the secure heap if it is enabled.
2600 - The public and private DRBG instance are per thread for lock free
2603 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2605 * Changed Configure so it only says what it does and doesn't dump
2606 so much data. Instead, ./configdata.pm should be used as a script
2607 to display all sorts of configuration data.
2611 * Added processing of "make variables" to Configure.
2615 * Added SHA512/224 and SHA512/256 algorithm support.
2619 * The last traces of Netware support, first removed in 1.1.0, have
2624 * Get rid of Makefile.shared, and in the process, make the processing
2625 of certain files (rc.obj, or the .def/.map/.opt files produced from
2626 the ordinal files) more visible and hopefully easier to trace and
2627 debug (or make silent).
2631 * Make it possible to have environment variable assignments as
2632 arguments to config / Configure.
2636 * Add multi-prime RSA (RFC 8017) support.
2640 * Add SM3 implemented according to GB/T 32905-2016
2641 *Jack Lloyd <jack.lloyd@ribose.com>,*
2642 *Ronald Tse <ronald.tse@ribose.com>,*
2643 *Erick Borsboom <erick.borsboom@ribose.com>*
2645 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2646 as documented in RFC6066.
2647 Based on a patch from Tomasz Moń
2649 *Filipe Raimundo da Silva*
2651 * Add SM4 implemented according to GB/T 32907-2016.
2652 *Jack Lloyd <jack.lloyd@ribose.com>,*
2653 *Ronald Tse <ronald.tse@ribose.com>,*
2654 *Erick Borsboom <erick.borsboom@ribose.com>*
2656 * Reimplement -newreq-nodes and ERR_error_string_n; the
2657 original author does not agree with the license change.
2661 * Add ARIA AEAD TLS support.
2665 * Some macro definitions to support VS6 have been removed. Visual
2666 Studio 6 has not worked since 1.1.0
2670 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2671 without clearing the errors.
2675 * Add "atfork" functions. If building on a system that without
2676 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2677 requirements. The RAND facility now uses/requires this.
2685 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2686 not possible to disable entirely. However, it's still possible to
2687 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2690 To disable, configure with 'no-ui-console'. 'no-ui' is still
2691 possible to use as an alias. Check at compile time with the
2692 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2693 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2697 * Add a STORE module, which implements a uniform and URI based reader of
2698 stores that can contain keys, certificates, CRLs and numerous other
2699 objects. The main API is loosely based on a few stdio functions,
2700 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2701 OSSL_STORE_error and OSSL_STORE_close.
2702 The implementation uses backends called "loaders" to implement arbitrary
2703 URI schemes. There is one built in "loader" for the 'file' scheme.
2707 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2708 then adjusted to work on FreeBSD 8.4 as well.
2709 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2710 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2714 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2715 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2716 error code calls like this:
2718 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2720 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2721 that can be encoded in C. For the foreseeable future, this will only
2724 *Richard Levitte and Tim Hudson*
2726 * Removed BSD cryptodev engine.
2730 * Add a build target 'build_all_generated', to build all generated files
2731 and only that. This can be used to prepare everything that requires
2732 things like perl for a system that lacks perl and then move everything
2733 to that system and do the rest of the build there.
2737 * In the UI interface, make it possible to duplicate the user data. This
2738 can be used by engines that need to retain the data for a longer time
2739 than just the call where this user data is passed.
2743 * Ignore the '-named_curve auto' value for compatibility of applications
2746 *Tomáš Mráz <tmraz@fedoraproject.org>*
2748 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2749 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2750 alerts across multiple records (some of which could be empty). In practice
2751 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2752 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2753 support this at all. Supporting it adds significant complexity to the
2754 record layer, and its removal is unlikely to cause interoperability
2759 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2760 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2761 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2766 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2767 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2769 *Richard Levitte, Andy Polyakov*
2771 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2776 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2777 platform rather than 'mingw'.
2781 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2782 success if they are asked to add an object which already exists
2783 in the store. This change cascades to other functions which load
2784 certificates and CRLs.
2788 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2789 facilitate stack unwinding even from assembly subroutines.
2793 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2794 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2798 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2799 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2800 which is the minimum version we support.
2804 * Certificate time validation (X509_cmp_time) enforces stricter
2805 compliance with RFC 5280. Fractional seconds and timezone offsets
2806 are no longer allowed.
2810 * Add support for ARIA
2814 * s_client will now send the Server Name Indication (SNI) extension by
2815 default unless the new "-noservername" option is used. The server name is
2816 based on the host provided to the "-connect" option unless overridden by
2817 using "-servername".
2821 * Add support for SipHash
2825 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2826 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2827 prevent issues where no progress is being made and the peer continually
2828 sends unrecognised record types, using up resources processing them.
2832 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2833 using the algorithm defined in
2834 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2838 * Heartbeat support has been removed; the ABI is changed for now.
2840 *Richard Levitte, Rich Salz*
2842 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2846 * The RSA "null" method, which was partially supported to avoid patent
2847 issues, has been replaced to always returns NULL.
2854 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2856 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2857 used even when parsing explicit parameters, when loading a encoded key
2858 or calling `EC_GROUP_new_from_ecpkparameters()`/
2859 `EC_GROUP_new_from_ecparameters()`.
2860 This prevents bypass of security hardening and performance gains,
2861 especially for curves with specialized EC_METHODs.
2862 By default, if a key encoded with explicit parameters is loaded and later
2863 encoded, the output is still encoded with explicit parameters, even if
2864 internally a "named" EC_GROUP is used for computation.
2868 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2869 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2870 NULL. After this change, only the cofactor parameter can be NULL. It also
2871 does some minimal sanity checks on the passed order.
2876 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2877 An attack is simple, if the first CMS_recipientInfo is valid but the
2878 second CMS_recipientInfo is chosen ciphertext. If the second
2879 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2880 encryption key will be replaced by garbage, and the message cannot be
2881 decoded, but if the RSA decryption fails, the correct encryption key is
2882 used and the recipient will not notice the attack.
2883 As a work around for this potential attack the length of the decrypted
2884 key must be equal to the cipher default key length, in case the
2885 certifiate is not given and all recipientInfo are tried out.
2886 The old behaviour can be re-enabled in the CMS code by setting the
2887 CMS_DEBUG_DECRYPT flag.
2892 * Use Windows installation paths in the mingw builds
2894 Mingw isn't a POSIX environment per se, which means that Windows
2895 paths should be used for installation.
2900 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2902 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2903 This changes the size when using the `genpkey` command when no size is given.
2904 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2905 generation commands to use 2048 bits by default.
2909 * Prevent over long nonces in ChaCha20-Poly1305.
2911 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2912 for every encryption operation. RFC 7539 specifies that the nonce value
2913 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2914 and front pads the nonce with 0 bytes if it is less than 12
2915 bytes. However it also incorrectly allows a nonce to be set of up to 16
2916 bytes. In this case only the last 12 bytes are significant and any
2917 additional leading bytes are ignored.
2919 It is a requirement of using this cipher that nonce values are
2920 unique. Messages encrypted using a reused nonce value are susceptible to
2921 serious confidentiality and integrity attacks. If an application changes
2922 the default nonce length to be longer than 12 bytes and then makes a
2923 change to the leading bytes of the nonce expecting the new value to be a
2924 new unique nonce then such an application could inadvertently encrypt
2925 messages with a reused nonce.
2927 Additionally the ignored bytes in a long nonce are not covered by the
2928 integrity guarantee of this cipher. Any application that relies on the
2929 integrity of these ignored leading bytes of a long nonce may be further
2930 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2931 is safe because no such use sets such a long nonce value. However user
2932 applications that use this cipher directly and set a non-default nonce
2933 length to be longer than 12 bytes may be vulnerable.
2935 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2941 * Added SCA hardening for modular field inversion in EC_GROUP through
2942 a new dedicated field_inv() pointer in EC_METHOD.
2943 This also addresses a leakage affecting conversions from projective
2944 to affine coordinates.
2946 *Billy Bob Brumley, Nicola Tuveri*
2948 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2949 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2953 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2957 * Remove the 'dist' target and add a tarball building script. The
2958 'dist' target has fallen out of use, and it shouldn't be
2959 necessary to configure just to create a source distribution.
2963 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2965 * Timing vulnerability in DSA signature generation
2967 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2968 timing side channel attack. An attacker could use variations in the signing
2969 algorithm to recover the private key.
2971 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2976 * Timing vulnerability in ECDSA signature generation
2978 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2979 timing side channel attack. An attacker could use variations in the signing
2980 algorithm to recover the private key.
2982 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2987 * Add coordinate blinding for EC_POINT and implement projective
2988 coordinate blinding for generic prime curves as a countermeasure to
2989 chosen point SCA attacks.
2991 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2993 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2995 * Client DoS due to large DH parameter
2997 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2998 malicious server can send a very large prime value to the client. This will
2999 cause the client to spend an unreasonably long period of time generating a
3000 key for this prime resulting in a hang until the client has finished. This
3001 could be exploited in a Denial Of Service attack.
3003 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3008 * Cache timing vulnerability in RSA Key Generation
3010 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3011 a cache timing side channel attack. An attacker with sufficient access to
3012 mount cache timing attacks during the RSA key generation process could
3013 recover the private key.
3015 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3016 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3021 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3022 parameter is no longer accepted, as it leads to a corrupt table. NULL
3023 pem_str is reserved for alias entries only.
3027 * Revert blinding in ECDSA sign and instead make problematic addition
3028 length-invariant. Switch even to fixed-length Montgomery multiplication.
3032 * Change generating and checking of primes so that the error rate of not
3033 being prime depends on the intended use based on the size of the input.
3034 For larger primes this will result in more rounds of Miller-Rabin.
3035 The maximal error rate for primes with more than 1080 bits is lowered
3038 *Kurt Roeckx, Annie Yousar*
3040 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3044 * Add blinding to ECDSA and DSA signatures to protect against side channel
3045 attacks discovered by Keegan Ryan (NCC Group).
3049 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3050 now allow empty (zero character) pass phrases.
3054 * Certificate time validation (X509_cmp_time) enforces stricter
3055 compliance with RFC 5280. Fractional seconds and timezone offsets
3056 are no longer allowed.
3060 * Fixed a text canonicalisation bug in CMS
3062 Where a CMS detached signature is used with text content the text goes
3063 through a canonicalisation process first prior to signing or verifying a
3064 signature. This process strips trailing space at the end of lines, converts
3065 line terminators to CRLF and removes additional trailing line terminators
3066 at the end of a file. A bug in the canonicalisation process meant that
3067 some characters, such as form-feed, were incorrectly treated as whitespace
3068 and removed. This is contrary to the specification (RFC5485). This fix
3069 could mean that detached text data signed with an earlier version of
3070 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3071 signed with a fixed OpenSSL may fail to verify with an earlier version of
3072 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3073 and use the "-binary" flag (for the "cms" command line application) or set
3074 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3078 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3080 * Constructed ASN.1 types with a recursive definition could exceed the stack
3082 Constructed ASN.1 types with a recursive definition (such as can be found
3083 in PKCS7) could eventually exceed the stack given malicious input with
3084 excessive recursion. This could result in a Denial Of Service attack. There
3085 are no such structures used within SSL/TLS that come from untrusted sources
3086 so this is considered safe.
3088 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3094 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3096 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3097 effectively reduced to only comparing the least significant bit of each
3098 byte. This allows an attacker to forge messages that would be considered as
3099 authenticated in an amount of tries lower than that guaranteed by the
3100 security claims of the scheme. The module can only be compiled by the
3101 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3103 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3109 * Add a build target 'build_all_generated', to build all generated files
3110 and only that. This can be used to prepare everything that requires
3111 things like perl for a system that lacks perl and then move everything
3112 to that system and do the rest of the build there.
3116 * Backport SSL_OP_NO_RENGOTIATION
3118 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3119 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3120 changes this is no longer possible in 1.1.0. Therefore the new
3121 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3122 1.1.0 to provide equivalent functionality.
3124 Note that if an application built against 1.1.0h headers (or above) is run
3125 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3126 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3130 * Removed the OS390-Unix config target. It relied on a script that doesn't
3135 * rsaz_1024_mul_avx2 overflow bug on x86_64
3137 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3138 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3139 Analysis suggests that attacks against RSA and DSA as a result of this
3140 defect would be very difficult to perform and are not believed likely.
3141 Attacks against DH1024 are considered just feasible, because most of the
3142 work necessary to deduce information about a private key may be performed
3143 offline. The amount of resources required for such an attack would be
3144 significant. However, for an attack on TLS to be meaningful, the server
3145 would have to share the DH1024 private key among multiple clients, which is
3146 no longer an option since CVE-2016-0701.
3148 This only affects processors that support the AVX2 but not ADX extensions
3149 like Intel Haswell (4th generation).
3151 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3152 was originally found via the OSS-Fuzz project.
3157 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3159 * bn_sqrx8x_internal carry bug on x86_64
3161 There is a carry propagating bug in the x86_64 Montgomery squaring
3162 procedure. No EC algorithms are affected. Analysis suggests that attacks
3163 against RSA and DSA as a result of this defect would be very difficult to
3164 perform and are not believed likely. Attacks against DH are considered just
3165 feasible (although very difficult) because most of the work necessary to
3166 deduce information about a private key may be performed offline. The amount
3167 of resources required for such an attack would be very significant and
3168 likely only accessible to a limited number of attackers. An attacker would
3169 additionally need online access to an unpatched system using the target
3170 private key in a scenario with persistent DH parameters and a private
3171 key that is shared between multiple clients.
3173 This only affects processors that support the BMI1, BMI2 and ADX extensions
3174 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3176 This issue was reported to OpenSSL by the OSS-Fuzz project.
3181 * Malformed X.509 IPAddressFamily could cause OOB read
3183 If an X.509 certificate has a malformed IPAddressFamily extension,
3184 OpenSSL could do a one-byte buffer overread. The most likely result
3185 would be an erroneous display of the certificate in text format.
3187 This issue was reported to OpenSSL by the OSS-Fuzz project.
3192 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3194 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3195 platform rather than 'mingw'.
3199 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3200 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3201 which is the minimum version we support.
3205 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3207 * Encrypt-Then-Mac renegotiation crash
3209 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3210 negotiated where it was not in the original handshake (or vice-versa) then
3211 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3212 and servers are affected.
3214 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3219 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3221 * Truncated packet could crash via OOB read
3223 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3224 cipher is being used, then a truncated packet can cause that host to
3225 perform an out-of-bounds read, usually resulting in a crash.
3227 This issue was reported to OpenSSL by Robert Święcki of Google.
3232 * Bad (EC)DHE parameters cause a client crash
3234 If a malicious server supplies bad parameters for a DHE or ECDHE key
3235 exchange then this can result in the client attempting to dereference a
3236 NULL pointer leading to a client crash. This could be exploited in a Denial
3239 This issue was reported to OpenSSL by Guido Vranken.
3244 * BN_mod_exp may produce incorrect results on x86_64
3246 There is a carry propagating bug in the x86_64 Montgomery squaring
3247 procedure. No EC algorithms are affected. Analysis suggests that attacks
3248 against RSA and DSA as a result of this defect would be very difficult to
3249 perform and are not believed likely. Attacks against DH are considered just
3250 feasible (although very difficult) because most of the work necessary to
3251 deduce information about a private key may be performed offline. The amount
3252 of resources required for such an attack would be very significant and
3253 likely only accessible to a limited number of attackers. An attacker would
3254 additionally need online access to an unpatched system using the target
3255 private key in a scenario with persistent DH parameters and a private
3256 key that is shared between multiple clients. For example this can occur by
3257 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3258 similar to CVE-2015-3193 but must be treated as a separate problem.
3260 This issue was reported to OpenSSL by the OSS-Fuzz project.
3265 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3267 * ChaCha20/Poly1305 heap-buffer-overflow
3269 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3270 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3271 crash. This issue is not considered to be exploitable beyond a DoS.
3273 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3278 * CMS Null dereference
3280 Applications parsing invalid CMS structures can crash with a NULL pointer
3281 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3282 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3283 structure callback if an attempt is made to free certain invalid encodings.
3284 Only CHOICE structures using a callback which do not handle NULL value are
3287 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3292 * Montgomery multiplication may produce incorrect results
3294 There is a carry propagating bug in the Broadwell-specific Montgomery
3295 multiplication procedure that handles input lengths divisible by, but
3296 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3297 and DH private keys are impossible. This is because the subroutine in
3298 question is not used in operations with the private key itself and an input
3299 of the attacker's direct choice. Otherwise the bug can manifest itself as
3300 transient authentication and key negotiation failures or reproducible
3301 erroneous outcome of public-key operations with specially crafted input.
3302 Among EC algorithms only Brainpool P-512 curves are affected and one
3303 presumably can attack ECDH key negotiation. Impact was not analyzed in
3304 detail, because pre-requisites for attack are considered unlikely. Namely
3305 multiple clients have to choose the curve in question and the server has to
3306 share the private key among them, neither of which is default behaviour.
3307 Even then only clients that chose the curve will be affected.
3309 This issue was publicly reported as transient failures and was not
3310 initially recognized as a security issue. Thanks to Richard Morgan for
3311 providing reproducible case.
3316 * Removed automatic addition of RPATH in shared libraries and executables,
3317 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3321 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3323 * Fix Use After Free for large message sizes
3325 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3326 message larger than approx 16k is received then the underlying buffer to
3327 store the incoming message is reallocated and moved. Unfortunately a
3328 dangling pointer to the old location is left which results in an attempt to
3329 write to the previously freed location. This is likely to result in a
3330 crash, however it could potentially lead to execution of arbitrary code.
3332 This issue only affects OpenSSL 1.1.0a.
3334 This issue was reported to OpenSSL by Robert Święcki.
3339 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3341 * OCSP Status Request extension unbounded memory growth
3343 A malicious client can send an excessively large OCSP Status Request
3344 extension. If that client continually requests renegotiation, sending a
3345 large OCSP Status Request extension each time, then there will be unbounded
3346 memory growth on the server. This will eventually lead to a Denial Of
3347 Service attack through memory exhaustion. Servers with a default
3348 configuration are vulnerable even if they do not support OCSP. Builds using
3349 the "no-ocsp" build time option are not affected.
3351 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3356 * SSL_peek() hang on empty record
3358 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3359 sends an empty record. This could be exploited by a malicious peer in a
3360 Denial Of Service attack.
3362 This issue was reported to OpenSSL by Alex Gaynor.
3367 * Excessive allocation of memory in tls_get_message_header() and
3368 dtls1_preprocess_fragment()
3370 A (D)TLS message includes 3 bytes for its length in the header for the
3371 message. This would allow for messages up to 16Mb in length. Messages of
3372 this length are excessive and OpenSSL includes a check to ensure that a
3373 peer is sending reasonably sized messages in order to avoid too much memory
3374 being consumed to service a connection. A flaw in the logic of version
3375 1.1.0 means that memory for the message is allocated too early, prior to
3376 the excessive message length check. Due to way memory is allocated in
3377 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3378 to service a connection. This could lead to a Denial of Service through
3379 memory exhaustion. However, the excessive message length check still takes
3380 place, and this would cause the connection to immediately fail. Assuming
3381 that the application calls SSL_free() on the failed connection in a timely
3382 manner then the 21Mb of allocated memory will then be immediately freed
3383 again. Therefore the excessive memory allocation will be transitory in
3384 nature. This then means that there is only a security impact if:
3386 1) The application does not call SSL_free() in a timely manner in the event
3387 that the connection fails
3389 2) The application is working in a constrained environment where there is
3390 very little free memory
3392 3) The attacker initiates multiple connection attempts such that there are
3393 multiple connections in a state where memory has been allocated for the
3394 connection; SSL_free() has not yet been called; and there is insufficient
3395 memory to service the multiple requests.
3397 Except in the instance of (1) above any Denial Of Service is likely to be
3398 transitory because as soon as the connection fails the memory is
3399 subsequently freed again in the SSL_free() call. However there is an
3400 increased risk during this period of application crashes due to the lack of
3401 memory - which would then mean a more serious Denial of Service.
3403 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3404 (CVE-2016-6307 and CVE-2016-6308)
3408 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3409 had to be removed. Primary reason is that vendor assembler can't
3410 assemble our modules with -KPIC flag. As result it, assembly
3411 support, was not even available as option. But its lack means
3412 lack of side-channel resistant code, which is incompatible with
3413 security by todays standards. Fortunately gcc is readily available
3414 prepackaged option, which we firmly point at...
3418 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3420 * Windows command-line tool supports UTF-8 opt-in option for arguments
3421 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3422 (to any value) allows Windows user to access PKCS#12 file generated
3423 with Windows CryptoAPI and protected with non-ASCII password, as well
3424 as files generated under UTF-8 locale on Linux also protected with
3429 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3430 have been disabled by default and removed from DEFAULT, just like RC4.
3431 See the RC4 item below to re-enable both.
3435 * The method for finding the storage location for the Windows RAND seed file
3436 has changed. First we check %RANDFILE%. If that is not set then we check
3437 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3438 all else fails we fall back to C:\.
3442 * The EVP_EncryptUpdate() function has had its return type changed from void
3443 to int. A return of 0 indicates and error while a return of 1 indicates
3448 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3449 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3450 off the constant time implementation for RSA, DSA and DH have been made
3451 no-ops and deprecated.
3455 * Windows RAND implementation was simplified to only get entropy by
3456 calling CryptGenRandom(). Various other RAND-related tickets
3459 *Joseph Wylie Yandle, Rich Salz*
3461 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3462 and `OPENSSL_LH_`, respectively. The old names are available
3463 with API compatibility. They new names are now completely documented.
3467 * Unify TYPE_up_ref(obj) methods signature.
3468 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3469 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3470 int (instead of void) like all others TYPE_up_ref() methods.
3471 So now these methods also check the return value of CRYPTO_atomic_add(),
3472 and the validity of object reference counter.
3474 *fdasilvayy@gmail.com*
3476 * With Windows Visual Studio builds, the .pdb files are installed
3477 alongside the installed libraries and executables. For a static
3478 library installation, ossl_static.pdb is the associate compiler
3479 generated .pdb file to be used when linking programs.
3483 * Remove openssl.spec. Packaging files belong with the packagers.
3487 * Automatic Darwin/OSX configuration has had a refresh, it will now
3488 recognise x86_64 architectures automatically. You can still decide
3489 to build for a different bitness with the environment variable
3490 KERNEL_BITS (can be 32 or 64), for example:
3492 KERNEL_BITS=32 ./config
3496 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3497 256 bit AES and HMAC with SHA256.
3501 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3505 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3509 * To enable users to have their own config files and build file templates,
3510 Configure looks in the directory indicated by the environment variable
3511 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3512 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3513 name and is used as is.
3517 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3518 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3519 X509_CERT_FILE_CTX was removed.
3523 * "shared" builds are now the default. To create only static libraries use
3524 the "no-shared" Configure option.
3528 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3529 All of these option have not worked for some while and are fundamental
3534 * Make various cleanup routines no-ops and mark them as deprecated. Most
3535 global cleanup functions are no longer required because they are handled
3536 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3537 Explicitly de-initing can cause problems (e.g. where a library that uses
3538 OpenSSL de-inits, but an application is still using it). The affected
3539 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3540 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3541 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3542 COMP_zlib_cleanup().
3546 * --strict-warnings no longer enables runtime debugging options
3547 such as REF_DEBUG. Instead, debug options are automatically
3548 enabled with '--debug' builds.
3550 *Andy Polyakov, Emilia Käsper*
3552 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3553 have been moved out of the public header files. New functions for managing
3554 these have been added.
3558 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3559 objects have been moved out of the public header files. New
3560 functions for managing these have been added.
3564 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3565 have been moved out of the public header files. New functions for managing
3566 these have been added.
3570 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3571 moved out of the public header files. New functions for managing these
3576 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3580 * Removed the mk1mf build scripts.
3584 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3585 it is always safe to #include a header now.
3589 * Removed the aged BC-32 config and all its supporting scripts
3593 * Removed support for Ultrix, Netware, and OS/2.
3597 * Add support for HKDF.
3599 *Alessandro Ghedini*
3601 * Add support for blake2b and blake2s
3605 * Added support for "pipelining". Ciphers that have the
3606 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3607 encryptions/decryptions simultaneously. There are currently no built-in
3608 ciphers with this property but the expectation is that engines will be able
3609 to offer it to significantly improve throughput. Support has been extended
3610 into libssl so that multiple records for a single connection can be
3611 processed in one go (for >=TLS 1.1).
3615 * Added the AFALG engine. This is an async capable engine which is able to
3616 offload work to the Linux kernel. In this initial version it only supports
3617 AES128-CBC. The kernel must be version 4.1.0 or greater.
3621 * OpenSSL now uses a new threading API. It is no longer necessary to
3622 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3623 are two supported threading models: pthreads and windows threads. It is
3624 also possible to configure OpenSSL at compile time for "no-threads". The
3625 old threading API should no longer be used. The functions have been
3626 replaced with "no-op" compatibility macros.
3628 *Alessandro Ghedini, Matt Caswell*
3630 * Modify behavior of ALPN to invoke callback after SNI/servername
3631 callback, such that updates to the SSL_CTX affect ALPN.
3635 * Add SSL_CIPHER queries for authentication and key-exchange.
3639 * Changes to the DEFAULT cipherlist:
3640 - Prefer (EC)DHE handshakes over plain RSA.
3641 - Prefer AEAD ciphers over legacy ciphers.
3642 - Prefer ECDSA over RSA when both certificates are available.
3643 - Prefer TLSv1.2 ciphers/PRF.
3644 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3649 * Change the ECC default curve list to be this, in order: x25519,
3650 secp256r1, secp521r1, secp384r1.
3654 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3655 disabled by default. They can be re-enabled using the
3656 enable-weak-ssl-ciphers option to Configure.
3660 * If the server has ALPN configured, but supports no protocols that the
3661 client advertises, send a fatal "no_application_protocol" alert.
3662 This behaviour is SHALL in RFC 7301, though it isn't universally
3663 implemented by other servers.
3667 * Add X25519 support.
3668 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3669 for public and private key encoding using the format documented in
3670 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3671 key generation and key derivation.
3673 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3678 * Deprecate SRP_VBASE_get_by_user.
3679 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3680 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3681 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3682 seed, even if the seed is configured.
3684 Users should use SRP_VBASE_get1_by_user instead. Note that in
3685 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3686 also that even though configuring the SRP seed attempts to hide
3687 invalid usernames by continuing the handshake with fake
3688 credentials, this behaviour is not constant time and no strong
3689 guarantees are made that the handshake is indistinguishable from
3690 that of a valid user.
3694 * Configuration change; it's now possible to build dynamic engines
3695 without having to build shared libraries and vice versa. This
3696 only applies to the engines in `engines/`, those in `crypto/engine/`
3697 will always be built into libcrypto (i.e. "static").
3699 Building dynamic engines is enabled by default; to disable, use
3700 the configuration option "disable-dynamic-engine".
3702 The only requirements for building dynamic engines are the
3703 presence of the DSO module and building with position independent
3704 code, so they will also automatically be disabled if configuring
3705 with "disable-dso" or "disable-pic".
3707 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3708 are also taken away from openssl/opensslconf.h, as they are
3713 * Configuration change; if there is a known flag to compile
3714 position independent code, it will always be applied on the
3715 libcrypto and libssl object files, and never on the application
3716 object files. This means other libraries that use routines from
3717 libcrypto / libssl can be made into shared libraries regardless
3718 of how OpenSSL was configured.
3720 If this isn't desirable, the configuration options "disable-pic"
3721 or "no-pic" can be used to disable the use of PIC. This will
3722 also disable building shared libraries and dynamic engines.
3726 * Removed JPAKE code. It was experimental and has no wide use.
3730 * The INSTALL_PREFIX Makefile variable has been renamed to
3731 DESTDIR. That makes for less confusion on what this variable
3732 is for. Also, the configuration option --install_prefix is
3737 * Heartbeat for TLS has been removed and is disabled by default
3738 for DTLS; configure with enable-heartbeats. Code that uses the
3739 old #define's might need to be updated.
3741 *Emilia Käsper, Rich Salz*
3743 * Rename REF_CHECK to REF_DEBUG.
3747 * New "unified" build system
3749 The "unified" build system is aimed to be a common system for all
3750 platforms we support. With it comes new support for VMS.
3752 This system builds supports building in a different directory tree
3753 than the source tree. It produces one Makefile (for unix family
3754 or lookalikes), or one descrip.mms (for VMS).
3756 The source of information to make the Makefile / descrip.mms is
3757 small files called 'build.info', holding the necessary
3758 information for each directory with source to compile, and a
3759 template in Configurations, like unix-Makefile.tmpl or
3762 With this change, the library names were also renamed on Windows
3763 and on VMS. They now have names that are closer to the standard
3764 on Unix, and include the major version number, and in certain
3765 cases, the architecture they are built for. See "Notes on shared
3766 libraries" in INSTALL.
3768 We rely heavily on the perl module Text::Template.
3772 * Added support for auto-initialisation and de-initialisation of the library.
3773 OpenSSL no longer requires explicit init or deinit routines to be called,
3774 except in certain circumstances. See the OPENSSL_init_crypto() and
3775 OPENSSL_init_ssl() man pages for further information.
3779 * The arguments to the DTLSv1_listen function have changed. Specifically the
3780 "peer" argument is now expected to be a BIO_ADDR object.
3782 * Rewrite of BIO networking library. The BIO library lacked consistent
3783 support of IPv6, and adding it required some more extensive
3784 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3785 which hold all types of addresses and chains of address information.
3786 It also introduces a new API, with functions like BIO_socket,
3787 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3788 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3789 have been adapted accordingly.
3793 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3798 * CRIME protection: disable compression by default, even if OpenSSL is
3799 compiled with zlib enabled. Applications can still enable compression
3800 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3801 using the SSL_CONF library to configure compression.
3805 * The signature of the session callback configured with
3806 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3807 was explicitly marked as `const unsigned char*` instead of
3812 * Always DPURIFY. Remove the use of uninitialized memory in the
3813 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3817 * Removed many obsolete configuration items, including
3818 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3819 MD2_CHAR, MD2_INT, MD2_LONG
3821 IDEA_SHORT, IDEA_LONG
3822 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3824 *Rich Salz, with advice from Andy Polyakov*
3826 * Many BN internals have been moved to an internal header file.
3828 *Rich Salz with help from Andy Polyakov*
3830 * Configuration and writing out the results from it has changed.
3831 Files such as Makefile include/openssl/opensslconf.h and are now
3832 produced through general templates, such as Makefile.in and
3833 crypto/opensslconf.h.in and some help from the perl module
3836 Also, the center of configuration information is no longer
3837 Makefile. Instead, Configure produces a perl module in
3838 configdata.pm which holds most of the config data (in the hash
3839 table %config), the target data that comes from the target
3840 configuration in one of the `Configurations/*.conf` files (in
3845 * To clarify their intended purposes, the Configure options
3846 --prefix and --openssldir change their semantics, and become more
3847 straightforward and less interdependent.
3849 --prefix shall be used exclusively to give the location INSTALLTOP
3850 where programs, scripts, libraries, include files and manuals are
3851 going to be installed. The default is now /usr/local.
3853 --openssldir shall be used exclusively to give the default
3854 location OPENSSLDIR where certificates, private keys, CRLs are
3855 managed. This is also where the default openssl.cnf gets
3857 If the directory given with this option is a relative path, the
3858 values of both the --prefix value and the --openssldir value will
3859 be combined to become OPENSSLDIR.
3860 The default for --openssldir is INSTALLTOP/ssl.
3862 Anyone who uses --openssldir to specify where OpenSSL is to be
3863 installed MUST change to use --prefix instead.
3867 * The GOST engine was out of date and therefore it has been removed. An up
3868 to date GOST engine is now being maintained in an external repository.
3869 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3870 support for GOST ciphersuites (these are only activated if a GOST engine
3875 * EGD is no longer supported by default; use enable-egd when
3878 *Ben Kaduk and Rich Salz*
3880 * The distribution now has Makefile.in files, which are used to
3881 create Makefile's when Configure is run. *Configure must be run
3882 before trying to build now.*
3886 * The return value for SSL_CIPHER_description() for error conditions
3891 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3893 Obtaining and performing DNSSEC validation of TLSA records is
3894 the application's responsibility. The application provides
3895 the TLSA records of its choice to OpenSSL, and these are then
3896 used to authenticate the peer.
3898 The TLSA records need not even come from DNS. They can, for
3899 example, be used to implement local end-entity certificate or
3900 trust-anchor "pinning", where the "pin" data takes the form
3901 of TLSA records, which can augment or replace verification
3902 based on the usual WebPKI public certification authorities.
3906 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3907 continues to support deprecated interfaces in default builds.
3908 However, applications are strongly advised to compile their
3909 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3910 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3911 or the 1.1.0 releases.
3913 In environments in which all applications have been ported to
3914 not use any deprecated interfaces OpenSSL's Configure script
3915 should be used with the --api=1.1.0 option to entirely remove
3916 support for the deprecated features from the library and
3917 unconditionally disable them in the installed headers.
3918 Essentially the same effect can be achieved with the "no-deprecated"
3919 argument to Configure, except that this will always restrict
3920 the build to just the latest API, rather than a fixed API
3923 As applications are ported to future revisions of the API,
3924 they should update their compile-time OPENSSL_API_COMPAT define
3925 accordingly, but in most cases should be able to continue to
3926 compile with later releases.
3928 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3929 0x10000000L and 0x00908000L, respectively. However those
3930 versions did not support the OPENSSL_API_COMPAT feature, and
3931 so applications are not typically tested for explicit support
3932 of just the undeprecated features of either release.
3936 * Add support for setting the minimum and maximum supported protocol.
3937 It can bet set via the SSL_set_min_proto_version() and
3938 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3939 MaxProtocol. It's recommended to use the new APIs to disable
3940 protocols instead of disabling individual protocols using
3941 SSL_set_options() or SSL_CONF's Protocol. This change also
3942 removes support for disabling TLS 1.2 in the OpenSSL TLS
3943 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3947 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3951 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3952 and integrates ECDSA and ECDH functionality into EC. Implementations can
3953 now redirect key generation and no longer need to convert to or from
3956 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3957 include the ec.h header file instead.
3961 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3962 ciphers who are no longer supported and drops support the ephemeral RSA key
3963 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3967 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3968 opaque. For HMAC_CTX, the following constructors and destructors
3971 HMAC_CTX *HMAC_CTX_new(void);
3972 void HMAC_CTX_free(HMAC_CTX *ctx);
3974 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3975 destroy such methods has been added. See EVP_MD_meth_new(3) and
3976 EVP_CIPHER_meth_new(3) for documentation.
3979 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3980 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3981 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3982 an already created structure.
3983 2) For consistency with the majority of our object creators and
3984 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3985 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3986 for deprecated builds.
3990 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3991 cryptographic operations to be performed asynchronously as long as an
3992 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3993 further details. Libssl has also had this capability integrated with the
3994 introduction of the new mode SSL_MODE_ASYNC and associated error
3995 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3996 pages. This work was developed in partnership with Intel Corp.
4000 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4001 always enabled now. If you want to disable the support you should
4002 exclude it using the list of supported ciphers. This also means that the
4003 "-no_ecdhe" option has been removed from s_server.
4007 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4008 SSL_{CTX_}set1_curves() which can set a list.
4012 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4013 curve you want to support using SSL_{CTX_}set1_curves().
4017 * State machine rewrite. The state machine code has been significantly
4018 refactored in order to remove much duplication of code and solve issues
4019 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4020 further details). This change does have some associated API changes.
4021 Notably the SSL_state() function has been removed and replaced by
4022 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4023 SSL_set_state() has been removed altogether. The previous handshake states
4024 defined in ssl.h and ssl3.h have also been removed.
4028 * All instances of the string "ssleay" in the public API were replaced
4029 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4030 Some error codes related to internal RSA_eay API's were renamed.
4034 * The demo files in crypto/threads were moved to demo/threads.
4038 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4041 *Matt Caswell, Rich Salz*
4043 * New ASN.1 embed macro.
4045 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4046 structure is not allocated: it is part of the parent. That is instead of
4054 This reduces memory fragmentation and make it impossible to accidentally
4055 set a mandatory field to NULL.
4057 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4058 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4059 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4064 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4068 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4069 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4070 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4071 DES and RC4 ciphersuites.
4075 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4076 This changes the decoding behaviour for some invalid messages,
4077 though the change is mostly in the more lenient direction, and
4078 legacy behaviour is preserved as much as possible.
4082 * Fix no-stdio build.
4083 *David Woodhouse <David.Woodhouse@intel.com> and also*
4084 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4086 * New testing framework
4087 The testing framework has been largely rewritten and is now using
4088 perl and the perl modules Test::Harness and an extended variant of
4089 Test::More called OpenSSL::Test to do its work. All test scripts in
4090 test/ have been rewritten into test recipes, and all direct calls to
4091 executables in test/Makefile have become individual recipes using the
4092 simplified testing OpenSSL::Test::Simple.
4094 For documentation on our testing modules, do:
4096 perldoc test/testlib/OpenSSL/Test/Simple.pm
4097 perldoc test/testlib/OpenSSL/Test.pm
4101 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4102 are used; the latter aborts on memory leaks (usually checked on exit).
4103 Some undocumented "set malloc, etc., hooks" functions were removed
4104 and others were changed. All are now documented.
4108 * In DSA_generate_parameters_ex, if the provided seed is too short,
4111 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4113 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4114 from RFC4279, RFC4785, RFC5487, RFC5489.
4116 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4117 original RSA_PSK patch.
4121 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4122 era flag was never set throughout the codebase (only read). Also removed
4123 SSL3_FLAGS_POP_BUFFER which was only used if
4124 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4128 * Changed the default name options in the "ca", "crl", "req" and "x509"
4129 to be "oneline" instead of "compat".
4133 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4134 not aware of clients that still exhibit this bug, and the workaround
4135 hasn't been working properly for a while.
4139 * The return type of BIO_number_read() and BIO_number_written() as well as
4140 the corresponding num_read and num_write members in the BIO structure has
4141 changed from unsigned long to uint64_t. On platforms where an unsigned
4142 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4147 * Given the pervasive nature of TLS extensions it is inadvisable to run
4148 OpenSSL without support for them. It also means that maintaining
4149 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4150 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4154 * Removed support for the two export grade static DH ciphersuites
4155 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4156 were newly added (along with a number of other static DH ciphersuites) to
4157 1.0.2. However the two export ones have *never* worked since they were
4158 introduced. It seems strange in any case to be adding new export
4159 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4163 * Version negotiation has been rewritten. In particular SSLv23_method(),
4164 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4165 and turned into macros which simply call the new preferred function names
4166 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4167 should use the new names instead. Also as part of this change the ssl23.h
4168 header file has been removed.
4172 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4173 code and the associated standard is no longer considered fit-for-purpose.
4177 * RT2547 was closed. When generating a private key, try to make the
4178 output file readable only by the owner. This behavior change might
4179 be noticeable when interacting with other software.
4181 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4186 * Added HTTP GET support to the ocsp command.
4190 * Changed default digest for the dgst and enc commands from MD5 to
4195 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4199 * Added support for TLS extended master secret from
4200 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4201 initial patch which was a great help during development.
4205 * All libssl internal structures have been removed from the public header
4206 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4207 now redundant). Users should not attempt to access internal structures
4208 directly. Instead they should use the provided API functions.
4212 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4213 Access to deprecated functions can be re-enabled by running config with
4214 "enable-deprecated". In addition applications wishing to use deprecated
4215 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4216 will, by default, disable some transitive includes that previously existed
4217 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4221 * Added support for OCB mode. OpenSSL has been granted a patent license
4222 compatible with the OpenSSL license for use of OCB. Details are available
4223 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4224 for OCB can be removed by calling config with no-ocb.
4228 * SSLv2 support has been removed. It still supports receiving a SSLv2
4229 compatible client hello.
4233 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4234 done while fixing the error code for the key-too-small case.
4236 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4238 * CA.sh has been removed; use CA.pl instead.
4242 * Removed old DES API.
4246 * Remove various unsupported platforms:
4252 Sinix/ReliantUNIX RM400
4257 16-bit platforms such as WIN16
4261 * Clean up OPENSSL_NO_xxx #define's
4262 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4263 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4264 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4265 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4266 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4267 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4268 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4269 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4270 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4271 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4275 * Cleaned up dead code
4276 Remove all but one '#ifdef undef' which is to be looked at.
4280 * Clean up calling of xxx_free routines.
4281 Just like free(), fix most of the xxx_free routines to accept
4282 NULL. Remove the non-null checks from callers. Save much code.
4286 * Add secure heap for storage of private keys (when possible).
4287 Add BIO_s_secmem(), CBIGNUM, etc.
4288 Contributed by Akamai Technologies under our Corporate CLA.
4292 * Experimental support for a new, fast, unbiased prime candidate generator,
4293 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4295 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4297 * New output format NSS in the sess_id command line tool. This allows
4298 exporting the session id and the master key in NSS keylog format.
4300 *Martin Kaiser <martin@kaiser.cx>*
4302 * Harmonize version and its documentation. -f flag is used to display
4305 *mancha <mancha1@zoho.com>*
4307 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4308 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4310 *mancha <mancha1@zoho.com>*
4312 * Fix some double frees. These are not thought to be exploitable.
4314 *mancha <mancha1@zoho.com>*
4316 * A missing bounds check in the handling of the TLS heartbeat extension
4317 can be used to reveal up to 64k of memory to a connected client or
4320 Thanks for Neel Mehta of Google Security for discovering this bug and to
4321 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4322 preparing the fix ([CVE-2014-0160])
4324 *Adam Langley, Bodo Moeller*
4326 * Fix for the attack described in the paper "Recovering OpenSSL
4327 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4328 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4329 <http://eprint.iacr.org/2014/140>
4331 Thanks to Yuval Yarom and Naomi Benger for discovering this
4332 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4334 *Yuval Yarom and Naomi Benger*
4336 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4337 this fixes a limitation in previous versions of OpenSSL.
4341 * Experimental encrypt-then-mac support.
4343 Experimental support for encrypt then mac from
4344 draft-gutmann-tls-encrypt-then-mac-02.txt
4346 To enable it set the appropriate extension number (0x42 for the test
4347 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4349 For non-compliant peers (i.e. just about everything) this should have no
4352 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4356 * Add EVP support for key wrapping algorithms, to avoid problems with
4357 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4358 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4359 algorithms and include tests cases.
4363 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4368 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4369 MGF1 digest and OAEP label.
4373 * Make openssl verify return errors.
4375 *Chris Palmer <palmer@google.com> and Ben Laurie*
4377 * New function ASN1_TIME_diff to calculate the difference between two
4378 ASN1_TIME structures or one structure and the current time.
4382 * Update fips_test_suite to support multiple command line options. New
4383 test to induce all self test errors in sequence and check expected
4388 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4389 sign or verify all in one operation.
4393 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4394 test programs and fips_test_suite. Includes functionality to parse
4395 the minimal script output of fipsalgest.pl directly.
4399 * Add authorisation parameter to FIPS_module_mode_set().
4403 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4407 * Use separate DRBG fields for internal and external flags. New function
4408 FIPS_drbg_health_check() to perform on demand health checking. Add
4409 generation tests to fips_test_suite with reduced health check interval to
4410 demonstrate periodic health checking. Add "nodh" option to
4411 fips_test_suite to skip very slow DH test.
4415 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4420 * More extensive health check for DRBG checking many more failure modes.
4421 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4422 combination: call this in fips_test_suite.
4426 * Add support for canonical generation of DSA parameter 'g'. See
4429 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4430 POST to handle HMAC cases.
4434 * Add functions FIPS_module_version() and FIPS_module_version_text()
4435 to return numerical and string versions of the FIPS module number.
4439 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4440 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4441 outside the validated module in the FIPS capable OpenSSL.
4445 * Minor change to DRBG entropy callback semantics. In some cases
4446 there is no multiple of the block length between min_len and
4447 max_len. Allow the callback to return more than max_len bytes
4448 of entropy but discard any extra: it is the callback's responsibility
4449 to ensure that the extra data discarded does not impact the
4450 requested amount of entropy.
4454 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4455 information in FIPS186-3, SP800-57 and SP800-131A.
4459 * CCM support via EVP. Interface is very similar to GCM case except we
4460 must supply all data in one chunk (i.e. no update, final) and the
4461 message length must be supplied if AAD is used. Add algorithm test
4466 * Initial version of POST overhaul. Add POST callback to allow the status
4467 of POST to be monitored and/or failures induced. Modify fips_test_suite
4468 to use callback. Always run all selftests even if one fails.
4472 * XTS support including algorithm test driver in the fips_gcmtest program.
4473 Note: this does increase the maximum key length from 32 to 64 bytes but
4474 there should be no binary compatibility issues as existing applications
4475 will never use XTS mode.
4479 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4480 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4481 performs algorithm blocking for unapproved PRNG types. Also do not
4482 set PRNG type in FIPS_mode_set(): leave this to the application.
4483 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4484 the standard OpenSSL PRNG: set additional data to a date time vector.
4488 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4489 This shouldn't present any incompatibility problems because applications
4490 shouldn't be using these directly and any that are will need to rethink
4491 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4495 * Extensive self tests and health checking required by SP800-90 DRBG.
4496 Remove strength parameter from FIPS_drbg_instantiate and always
4497 instantiate at maximum supported strength.
4501 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4505 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4509 * New function DH_compute_key_padded() to compute a DH key and pad with
4510 leading zeroes if needed: this complies with SP800-56A et al.
4514 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4515 anything, incomplete, subject to change and largely untested at present.
4519 * Modify fipscanisteronly build option to only build the necessary object
4520 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4524 * Add experimental option FIPSSYMS to give all symbols in
4525 fipscanister.o and FIPS or fips prefix. This will avoid
4526 conflicts with future versions of OpenSSL. Add perl script
4527 util/fipsas.pl to preprocess assembly language source files
4528 and rename any affected symbols.
4532 * Add selftest checks and algorithm block of non-fips algorithms in
4533 FIPS mode. Remove DES2 from selftests.
4537 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4538 return internal method without any ENGINE dependencies. Add new
4539 tiny fips sign and verify functions.
4543 * New build option no-ec2m to disable characteristic 2 code.
4547 * New build option "fipscanisteronly". This only builds fipscanister.o
4548 and (currently) associated fips utilities. Uses the file Makefile.fips
4549 instead of Makefile.org as the prototype.
4553 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4554 Update fips_gcmtest to use IV generator.
4558 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4559 setting output buffer to NULL. The `*Final` function must be
4560 called although it will not retrieve any additional data. The tag
4561 can be set or retrieved with a ctrl. The IV length is by default 12
4562 bytes (96 bits) but can be set to an alternative value. If the IV
4563 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4568 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4569 underlying do_cipher function handles all cipher semantics itself
4570 including padding and finalisation. This is useful if (for example)
4571 an ENGINE cipher handles block padding itself. The behaviour of
4572 do_cipher is subtly changed if this flag is set: the return value
4573 is the number of characters written to the output buffer (zero is
4574 no longer an error code) or a negative error code. Also if the
4575 input buffer is NULL and length 0 finalisation should be performed.
4579 * If a candidate issuer certificate is already part of the constructed
4580 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4584 * Improve forward-security support: add functions
4586 void SSL_CTX_set_not_resumable_session_callback(
4587 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4588 void SSL_set_not_resumable_session_callback(
4589 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4591 for use by SSL/TLS servers; the callback function will be called whenever a
4592 new session is created, and gets to decide whether the session may be
4593 cached to make it resumable (return 0) or not (return 1). (As by the
4594 SSL/TLS protocol specifications, the session_id sent by the server will be
4595 empty to indicate that the session is not resumable; also, the server will
4596 not generate RFC 4507 (RFC 5077) session tickets.)
4598 A simple reasonable callback implementation is to return is_forward_secure.
4599 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4600 by the SSL/TLS server library, indicating whether it can provide forward
4603 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4605 * New -verify_name option in command line utilities to set verification
4610 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4611 Add CMAC pkey methods.
4615 * Experimental renegotiation in s_server -www mode. If the client
4616 browses /reneg connection is renegotiated. If /renegcert it is
4617 renegotiated requesting a certificate.
4621 * Add an "external" session cache for debugging purposes to s_server. This
4622 should help trace issues which normally are only apparent in deployed
4623 multi-process servers.
4627 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4628 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4629 BIO_set_cipher() and some obscure PEM functions were changed so they
4630 can now return an error. The RAND changes required a change to the
4631 RAND_METHOD structure.
4635 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4636 a gcc attribute to warn if the result of a function is ignored. This
4637 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4638 whose return value is often ignored.
4642 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4643 These allow SCTs (signed certificate timestamps) to be requested and
4644 validated when establishing a connection.
4646 *Rob Percival <robpercival@google.com>*
4651 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4653 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4654 used even when parsing explicit parameters, when loading a encoded key
4655 or calling `EC_GROUP_new_from_ecpkparameters()`/
4656 `EC_GROUP_new_from_ecparameters()`.
4657 This prevents bypass of security hardening and performance gains,
4658 especially for curves with specialized EC_METHODs.
4659 By default, if a key encoded with explicit parameters is loaded and later
4660 encoded, the output is still encoded with explicit parameters, even if
4661 internally a "named" EC_GROUP is used for computation.
4665 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4666 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4667 NULL. After this change, only the cofactor parameter can be NULL. It also
4668 does some minimal sanity checks on the passed order.
4673 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4674 An attack is simple, if the first CMS_recipientInfo is valid but the
4675 second CMS_recipientInfo is chosen ciphertext. If the second
4676 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4677 encryption key will be replaced by garbage, and the message cannot be
4678 decoded, but if the RSA decryption fails, the correct encryption key is
4679 used and the recipient will not notice the attack.
4680 As a work around for this potential attack the length of the decrypted
4681 key must be equal to the cipher default key length, in case the
4682 certifiate is not given and all recipientInfo are tried out.
4683 The old behaviour can be re-enabled in the CMS code by setting the
4684 CMS_DEBUG_DECRYPT flag.
4689 * Document issue with installation paths in diverse Windows builds
4691 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4692 binaries and run-time config file.
4697 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4699 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4700 This changes the size when using the `genpkey` command when no size is given.
4701 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4702 generation commands to use 2048 bits by default.
4706 * Add FIPS support for Android Arm 64-bit
4708 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4709 Module in Version 2.0.10. For some reason, the corresponding target
4710 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4711 built with FIPS support on Android Arm 64-bit. This omission has been
4714 *Matthias St. Pierre*
4716 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4718 * 0-byte record padding oracle
4720 If an application encounters a fatal protocol error and then calls
4721 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4722 then OpenSSL can respond differently to the calling application if a 0 byte
4723 record is received with invalid padding compared to if a 0 byte record is
4724 received with an invalid MAC. If the application then behaves differently
4725 based on that in a way that is detectable to the remote peer, then this
4726 amounts to a padding oracle that could be used to decrypt data.
4728 In order for this to be exploitable "non-stitched" ciphersuites must be in
4729 use. Stitched ciphersuites are optimised implementations of certain
4730 commonly used ciphersuites. Also the application must call SSL_shutdown()
4731 twice even if a protocol error has occurred (applications should not do
4732 this but some do anyway).
4734 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4735 Aviram, with additional investigation by Steven Collison and Andrew
4736 Hourselt. It was reported to OpenSSL on 10th December 2018.
4741 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4745 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4747 * Microarchitecture timing vulnerability in ECC scalar multiplication
4749 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4750 shown to be vulnerable to a microarchitecture timing side channel attack.
4751 An attacker with sufficient access to mount local timing attacks during
4752 ECDSA signature generation could recover the private key.
4754 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4755 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4761 * Timing vulnerability in DSA signature generation
4763 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4764 timing side channel attack. An attacker could use variations in the signing
4765 algorithm to recover the private key.
4767 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4772 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4773 Module, accidentally introduced while backporting security fixes from the
4774 development branch and hindering the use of ECC in FIPS mode.
4778 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4780 * Client DoS due to large DH parameter
4782 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4783 malicious server can send a very large prime value to the client. This will
4784 cause the client to spend an unreasonably long period of time generating a
4785 key for this prime resulting in a hang until the client has finished. This
4786 could be exploited in a Denial Of Service attack.
4788 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4793 * Cache timing vulnerability in RSA Key Generation
4795 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4796 a cache timing side channel attack. An attacker with sufficient access to
4797 mount cache timing attacks during the RSA key generation process could
4798 recover the private key.
4800 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4801 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4806 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4807 parameter is no longer accepted, as it leads to a corrupt table. NULL
4808 pem_str is reserved for alias entries only.
4812 * Revert blinding in ECDSA sign and instead make problematic addition
4813 length-invariant. Switch even to fixed-length Montgomery multiplication.
4817 * Change generating and checking of primes so that the error rate of not
4818 being prime depends on the intended use based on the size of the input.
4819 For larger primes this will result in more rounds of Miller-Rabin.
4820 The maximal error rate for primes with more than 1080 bits is lowered
4823 *Kurt Roeckx, Annie Yousar*
4825 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4829 * Add blinding to ECDSA and DSA signatures to protect against side channel
4830 attacks discovered by Keegan Ryan (NCC Group).
4834 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4835 now allow empty (zero character) pass phrases.
4839 * Certificate time validation (X509_cmp_time) enforces stricter
4840 compliance with RFC 5280. Fractional seconds and timezone offsets
4841 are no longer allowed.
4845 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4847 * Constructed ASN.1 types with a recursive definition could exceed the stack
4849 Constructed ASN.1 types with a recursive definition (such as can be found
4850 in PKCS7) could eventually exceed the stack given malicious input with
4851 excessive recursion. This could result in a Denial Of Service attack. There
4852 are no such structures used within SSL/TLS that come from untrusted sources
4853 so this is considered safe.
4855 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4861 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4863 * Read/write after SSL object in error state
4865 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4866 mechanism. The intent was that if a fatal error occurred during a handshake
4867 then OpenSSL would move into the error state and would immediately fail if
4868 you attempted to continue the handshake. This works as designed for the
4869 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4870 SSL_connect()), however due to a bug it does not work correctly if
4871 SSL_read() or SSL_write() is called directly. In that scenario, if the
4872 handshake fails then a fatal error will be returned in the initial function
4873 call. If SSL_read()/SSL_write() is subsequently called by the application
4874 for the same SSL object then it will succeed and the data is passed without
4875 being decrypted/encrypted directly from the SSL/TLS record layer.
4877 In order to exploit this issue an application bug would have to be present
4878 that resulted in a call to SSL_read()/SSL_write() being issued after having
4879 already received a fatal error.
4881 This issue was reported to OpenSSL by David Benjamin (Google).
4886 * rsaz_1024_mul_avx2 overflow bug on x86_64
4888 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4889 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4890 Analysis suggests that attacks against RSA and DSA as a result of this
4891 defect would be very difficult to perform and are not believed likely.
4892 Attacks against DH1024 are considered just feasible, because most of the
4893 work necessary to deduce information about a private key may be performed
4894 offline. The amount of resources required for such an attack would be
4895 significant. However, for an attack on TLS to be meaningful, the server
4896 would have to share the DH1024 private key among multiple clients, which is
4897 no longer an option since CVE-2016-0701.
4899 This only affects processors that support the AVX2 but not ADX extensions
4900 like Intel Haswell (4th generation).
4902 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4903 was originally found via the OSS-Fuzz project.
4908 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4910 * bn_sqrx8x_internal carry bug on x86_64
4912 There is a carry propagating bug in the x86_64 Montgomery squaring
4913 procedure. No EC algorithms are affected. Analysis suggests that attacks
4914 against RSA and DSA as a result of this defect would be very difficult to
4915 perform and are not believed likely. Attacks against DH are considered just
4916 feasible (although very difficult) because most of the work necessary to
4917 deduce information about a private key may be performed offline. The amount
4918 of resources required for such an attack would be very significant and
4919 likely only accessible to a limited number of attackers. An attacker would
4920 additionally need online access to an unpatched system using the target
4921 private key in a scenario with persistent DH parameters and a private
4922 key that is shared between multiple clients.
4924 This only affects processors that support the BMI1, BMI2 and ADX extensions
4925 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4927 This issue was reported to OpenSSL by the OSS-Fuzz project.
4932 * Malformed X.509 IPAddressFamily could cause OOB read
4934 If an X.509 certificate has a malformed IPAddressFamily extension,
4935 OpenSSL could do a one-byte buffer overread. The most likely result
4936 would be an erroneous display of the certificate in text format.
4938 This issue was reported to OpenSSL by the OSS-Fuzz project.
4942 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4944 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4945 platform rather than 'mingw'.
4949 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4951 * Truncated packet could crash via OOB read
4953 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4954 cipher is being used, then a truncated packet can cause that host to
4955 perform an out-of-bounds read, usually resulting in a crash.
4957 This issue was reported to OpenSSL by Robert Święcki of Google.
4962 * BN_mod_exp may produce incorrect results on x86_64
4964 There is a carry propagating bug in the x86_64 Montgomery squaring
4965 procedure. No EC algorithms are affected. Analysis suggests that attacks
4966 against RSA and DSA as a result of this defect would be very difficult to
4967 perform and are not believed likely. Attacks against DH are considered just
4968 feasible (although very difficult) because most of the work necessary to
4969 deduce information about a private key may be performed offline. The amount
4970 of resources required for such an attack would be very significant and
4971 likely only accessible to a limited number of attackers. An attacker would
4972 additionally need online access to an unpatched system using the target
4973 private key in a scenario with persistent DH parameters and a private
4974 key that is shared between multiple clients. For example this can occur by
4975 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4976 similar to CVE-2015-3193 but must be treated as a separate problem.
4978 This issue was reported to OpenSSL by the OSS-Fuzz project.
4983 * Montgomery multiplication may produce incorrect results
4985 There is a carry propagating bug in the Broadwell-specific Montgomery
4986 multiplication procedure that handles input lengths divisible by, but
4987 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4988 and DH private keys are impossible. This is because the subroutine in
4989 question is not used in operations with the private key itself and an input
4990 of the attacker's direct choice. Otherwise the bug can manifest itself as
4991 transient authentication and key negotiation failures or reproducible
4992 erroneous outcome of public-key operations with specially crafted input.
4993 Among EC algorithms only Brainpool P-512 curves are affected and one
4994 presumably can attack ECDH key negotiation. Impact was not analyzed in
4995 detail, because pre-requisites for attack are considered unlikely. Namely
4996 multiple clients have to choose the curve in question and the server has to
4997 share the private key among them, neither of which is default behaviour.
4998 Even then only clients that chose the curve will be affected.
5000 This issue was publicly reported as transient failures and was not
5001 initially recognized as a security issue. Thanks to Richard Morgan for
5002 providing reproducible case.
5007 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5008 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5009 prevent issues where no progress is being made and the peer continually
5010 sends unrecognised record types, using up resources processing them.
5014 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5016 * Missing CRL sanity check
5018 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5019 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5020 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5022 This issue only affects the OpenSSL 1.0.2i
5027 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5029 * OCSP Status Request extension unbounded memory growth
5031 A malicious client can send an excessively large OCSP Status Request
5032 extension. If that client continually requests renegotiation, sending a
5033 large OCSP Status Request extension each time, then there will be unbounded
5034 memory growth on the server. This will eventually lead to a Denial Of
5035 Service attack through memory exhaustion. Servers with a default
5036 configuration are vulnerable even if they do not support OCSP. Builds using
5037 the "no-ocsp" build time option are not affected.
5039 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5044 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5047 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5053 * OOB write in MDC2_Update()
5055 An overflow can occur in MDC2_Update() either if called directly or
5056 through the EVP_DigestUpdate() function using MDC2. If an attacker
5057 is able to supply very large amounts of input data after a previous
5058 call to EVP_EncryptUpdate() with a partial block then a length check
5059 can overflow resulting in a heap corruption.
5061 The amount of data needed is comparable to SIZE_MAX which is impractical
5064 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5069 * Malformed SHA512 ticket DoS
5071 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5072 DoS attack where a malformed ticket will result in an OOB read which will
5075 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5076 a custom server callback and ticket lookup mechanism.
5078 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5083 * OOB write in BN_bn2dec()
5085 The function BN_bn2dec() does not check the return value of BN_div_word().
5086 This can cause an OOB write if an application uses this function with an
5087 overly large BIGNUM. This could be a problem if an overly large certificate
5088 or CRL is printed out from an untrusted source. TLS is not affected because
5089 record limits will reject an oversized certificate before it is parsed.
5091 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5096 * OOB read in TS_OBJ_print_bio()
5098 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5099 the total length the OID text representation would use and not the amount
5100 of data written. This will result in OOB reads when large OIDs are
5103 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5108 * Pointer arithmetic undefined behaviour
5110 Avoid some undefined pointer arithmetic
5112 A common idiom in the codebase is to check limits in the following manner:
5115 Where "p" points to some malloc'd data of SIZE bytes and
5118 "len" here could be from some externally supplied data (e.g. from a TLS
5121 The rules of C pointer arithmetic are such that "p + len" is only well
5122 defined where len <= SIZE. Therefore the above idiom is actually
5123 undefined behaviour.
5125 For example this could cause problems if some malloc implementation
5126 provides an address for "p" such that "p + len" actually overflows for
5127 values of len that are too big and therefore p + len < limit.
5129 This issue was reported to OpenSSL by Guido Vranken
5134 * Constant time flag not preserved in DSA signing
5136 Operations in the DSA signing algorithm should run in constant time in
5137 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5138 implementation means that a non-constant time codepath is followed for
5139 certain operations. This has been demonstrated through a cache-timing
5140 attack to be sufficient for an attacker to recover the private DSA key.
5142 This issue was reported by César Pereida (Aalto University), Billy Brumley
5143 (Tampere University of Technology), and Yuval Yarom (The University of
5144 Adelaide and NICTA).
5149 * DTLS buffered message DoS
5151 In a DTLS connection where handshake messages are delivered out-of-order
5152 those messages that OpenSSL is not yet ready to process will be buffered
5153 for later use. Under certain circumstances, a flaw in the logic means that
5154 those messages do not get removed from the buffer even though the handshake
5155 has been completed. An attacker could force up to approx. 15 messages to
5156 remain in the buffer when they are no longer required. These messages will
5157 be cleared when the DTLS connection is closed. The default maximum size for
5158 a message is 100k. Therefore the attacker could force an additional 1500k
5159 to be consumed per connection. By opening many simulataneous connections an
5160 attacker could cause a DoS attack through memory exhaustion.
5162 This issue was reported to OpenSSL by Quan Luo.
5167 * DTLS replay protection DoS
5169 A flaw in the DTLS replay attack protection mechanism means that records
5170 that arrive for future epochs update the replay protection "window" before
5171 the MAC for the record has been validated. This could be exploited by an
5172 attacker by sending a record for the next epoch (which does not have to
5173 decrypt or have a valid MAC), with a very large sequence number. This means
5174 that all subsequent legitimate packets are dropped causing a denial of
5175 service for a specific DTLS connection.
5177 This issue was reported to OpenSSL by the OCAP audit team.
5182 * Certificate message OOB reads
5184 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5185 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5186 theoretical DoS risk but this has not been observed in practice on common
5189 The messages affected are client certificate, client certificate request
5190 and server certificate. As a result the attack can only be performed
5191 against a client or a server which enables client authentication.
5193 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5198 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5200 * Prevent padding oracle in AES-NI CBC MAC check
5202 A MITM attacker can use a padding oracle attack to decrypt traffic
5203 when the connection uses an AES CBC cipher and the server support
5206 This issue was introduced as part of the fix for Lucky 13 padding
5207 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5208 constant time by making sure that always the same bytes are read and
5209 compared against either the MAC or padding bytes. But it no longer
5210 checked that there was enough data to have both the MAC and padding
5213 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5217 * Fix EVP_EncodeUpdate overflow
5219 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5220 Base64 encoding of binary data. If an attacker is able to supply very large
5221 amounts of input data then a length check can overflow resulting in a heap
5224 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5225 the `PEM_write_bio*` family of functions. These are mainly used within the
5226 OpenSSL command line applications, so any application which processes data
5227 from an untrusted source and outputs it as a PEM file should be considered
5228 vulnerable to this issue. User applications that call these APIs directly
5229 with large amounts of untrusted data may also be vulnerable.
5231 This issue was reported by Guido Vranken.
5236 * Fix EVP_EncryptUpdate overflow
5238 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5239 is able to supply very large amounts of input data after a previous call to
5240 EVP_EncryptUpdate() with a partial block then a length check can overflow
5241 resulting in a heap corruption. Following an analysis of all OpenSSL
5242 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5243 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5244 the first called function after an EVP_EncryptInit(), and therefore that
5245 specific call must be safe. The second form is where the length passed to
5246 EVP_EncryptUpdate() can be seen from the code to be some small value and
5247 therefore there is no possibility of an overflow. Since all instances are
5248 one of these two forms, it is believed that there can be no overflows in
5249 internal code due to this problem. It should be noted that
5250 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5251 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5252 of these calls have also been analysed too and it is believed there are no
5253 instances in internal usage where an overflow could occur.
5255 This issue was reported by Guido Vranken.
5260 * Prevent ASN.1 BIO excessive memory allocation
5262 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5263 a short invalid encoding can cause allocation of large amounts of memory
5264 potentially consuming excessive resources or exhausting memory.
5266 Any application parsing untrusted data through d2i BIO functions is
5267 affected. The memory based functions such as d2i_X509() are *not* affected.
5268 Since the memory based functions are used by the TLS library, TLS
5269 applications are not affected.
5271 This issue was reported by Brian Carpenter.
5278 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5279 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5280 in arbitrary stack data being returned in the buffer.
5282 This issue was reported by Guido Vranken.
5287 * Modify behavior of ALPN to invoke callback after SNI/servername
5288 callback, such that updates to the SSL_CTX affect ALPN.
5292 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5297 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5298 methods are enabled and ssl2 is disabled the methods return NULL.
5302 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5304 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5305 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5306 provide any "EXPORT" or "LOW" strength ciphers.
5310 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5311 is by default disabled at build-time. Builds that are not configured with
5312 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5313 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5314 will need to explicitly call either of:
5316 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5318 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5320 as appropriate. Even if either of those is used, or the application
5321 explicitly uses the version-specific SSLv2_method() or its client and
5322 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5323 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5324 ciphers, and SSLv2 56-bit DES are no longer available.
5329 * Fix a double-free in DSA code
5331 A double free bug was discovered when OpenSSL parses malformed DSA private
5332 keys and could lead to a DoS attack or memory corruption for applications
5333 that receive DSA private keys from untrusted sources. This scenario is
5336 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5342 * Disable SRP fake user seed to address a server memory leak.
5344 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5346 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5347 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5348 was changed to ignore the "fake user" SRP seed, even if the seed
5351 Users should use SRP_VBASE_get1_by_user instead. Note that in
5352 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5353 also that even though configuring the SRP seed attempts to hide
5354 invalid usernames by continuing the handshake with fake
5355 credentials, this behaviour is not constant time and no strong
5356 guarantees are made that the handshake is indistinguishable from
5357 that of a valid user.
5362 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5364 In the BN_hex2bn function the number of hex digits is calculated using an
5365 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5366 large values of `i` this can result in `bn_expand` not allocating any
5367 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5368 field as NULL leading to a subsequent NULL ptr deref. For very large values
5369 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5370 In this case memory is allocated to the internal BIGNUM data field, but it
5371 is insufficiently sized leading to heap corruption. A similar issue exists
5372 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5373 is ever called by user applications with very large untrusted hex/dec data.
5374 This is anticipated to be a rare occurrence.
5376 All OpenSSL internal usage of these functions use data that is not expected
5377 to be untrusted, e.g. config file data or application command line
5378 arguments. If user developed applications generate config file data based
5379 on untrusted data then it is possible that this could also lead to security
5380 consequences. This is also anticipated to be rare.
5382 This issue was reported to OpenSSL by Guido Vranken.
5387 * Fix memory issues in `BIO_*printf` functions
5389 The internal `fmtstr` function used in processing a "%s" format string in
5390 the `BIO_*printf` functions could overflow while calculating the length of a
5391 string and cause an OOB read when printing very long strings.
5393 Additionally the internal `doapr_outch` function can attempt to write to an
5394 OOB memory location (at an offset from the NULL pointer) in the event of a
5395 memory allocation failure. In 1.0.2 and below this could be caused where
5396 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5397 could be in processing a very long "%s" format string. Memory leaks can
5400 The first issue may mask the second issue dependent on compiler behaviour.
5401 These problems could enable attacks where large amounts of untrusted data
5402 is passed to the `BIO_*printf` functions. If applications use these functions
5403 in this way then they could be vulnerable. OpenSSL itself uses these
5404 functions when printing out human-readable dumps of ASN.1 data. Therefore
5405 applications that print this data could be vulnerable if the data is from
5406 untrusted sources. OpenSSL command line applications could also be
5407 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5408 as command line arguments.
5410 Libssl is not considered directly vulnerable. Additionally certificates etc
5411 received via remote connections via libssl are also unlikely to be able to
5412 trigger these issues because of message size limits enforced within libssl.
5414 This issue was reported to OpenSSL Guido Vranken.
5419 * Side channel attack on modular exponentiation
5421 A side-channel attack was found which makes use of cache-bank conflicts on
5422 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5423 of RSA keys. The ability to exploit this issue is limited as it relies on
5424 an attacker who has control of code in a thread running on the same
5425 hyper-threaded core as the victim thread which is performing decryptions.
5427 This issue was reported to OpenSSL by Yuval Yarom, The University of
5428 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5429 Nadia Heninger, University of Pennsylvania with more information at
5430 <http://cachebleed.info>.
5435 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5436 if no keysize is specified with default_bits. This fixes an
5437 omission in an earlier change that changed all RSA/DSA key generation
5438 commands to use 2048 bits by default.
5442 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5444 * DH small subgroups
5446 Historically OpenSSL only ever generated DH parameters based on "safe"
5447 primes. More recently (in version 1.0.2) support was provided for
5448 generating X9.42 style parameter files such as those required for RFC 5114
5449 support. The primes used in such files may not be "safe". Where an
5450 application is using DH configured with parameters based on primes that are
5451 not "safe" then an attacker could use this fact to find a peer's private
5452 DH exponent. This attack requires that the attacker complete multiple
5453 handshakes in which the peer uses the same private DH exponent. For example
5454 this could be used to discover a TLS server's private DH exponent if it's
5455 reusing the private DH exponent or it's using a static DH ciphersuite.
5457 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5458 TLS. It is not on by default. If the option is not set then the server
5459 reuses the same private DH exponent for the life of the server process and
5460 would be vulnerable to this attack. It is believed that many popular
5461 applications do set this option and would therefore not be at risk.
5463 The fix for this issue adds an additional check where a "q" parameter is
5464 available (as is the case in X9.42 based parameters). This detects the
5465 only known attack, and is the only possible defense for static DH
5466 ciphersuites. This could have some performance impact.
5468 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5469 default and cannot be disabled. This could have some performance impact.
5471 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5476 * SSLv2 doesn't block disabled ciphers
5478 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5479 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5480 been disabled, provided that the SSLv2 protocol was not also disabled via
5483 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5484 and Sebastian Schinzel.
5489 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5491 * BN_mod_exp may produce incorrect results on x86_64
5493 There is a carry propagating bug in the x86_64 Montgomery squaring
5494 procedure. No EC algorithms are affected. Analysis suggests that attacks
5495 against RSA and DSA as a result of this defect would be very difficult to
5496 perform and are not believed likely. Attacks against DH are considered just
5497 feasible (although very difficult) because most of the work necessary to
5498 deduce information about a private key may be performed offline. The amount
5499 of resources required for such an attack would be very significant and
5500 likely only accessible to a limited number of attackers. An attacker would
5501 additionally need online access to an unpatched system using the target
5502 private key in a scenario with persistent DH parameters and a private
5503 key that is shared between multiple clients. For example this can occur by
5504 default in OpenSSL DHE based SSL/TLS ciphersuites.
5506 This issue was reported to OpenSSL by Hanno Böck.
5511 * Certificate verify crash with missing PSS parameter
5513 The signature verification routines will crash with a NULL pointer
5514 dereference if presented with an ASN.1 signature using the RSA PSS
5515 algorithm and absent mask generation function parameter. Since these
5516 routines are used to verify certificate signature algorithms this can be
5517 used to crash any certificate verification operation and exploited in a
5518 DoS attack. Any application which performs certificate verification is
5519 vulnerable including OpenSSL clients and servers which enable client
5522 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5527 * X509_ATTRIBUTE memory leak
5529 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5530 memory. This structure is used by the PKCS#7 and CMS routines so any
5531 application which reads PKCS#7 or CMS data from untrusted sources is
5532 affected. SSL/TLS is not affected.
5534 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5540 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5541 This changes the decoding behaviour for some invalid messages,
5542 though the change is mostly in the more lenient direction, and
5543 legacy behaviour is preserved as much as possible.
5547 * In DSA_generate_parameters_ex, if the provided seed is too short,
5550 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5552 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5554 * Alternate chains certificate forgery
5556 During certificate verification, OpenSSL will attempt to find an
5557 alternative certificate chain if the first attempt to build such a chain
5558 fails. An error in the implementation of this logic can mean that an
5559 attacker could cause certain checks on untrusted certificates to be
5560 bypassed, such as the CA flag, enabling them to use a valid leaf
5561 certificate to act as a CA and "issue" an invalid certificate.
5563 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5568 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5570 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5571 incompatibility in the handling of HMAC. The previous ABI has now been
5576 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5578 * Malformed ECParameters causes infinite loop
5580 When processing an ECParameters structure OpenSSL enters an infinite loop
5581 if the curve specified is over a specially malformed binary polynomial
5584 This can be used to perform denial of service against any
5585 system which processes public keys, certificate requests or
5586 certificates. This includes TLS clients and TLS servers with
5587 client authentication enabled.
5589 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5594 * Exploitable out-of-bounds read in X509_cmp_time
5596 X509_cmp_time does not properly check the length of the ASN1_TIME
5597 string and can read a few bytes out of bounds. In addition,
5598 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5601 An attacker can use this to craft malformed certificates and CRLs of
5602 various sizes and potentially cause a segmentation fault, resulting in
5603 a DoS on applications that verify certificates or CRLs. TLS clients
5604 that verify CRLs are affected. TLS clients and servers with client
5605 authentication enabled may be affected if they use custom verification
5608 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5609 independently by Hanno Böck.
5614 * PKCS7 crash with missing EnvelopedContent
5616 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5617 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5618 with missing content and trigger a NULL pointer dereference on parsing.
5620 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5621 structures from untrusted sources are affected. OpenSSL clients and
5622 servers are not affected.
5624 This issue was reported to OpenSSL by Michal Zalewski (Google).
5629 * CMS verify infinite loop with unknown hash function
5631 When verifying a signedData message the CMS code can enter an infinite loop
5632 if presented with an unknown hash function OID. This can be used to perform
5633 denial of service against any system which verifies signedData messages using
5635 This issue was reported to OpenSSL by Johannes Bauer.
5640 * Race condition handling NewSessionTicket
5642 If a NewSessionTicket is received by a multi-threaded client when attempting to
5643 reuse a previous ticket then a race condition can occur potentially leading to
5644 a double free of the ticket data.
5649 * Only support 256-bit or stronger elliptic curves with the
5650 'ecdh_auto' setting (server) or by default (client). Of supported
5651 curves, prefer P-256 (both).
5655 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5657 * ClientHello sigalgs DoS fix
5659 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5660 invalid signature algorithms extension a NULL pointer dereference will
5661 occur. This can be exploited in a DoS attack against the server.
5663 This issue was was reported to OpenSSL by David Ramos of Stanford
5667 *Stephen Henson and Matt Caswell*
5669 * Multiblock corrupted pointer fix
5671 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5672 feature only applies on 64 bit x86 architecture platforms that support AES
5673 NI instructions. A defect in the implementation of "multiblock" can cause
5674 OpenSSL's internal write buffer to become incorrectly set to NULL when
5675 using non-blocking IO. Typically, when the user application is using a
5676 socket BIO for writing, this will only result in a failed connection.
5677 However if some other BIO is used then it is likely that a segmentation
5678 fault will be triggered, thus enabling a potential DoS attack.
5680 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5685 * Segmentation fault in DTLSv1_listen fix
5687 The DTLSv1_listen function is intended to be stateless and processes the
5688 initial ClientHello from many peers. It is common for user code to loop
5689 over the call to DTLSv1_listen until a valid ClientHello is received with
5690 an associated cookie. A defect in the implementation of DTLSv1_listen means
5691 that state is preserved in the SSL object from one invocation to the next
5692 that can lead to a segmentation fault. Errors processing the initial
5693 ClientHello can trigger this scenario. An example of such an error could be
5694 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5697 This issue was reported to OpenSSL by Per Allansson.
5702 * Segmentation fault in ASN1_TYPE_cmp fix
5704 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5705 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5706 certificate signature algorithm consistency this can be used to crash any
5707 certificate verification operation and exploited in a DoS attack. Any
5708 application which performs certificate verification is vulnerable including
5709 OpenSSL clients and servers which enable client authentication.
5714 * Segmentation fault for invalid PSS parameters fix
5716 The signature verification routines will crash with a NULL pointer
5717 dereference if presented with an ASN.1 signature using the RSA PSS
5718 algorithm and invalid parameters. Since these routines are used to verify
5719 certificate signature algorithms this can be used to crash any
5720 certificate verification operation and exploited in a DoS attack. Any
5721 application which performs certificate verification is vulnerable including
5722 OpenSSL clients and servers which enable client authentication.
5724 This issue was was reported to OpenSSL by Brian Carpenter.
5729 * ASN.1 structure reuse memory corruption fix
5731 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5732 memory corruption via an invalid write. Such reuse is and has been
5733 strongly discouraged and is believed to be rare.
5735 Applications that parse structures containing CHOICE or ANY DEFINED BY
5736 components may be affected. Certificate parsing (d2i_X509 and related
5737 functions) are however not affected. OpenSSL clients and servers are
5743 * PKCS7 NULL pointer dereferences fix
5745 The PKCS#7 parsing code does not handle missing outer ContentInfo
5746 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5747 missing content and trigger a NULL pointer dereference on parsing.
5749 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5750 otherwise parse PKCS#7 structures from untrusted sources are
5751 affected. OpenSSL clients and servers are not affected.
5753 This issue was reported to OpenSSL by Michal Zalewski (Google).
5758 * DoS via reachable assert in SSLv2 servers fix
5760 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5761 servers that both support SSLv2 and enable export cipher suites by sending
5762 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5764 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5765 (OpenSSL development team).
5770 * Empty CKE with client auth and DHE fix
5772 If client auth is used then a server can seg fault in the event of a DHE
5773 ciphersuite being selected and a zero length ClientKeyExchange message
5774 being sent by the client. This could be exploited in a DoS attack.
5779 * Handshake with unseeded PRNG fix
5781 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5782 with an unseeded PRNG. The conditions are:
5783 - The client is on a platform where the PRNG has not been seeded
5784 automatically, and the user has not seeded manually
5785 - A protocol specific client method version has been used (i.e. not
5786 SSL_client_methodv23)
5787 - A ciphersuite is used that does not require additional random data from
5788 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5790 If the handshake succeeds then the client random that has been used will
5791 have been generated from a PRNG with insufficient entropy and therefore the
5792 output may be predictable.
5794 For example using the following command with an unseeded openssl will
5795 succeed on an unpatched platform:
5797 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5802 * Use After Free following d2i_ECPrivatekey error fix
5804 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5805 could cause a use after free condition. This, in turn, could cause a double
5806 free in several private key parsing functions (such as d2i_PrivateKey
5807 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5808 for applications that receive EC private keys from untrusted
5809 sources. This scenario is considered rare.
5811 This issue was discovered by the BoringSSL project and fixed in their
5817 * X509_to_X509_REQ NULL pointer deref fix
5819 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5820 the certificate key is invalid. This function is rarely used in practice.
5822 This issue was discovered by Brian Carpenter.
5827 * Removed the export ciphers from the DEFAULT ciphers
5831 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5833 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5834 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5835 So far those who have to target multiple platforms would compromise
5836 and argue that binary targeting say ARMv5 would still execute on
5837 ARMv8. "Universal" build resolves this compromise by providing
5838 near-optimal performance even on newer platforms.
5842 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5843 (other platforms pending).
5845 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5847 * Add support for the SignedCertificateTimestampList certificate and
5848 OCSP response extensions from RFC6962.
5852 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5853 for corner cases. (Certain input points at infinity could lead to
5854 bogus results, with non-infinity inputs mapped to infinity too.)
5858 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5859 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5860 common cases are optimized and there still is room for further
5861 improvements. Vector Permutation AES for Altivec is also added.
5865 * Add support for little-endian ppc64 Linux target.
5867 *Marcelo Cerri (IBM)*
5869 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5870 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5871 are optimized and there still is room for further improvements.
5872 Both 32- and 64-bit modes are supported.
5874 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5876 * Improved ARMv7 NEON support.
5880 * Support for SPARC Architecture 2011 crypto extensions, first
5881 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5882 SHA256/512, MD5, GHASH and modular exponentiation.
5884 *Andy Polyakov, David Miller*
5886 * Accelerated modular exponentiation for Intel processors, a.k.a.
5889 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5891 * Support for new and upcoming Intel processors, including AVX2,
5892 BMI and SHA ISA extensions. This includes additional "stitched"
5893 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5896 This work was sponsored by Intel Corp.
5900 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5901 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5902 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5906 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5907 this fixes a limitation in previous versions of OpenSSL.
5911 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5912 MGF1 digest and OAEP label.
5916 * Add EVP support for key wrapping algorithms, to avoid problems with
5917 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5918 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5919 algorithms and include tests cases.
5923 * Add functions to allocate and set the fields of an ECDSA_METHOD
5926 *Douglas E. Engert, Steve Henson*
5928 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5929 difference in days and seconds between two tm or ASN1_TIME structures.
5933 * Add -rev test option to s_server to just reverse order of characters
5934 received by client and send back to server. Also prints an abbreviated
5935 summary of the connection parameters.
5939 * New option -brief for s_client and s_server to print out a brief summary
5940 of connection parameters.
5944 * Add callbacks for arbitrary TLS extensions.
5946 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5948 * New option -crl_download in several openssl utilities to download CRLs
5949 from CRLDP extension in certificates.
5953 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5957 * New function X509_CRL_diff to generate a delta CRL from the difference
5958 of two full CRLs. Add support to "crl" utility.
5962 * New functions to set lookup_crls function and to retrieve
5963 X509_STORE from X509_STORE_CTX.
5967 * Print out deprecated issuer and subject unique ID fields in
5972 * Extend OCSP I/O functions so they can be used for simple general purpose
5973 HTTP as well as OCSP. New wrapper function which can be used to download
5974 CRLs using the OCSP API.
5978 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5982 * `SSL_CONF*` functions. These provide a common framework for application
5983 configuration using configuration files or command lines.
5987 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5988 message callback and prints the results. Needs compile time option
5989 "enable-ssl-trace". New options to s_client and s_server to enable
5994 * New ctrl and macro to retrieve supported points extensions.
5995 Print out extension in s_server and s_client.
5999 * New functions to retrieve certificate signature and signature
6004 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6009 * New Suite B modes for TLS code. These use and enforce the requirements
6010 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6011 only use Suite B curves. The Suite B modes can be set by using the
6012 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6016 * New chain verification flags for Suite B levels of security. Check
6017 algorithms are acceptable when flags are set in X509_verify_cert.
6021 * Make tls1_check_chain return a set of flags indicating checks passed
6022 by a certificate chain. Add additional tests to handle client
6023 certificates: checks for matching certificate type and issuer name
6028 * If an attempt is made to use a signature algorithm not in the peer
6029 preference list abort the handshake. If client has no suitable
6030 signature algorithms in response to a certificate request do not
6031 use the certificate.
6035 * If server EC tmp key is not in client preference list abort handshake.
6039 * Add support for certificate stores in CERT structure. This makes it
6040 possible to have different stores per SSL structure or one store in
6041 the parent SSL_CTX. Include distinct stores for certificate chain
6042 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6043 to build and store a certificate chain in CERT structure: returning
6044 an error if the chain cannot be built: this will allow applications
6045 to test if a chain is correctly configured.
6047 Note: if the CERT based stores are not set then the parent SSL_CTX
6048 store is used to retain compatibility with existing behaviour.
6052 * New function ssl_set_client_disabled to set a ciphersuite disabled
6053 mask based on the current session, check mask when sending client
6054 hello and checking the requested ciphersuite.
6058 * New ctrls to retrieve and set certificate types in a certificate
6059 request message. Print out received values in s_client. If certificate
6060 types is not set with custom values set sensible values based on
6061 supported signature algorithms.
6065 * Support for distinct client and server supported signature algorithms.
6069 * Add certificate callback. If set this is called whenever a certificate
6070 is required by client or server. An application can decide which
6071 certificate chain to present based on arbitrary criteria: for example
6072 supported signature algorithms. Add very simple example to s_server.
6073 This fixes many of the problems and restrictions of the existing client
6074 certificate callback: for example you can now clear an existing
6075 certificate and specify the whole chain.
6079 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6080 the certificate can be used for (if anything). Set valid_flags field
6081 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6082 to have similar checks in it.
6084 Add new "cert_flags" field to CERT structure and include a "strict mode".
6085 This enforces some TLS certificate requirements (such as only permitting
6086 certificate signature algorithms contained in the supported algorithms
6087 extension) which some implementations ignore: this option should be used
6088 with caution as it could cause interoperability issues.
6092 * Update and tidy signature algorithm extension processing. Work out
6093 shared signature algorithms based on preferences and peer algorithms
6094 and print them out in s_client and s_server. Abort handshake if no
6095 shared signature algorithms.
6099 * Add new functions to allow customised supported signature algorithms
6100 for SSL and SSL_CTX structures. Add options to s_client and s_server
6105 * New function SSL_certs_clear() to delete all references to certificates
6106 from an SSL structure. Before this once a certificate had been added
6107 it couldn't be removed.
6111 * Integrate hostname, email address and IP address checking with certificate
6112 verification. New verify options supporting checking in openssl utility.
6116 * Fixes and wildcard matching support to hostname and email checking
6117 functions. Add manual page.
6119 *Florian Weimer (Red Hat Product Security Team)*
6121 * New functions to check a hostname email or IP address against a
6122 certificate. Add options x509 utility to print results of checks against
6127 * Fix OCSP checking.
6129 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6131 * Initial experimental support for explicitly trusted non-root CAs.
6132 OpenSSL still tries to build a complete chain to a root but if an
6133 intermediate CA has a trust setting included that is used. The first
6134 setting is used: whether to trust (e.g., -addtrust option to the x509
6139 * Add -trusted_first option which attempts to find certificates in the
6140 trusted store even if an untrusted chain is also supplied.
6144 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6145 platform support for Linux and Android.
6149 * Support for linux-x32, ILP32 environment in x86_64 framework.
6153 * Experimental multi-implementation support for FIPS capable OpenSSL.
6154 When in FIPS mode the approved implementations are used as normal,
6155 when not in FIPS mode the internal unapproved versions are used instead.
6156 This means that the FIPS capable OpenSSL isn't forced to use the
6157 (often lower performance) FIPS implementations outside FIPS mode.
6161 * Transparently support X9.42 DH parameters when calling
6162 PEM_read_bio_DHparameters. This means existing applications can handle
6163 the new parameter format automatically.
6167 * Initial experimental support for X9.42 DH parameter format: mainly
6168 to support use of 'q' parameter for RFC5114 parameters.
6172 * Add DH parameters from RFC5114 including test data to dhtest.
6176 * Support for automatic EC temporary key parameter selection. If enabled
6177 the most preferred EC parameters are automatically used instead of
6178 hardcoded fixed parameters. Now a server just has to call:
6179 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6180 support ECDH and use the most appropriate parameters.
6184 * Enhance and tidy EC curve and point format TLS extension code. Use
6185 static structures instead of allocation if default values are used.
6186 New ctrls to set curves we wish to support and to retrieve shared curves.
6187 Print out shared curves in s_server. New options to s_server and s_client
6188 to set list of supported curves.
6192 * New ctrls to retrieve supported signature algorithms and
6193 supported curve values as an array of NIDs. Extend openssl utility
6194 to print out received values.
6198 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6199 between NIDs and the more common NIST names such as "P-256". Enhance
6200 ecparam utility and ECC method to recognise the NIST names for curves.
6204 * Enhance SSL/TLS certificate chain handling to support different
6205 chains for each certificate instead of one chain in the parent SSL_CTX.
6209 * Support for fixed DH ciphersuite client authentication: where both
6210 server and client use DH certificates with common parameters.
6214 * Support for fixed DH ciphersuites: those requiring DH server
6219 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6221 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6222 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6223 X509_CINF_get_signature were reverted post internal team review.
6228 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6230 * OCSP Status Request extension unbounded memory growth
6232 A malicious client can send an excessively large OCSP Status Request
6233 extension. If that client continually requests renegotiation, sending a
6234 large OCSP Status Request extension each time, then there will be unbounded
6235 memory growth on the server. This will eventually lead to a Denial Of
6236 Service attack through memory exhaustion. Servers with a default
6237 configuration are vulnerable even if they do not support OCSP. Builds using
6238 the "no-ocsp" build time option are not affected.
6240 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6245 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6248 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6254 * OOB write in MDC2_Update()
6256 An overflow can occur in MDC2_Update() either if called directly or
6257 through the EVP_DigestUpdate() function using MDC2. If an attacker
6258 is able to supply very large amounts of input data after a previous
6259 call to EVP_EncryptUpdate() with a partial block then a length check
6260 can overflow resulting in a heap corruption.
6262 The amount of data needed is comparable to SIZE_MAX which is impractical
6265 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6270 * Malformed SHA512 ticket DoS
6272 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6273 DoS attack where a malformed ticket will result in an OOB read which will
6276 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6277 a custom server callback and ticket lookup mechanism.
6279 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6284 * OOB write in BN_bn2dec()
6286 The function BN_bn2dec() does not check the return value of BN_div_word().
6287 This can cause an OOB write if an application uses this function with an
6288 overly large BIGNUM. This could be a problem if an overly large certificate
6289 or CRL is printed out from an untrusted source. TLS is not affected because
6290 record limits will reject an oversized certificate before it is parsed.
6292 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6297 * OOB read in TS_OBJ_print_bio()
6299 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6300 the total length the OID text representation would use and not the amount
6301 of data written. This will result in OOB reads when large OIDs are
6304 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6309 * Pointer arithmetic undefined behaviour
6311 Avoid some undefined pointer arithmetic
6313 A common idiom in the codebase is to check limits in the following manner:
6316 Where "p" points to some malloc'd data of SIZE bytes and
6319 "len" here could be from some externally supplied data (e.g. from a TLS
6322 The rules of C pointer arithmetic are such that "p + len" is only well
6323 defined where len <= SIZE. Therefore the above idiom is actually
6324 undefined behaviour.
6326 For example this could cause problems if some malloc implementation
6327 provides an address for "p" such that "p + len" actually overflows for
6328 values of len that are too big and therefore p + len < limit.
6330 This issue was reported to OpenSSL by Guido Vranken
6335 * Constant time flag not preserved in DSA signing
6337 Operations in the DSA signing algorithm should run in constant time in
6338 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6339 implementation means that a non-constant time codepath is followed for
6340 certain operations. This has been demonstrated through a cache-timing
6341 attack to be sufficient for an attacker to recover the private DSA key.
6343 This issue was reported by César Pereida (Aalto University), Billy Brumley
6344 (Tampere University of Technology), and Yuval Yarom (The University of
6345 Adelaide and NICTA).
6350 * DTLS buffered message DoS
6352 In a DTLS connection where handshake messages are delivered out-of-order
6353 those messages that OpenSSL is not yet ready to process will be buffered
6354 for later use. Under certain circumstances, a flaw in the logic means that
6355 those messages do not get removed from the buffer even though the handshake
6356 has been completed. An attacker could force up to approx. 15 messages to
6357 remain in the buffer when they are no longer required. These messages will
6358 be cleared when the DTLS connection is closed. The default maximum size for
6359 a message is 100k. Therefore the attacker could force an additional 1500k
6360 to be consumed per connection. By opening many simulataneous connections an
6361 attacker could cause a DoS attack through memory exhaustion.
6363 This issue was reported to OpenSSL by Quan Luo.
6368 * DTLS replay protection DoS
6370 A flaw in the DTLS replay attack protection mechanism means that records
6371 that arrive for future epochs update the replay protection "window" before
6372 the MAC for the record has been validated. This could be exploited by an
6373 attacker by sending a record for the next epoch (which does not have to
6374 decrypt or have a valid MAC), with a very large sequence number. This means
6375 that all subsequent legitimate packets are dropped causing a denial of
6376 service for a specific DTLS connection.
6378 This issue was reported to OpenSSL by the OCAP audit team.
6383 * Certificate message OOB reads
6385 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6386 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6387 theoretical DoS risk but this has not been observed in practice on common
6390 The messages affected are client certificate, client certificate request
6391 and server certificate. As a result the attack can only be performed
6392 against a client or a server which enables client authentication.
6394 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6399 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6401 * Prevent padding oracle in AES-NI CBC MAC check
6403 A MITM attacker can use a padding oracle attack to decrypt traffic
6404 when the connection uses an AES CBC cipher and the server support
6407 This issue was introduced as part of the fix for Lucky 13 padding
6408 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6409 constant time by making sure that always the same bytes are read and
6410 compared against either the MAC or padding bytes. But it no longer
6411 checked that there was enough data to have both the MAC and padding
6414 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6419 * Fix EVP_EncodeUpdate overflow
6421 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6422 Base64 encoding of binary data. If an attacker is able to supply very large
6423 amounts of input data then a length check can overflow resulting in a heap
6426 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6427 the `PEM_write_bio*` family of functions. These are mainly used within the
6428 OpenSSL command line applications, so any application which processes data
6429 from an untrusted source and outputs it as a PEM file should be considered
6430 vulnerable to this issue. User applications that call these APIs directly
6431 with large amounts of untrusted data may also be vulnerable.
6433 This issue was reported by Guido Vranken.
6438 * Fix EVP_EncryptUpdate overflow
6440 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6441 is able to supply very large amounts of input data after a previous call to
6442 EVP_EncryptUpdate() with a partial block then a length check can overflow
6443 resulting in a heap corruption. Following an analysis of all OpenSSL
6444 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6445 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6446 the first called function after an EVP_EncryptInit(), and therefore that
6447 specific call must be safe. The second form is where the length passed to
6448 EVP_EncryptUpdate() can be seen from the code to be some small value and
6449 therefore there is no possibility of an overflow. Since all instances are
6450 one of these two forms, it is believed that there can be no overflows in
6451 internal code due to this problem. It should be noted that
6452 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6453 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6454 of these calls have also been analysed too and it is believed there are no
6455 instances in internal usage where an overflow could occur.
6457 This issue was reported by Guido Vranken.
6462 * Prevent ASN.1 BIO excessive memory allocation
6464 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6465 a short invalid encoding can casuse allocation of large amounts of memory
6466 potentially consuming excessive resources or exhausting memory.
6468 Any application parsing untrusted data through d2i BIO functions is
6469 affected. The memory based functions such as d2i_X509() are *not* affected.
6470 Since the memory based functions are used by the TLS library, TLS
6471 applications are not affected.
6473 This issue was reported by Brian Carpenter.
6480 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6481 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6482 in arbitrary stack data being returned in the buffer.
6484 This issue was reported by Guido Vranken.
6489 * Modify behavior of ALPN to invoke callback after SNI/servername
6490 callback, such that updates to the SSL_CTX affect ALPN.
6494 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6499 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6500 methods are enabled and ssl2 is disabled the methods return NULL.
6504 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6506 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6507 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6508 provide any "EXPORT" or "LOW" strength ciphers.
6512 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6513 is by default disabled at build-time. Builds that are not configured with
6514 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6515 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6516 will need to explicitly call either of:
6518 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6520 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6522 as appropriate. Even if either of those is used, or the application
6523 explicitly uses the version-specific SSLv2_method() or its client and
6524 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6525 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6526 ciphers, and SSLv2 56-bit DES are no longer available.
6531 * Fix a double-free in DSA code
6533 A double free bug was discovered when OpenSSL parses malformed DSA private
6534 keys and could lead to a DoS attack or memory corruption for applications
6535 that receive DSA private keys from untrusted sources. This scenario is
6538 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6544 * Disable SRP fake user seed to address a server memory leak.
6546 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6548 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6549 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6550 was changed to ignore the "fake user" SRP seed, even if the seed
6553 Users should use SRP_VBASE_get1_by_user instead. Note that in
6554 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6555 also that even though configuring the SRP seed attempts to hide
6556 invalid usernames by continuing the handshake with fake
6557 credentials, this behaviour is not constant time and no strong
6558 guarantees are made that the handshake is indistinguishable from
6559 that of a valid user.
6564 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6566 In the BN_hex2bn function the number of hex digits is calculated using an
6567 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6568 large values of `i` this can result in `bn_expand` not allocating any
6569 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6570 field as NULL leading to a subsequent NULL ptr deref. For very large values
6571 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6572 In this case memory is allocated to the internal BIGNUM data field, but it
6573 is insufficiently sized leading to heap corruption. A similar issue exists
6574 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6575 is ever called by user applications with very large untrusted hex/dec data.
6576 This is anticipated to be a rare occurrence.
6578 All OpenSSL internal usage of these functions use data that is not expected
6579 to be untrusted, e.g. config file data or application command line
6580 arguments. If user developed applications generate config file data based
6581 on untrusted data then it is possible that this could also lead to security
6582 consequences. This is also anticipated to be rare.
6584 This issue was reported to OpenSSL by Guido Vranken.
6589 * Fix memory issues in `BIO_*printf` functions
6591 The internal `fmtstr` function used in processing a "%s" format string in
6592 the `BIO_*printf` functions could overflow while calculating the length of a
6593 string and cause an OOB read when printing very long strings.
6595 Additionally the internal `doapr_outch` function can attempt to write to an
6596 OOB memory location (at an offset from the NULL pointer) in the event of a
6597 memory allocation failure. In 1.0.2 and below this could be caused where
6598 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6599 could be in processing a very long "%s" format string. Memory leaks can
6602 The first issue may mask the second issue dependent on compiler behaviour.
6603 These problems could enable attacks where large amounts of untrusted data
6604 is passed to the `BIO_*printf` functions. If applications use these functions
6605 in this way then they could be vulnerable. OpenSSL itself uses these
6606 functions when printing out human-readable dumps of ASN.1 data. Therefore
6607 applications that print this data could be vulnerable if the data is from
6608 untrusted sources. OpenSSL command line applications could also be
6609 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6610 as command line arguments.
6612 Libssl is not considered directly vulnerable. Additionally certificates etc
6613 received via remote connections via libssl are also unlikely to be able to
6614 trigger these issues because of message size limits enforced within libssl.
6616 This issue was reported to OpenSSL Guido Vranken.
6621 * Side channel attack on modular exponentiation
6623 A side-channel attack was found which makes use of cache-bank conflicts on
6624 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6625 of RSA keys. The ability to exploit this issue is limited as it relies on
6626 an attacker who has control of code in a thread running on the same
6627 hyper-threaded core as the victim thread which is performing decryptions.
6629 This issue was reported to OpenSSL by Yuval Yarom, The University of
6630 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6631 Nadia Heninger, University of Pennsylvania with more information at
6632 <http://cachebleed.info>.
6637 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6638 if no keysize is specified with default_bits. This fixes an
6639 omission in an earlier change that changed all RSA/DSA key generation
6640 commands to use 2048 bits by default.
6644 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6646 * Protection for DH small subgroup attacks
6648 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6649 switched on by default and cannot be disabled. This could have some
6654 * SSLv2 doesn't block disabled ciphers
6656 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6657 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6658 been disabled, provided that the SSLv2 protocol was not also disabled via
6661 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6662 and Sebastian Schinzel.
6667 * Reject DH handshakes with parameters shorter than 1024 bits.
6671 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6673 * Certificate verify crash with missing PSS parameter
6675 The signature verification routines will crash with a NULL pointer
6676 dereference if presented with an ASN.1 signature using the RSA PSS
6677 algorithm and absent mask generation function parameter. Since these
6678 routines are used to verify certificate signature algorithms this can be
6679 used to crash any certificate verification operation and exploited in a
6680 DoS attack. Any application which performs certificate verification is
6681 vulnerable including OpenSSL clients and servers which enable client
6684 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6689 * X509_ATTRIBUTE memory leak
6691 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6692 memory. This structure is used by the PKCS#7 and CMS routines so any
6693 application which reads PKCS#7 or CMS data from untrusted sources is
6694 affected. SSL/TLS is not affected.
6696 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6702 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6703 This changes the decoding behaviour for some invalid messages,
6704 though the change is mostly in the more lenient direction, and
6705 legacy behaviour is preserved as much as possible.
6709 * In DSA_generate_parameters_ex, if the provided seed is too short,
6710 use a random seed, as already documented.
6712 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6714 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6716 * Alternate chains certificate forgery
6718 During certificate verfification, OpenSSL will attempt to find an
6719 alternative certificate chain if the first attempt to build such a chain
6720 fails. An error in the implementation of this logic can mean that an
6721 attacker could cause certain checks on untrusted certificates to be
6722 bypassed, such as the CA flag, enabling them to use a valid leaf
6723 certificate to act as a CA and "issue" an invalid certificate.
6725 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6731 * Race condition handling PSK identify hint
6733 If PSK identity hints are received by a multi-threaded client then
6734 the values are wrongly updated in the parent SSL_CTX structure. This can
6735 result in a race condition potentially leading to a double free of the
6741 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6743 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6744 incompatibility in the handling of HMAC. The previous ABI has now been
6747 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6749 * Malformed ECParameters causes infinite loop
6751 When processing an ECParameters structure OpenSSL enters an infinite loop
6752 if the curve specified is over a specially malformed binary polynomial
6755 This can be used to perform denial of service against any
6756 system which processes public keys, certificate requests or
6757 certificates. This includes TLS clients and TLS servers with
6758 client authentication enabled.
6760 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6765 * Exploitable out-of-bounds read in X509_cmp_time
6767 X509_cmp_time does not properly check the length of the ASN1_TIME
6768 string and can read a few bytes out of bounds. In addition,
6769 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6772 An attacker can use this to craft malformed certificates and CRLs of
6773 various sizes and potentially cause a segmentation fault, resulting in
6774 a DoS on applications that verify certificates or CRLs. TLS clients
6775 that verify CRLs are affected. TLS clients and servers with client
6776 authentication enabled may be affected if they use custom verification
6779 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6780 independently by Hanno Böck.
6785 * PKCS7 crash with missing EnvelopedContent
6787 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6788 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6789 with missing content and trigger a NULL pointer dereference on parsing.
6791 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6792 structures from untrusted sources are affected. OpenSSL clients and
6793 servers are not affected.
6795 This issue was reported to OpenSSL by Michal Zalewski (Google).
6800 * CMS verify infinite loop with unknown hash function
6802 When verifying a signedData message the CMS code can enter an infinite loop
6803 if presented with an unknown hash function OID. This can be used to perform
6804 denial of service against any system which verifies signedData messages using
6806 This issue was reported to OpenSSL by Johannes Bauer.
6811 * Race condition handling NewSessionTicket
6813 If a NewSessionTicket is received by a multi-threaded client when attempting to
6814 reuse a previous ticket then a race condition can occur potentially leading to
6815 a double free of the ticket data.
6820 * Reject DH handshakes with parameters shorter than 768 bits.
6822 *Kurt Roeckx and Emilia Kasper*
6824 * dhparam: generate 2048-bit parameters by default.
6826 *Kurt Roeckx and Emilia Kasper*
6828 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6830 * Segmentation fault in ASN1_TYPE_cmp fix
6832 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6833 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6834 certificate signature algorithm consistency this can be used to crash any
6835 certificate verification operation and exploited in a DoS attack. Any
6836 application which performs certificate verification is vulnerable including
6837 OpenSSL clients and servers which enable client authentication.
6842 * ASN.1 structure reuse memory corruption fix
6844 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6845 memory corruption via an invalid write. Such reuse is and has been
6846 strongly discouraged and is believed to be rare.
6848 Applications that parse structures containing CHOICE or ANY DEFINED BY
6849 components may be affected. Certificate parsing (d2i_X509 and related
6850 functions) are however not affected. OpenSSL clients and servers are
6856 * PKCS7 NULL pointer dereferences fix
6858 The PKCS#7 parsing code does not handle missing outer ContentInfo
6859 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6860 missing content and trigger a NULL pointer dereference on parsing.
6862 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6863 otherwise parse PKCS#7 structures from untrusted sources are
6864 affected. OpenSSL clients and servers are not affected.
6866 This issue was reported to OpenSSL by Michal Zalewski (Google).
6871 * DoS via reachable assert in SSLv2 servers fix
6873 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6874 servers that both support SSLv2 and enable export cipher suites by sending
6875 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6877 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6878 (OpenSSL development team).
6883 * Use After Free following d2i_ECPrivatekey error fix
6885 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6886 could cause a use after free condition. This, in turn, could cause a double
6887 free in several private key parsing functions (such as d2i_PrivateKey
6888 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6889 for applications that receive EC private keys from untrusted
6890 sources. This scenario is considered rare.
6892 This issue was discovered by the BoringSSL project and fixed in their
6898 * X509_to_X509_REQ NULL pointer deref fix
6900 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6901 the certificate key is invalid. This function is rarely used in practice.
6903 This issue was discovered by Brian Carpenter.
6908 * Removed the export ciphers from the DEFAULT ciphers
6912 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6914 * Build fixes for the Windows and OpenVMS platforms
6916 *Matt Caswell and Richard Levitte*
6918 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6920 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6921 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6922 dereference. This could lead to a Denial Of Service attack. Thanks to
6923 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6928 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6929 dtls1_buffer_record function under certain conditions. In particular this
6930 could occur if an attacker sent repeated DTLS records with the same
6931 sequence number but for the next epoch. The memory leak could be exploited
6932 by an attacker in a Denial of Service attack through memory exhaustion.
6933 Thanks to Chris Mueller for reporting this issue.
6938 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6939 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6940 method would be set to NULL which could later result in a NULL pointer
6941 dereference. Thanks to Frank Schmirler for reporting this issue.
6946 * Abort handshake if server key exchange message is omitted for ephemeral
6949 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6950 reporting this issue.
6955 * Remove non-export ephemeral RSA code on client and server. This code
6956 violated the TLS standard by allowing the use of temporary RSA keys in
6957 non-export ciphersuites and could be used by a server to effectively
6958 downgrade the RSA key length used to a value smaller than the server
6959 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6960 INRIA or reporting this issue.
6965 * Fixed issue where DH client certificates are accepted without verification.
6966 An OpenSSL server will accept a DH certificate for client authentication
6967 without the certificate verify message. This effectively allows a client to
6968 authenticate without the use of a private key. This only affects servers
6969 which trust a client certificate authority which issues certificates
6970 containing DH keys: these are extremely rare and hardly ever encountered.
6971 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6977 * Ensure that the session ID context of an SSL is updated when its
6978 SSL_CTX is updated via SSL_set_SSL_CTX.
6980 The session ID context is typically set from the parent SSL_CTX,
6981 and can vary with the CTX.
6985 * Fix various certificate fingerprint issues.
6987 By using non-DER or invalid encodings outside the signed portion of a
6988 certificate the fingerprint can be changed without breaking the signature.
6989 Although no details of the signed portion of the certificate can be changed
6990 this can cause problems with some applications: e.g. those using the
6991 certificate fingerprint for blacklists.
6993 1. Reject signatures with non zero unused bits.
6995 If the BIT STRING containing the signature has non zero unused bits reject
6996 the signature. All current signature algorithms require zero unused bits.
6998 2. Check certificate algorithm consistency.
7000 Check the AlgorithmIdentifier inside TBS matches the one in the
7001 certificate signature. NB: this will result in signature failure
7002 errors for some broken certificates.
7004 Thanks to Konrad Kraszewski from Google for reporting this issue.
7006 3. Check DSA/ECDSA signatures use DER.
7008 Re-encode DSA/ECDSA signatures and compare with the original received
7009 signature. Return an error if there is a mismatch.
7011 This will reject various cases including garbage after signature
7012 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7013 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7014 (negative or with leading zeroes).
7016 Further analysis was conducted and fixes were developed by Stephen Henson
7017 of the OpenSSL core team.
7023 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7024 results on some platforms, including x86_64. This bug occurs at random
7025 with a very low probability, and is not known to be exploitable in any
7026 way, though its exact impact is difficult to determine. Thanks to Pieter
7027 Wuille (Blockstream) who reported this issue and also suggested an initial
7028 fix. Further analysis was conducted by the OpenSSL development team and
7029 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7030 the OpenSSL core team.
7035 * Do not resume sessions on the server if the negotiated protocol
7036 version does not match the session's version. Resuming with a different
7037 version, while not strictly forbidden by the RFC, is of questionable
7038 sanity and breaks all known clients.
7040 *David Benjamin, Emilia Käsper*
7042 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7043 early CCS messages during renegotiation. (Note that because
7044 renegotiation is encrypted, this early CCS was not exploitable.)
7048 * Tighten client-side session ticket handling during renegotiation:
7049 ensure that the client only accepts a session ticket if the server sends
7050 the extension anew in the ServerHello. Previously, a TLS client would
7051 reuse the old extension state and thus accept a session ticket if one was
7052 announced in the initial ServerHello.
7054 Similarly, ensure that the client requires a session ticket if one
7055 was advertised in the ServerHello. Previously, a TLS client would
7056 ignore a missing NewSessionTicket message.
7060 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7064 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7065 sends a carefully crafted handshake message, to cause OpenSSL to fail
7066 to free up to 64k of memory causing a memory leak. This could be
7067 exploited in a Denial Of Service attack. This issue affects OpenSSL
7068 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7069 whether SRTP is used or configured. Implementations of OpenSSL that
7070 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7072 The fix was developed by the OpenSSL team.
7077 * Session Ticket Memory Leak.
7079 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7080 integrity of that ticket is first verified. In the event of a session
7081 ticket integrity check failing, OpenSSL will fail to free memory
7082 causing a memory leak. By sending a large number of invalid session
7083 tickets an attacker could exploit this issue in a Denial Of Service
7089 * Build option no-ssl3 is incomplete.
7091 When OpenSSL is configured with "no-ssl3" as a build option, servers
7092 could accept and complete a SSL 3.0 handshake, and clients could be
7093 configured to send them.
7096 *Akamai and the OpenSSL team*
7098 * Add support for TLS_FALLBACK_SCSV.
7099 Client applications doing fallback retries should call
7100 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7103 *Adam Langley, Bodo Moeller*
7105 * Add additional DigestInfo checks.
7107 Re-encode DigestInto in DER and check against the original when
7108 verifying RSA signature: this will reject any improperly encoded
7109 DigestInfo structures.
7111 Note: this is a precautionary measure and no attacks are currently known.
7115 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7117 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7118 SRP code can be overrun an internal buffer. Add sanity check that
7119 g, A, B < N to SRP code.
7121 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7122 Group for discovering this issue.
7127 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7128 TLS 1.0 instead of higher protocol versions when the ClientHello message
7129 is badly fragmented. This allows a man-in-the-middle attacker to force a
7130 downgrade to TLS 1.0 even if both the server and the client support a
7131 higher protocol version, by modifying the client's TLS records.
7133 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7134 researching this issue.
7139 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7140 to a denial of service attack. A malicious server can crash the client
7141 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7142 ciphersuite and sending carefully crafted handshake messages.
7144 Thanks to Felix Gröbert (Google) for discovering and researching this
7150 * By sending carefully crafted DTLS packets an attacker could cause openssl
7151 to leak memory. This can be exploited through a Denial of Service attack.
7152 Thanks to Adam Langley for discovering and researching this issue.
7157 * An attacker can force openssl to consume large amounts of memory whilst
7158 processing DTLS handshake messages. This can be exploited through a
7159 Denial of Service attack.
7160 Thanks to Adam Langley for discovering and researching this issue.
7165 * An attacker can force an error condition which causes openssl to crash
7166 whilst processing DTLS packets due to memory being freed twice. This
7167 can be exploited through a Denial of Service attack.
7168 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7174 * If a multithreaded client connects to a malicious server using a resumed
7175 session and the server sends an ec point format extension it could write
7176 up to 255 bytes to freed memory.
7178 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7184 * A malicious server can crash an OpenSSL client with a null pointer
7185 dereference (read) by specifying an SRP ciphersuite even though it was not
7186 properly negotiated with the client. This can be exploited through a
7187 Denial of Service attack.
7189 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7190 discovering and researching this issue.
7195 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7196 X509_name_oneline, X509_name_print_ex et al. to leak some information
7197 from the stack. Applications may be affected if they echo pretty printing
7198 output to the attacker.
7200 Thanks to Ivan Fratric (Google) for discovering this issue.
7203 *Emilia Käsper, and Steve Henson*
7205 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7206 for corner cases. (Certain input points at infinity could lead to
7207 bogus results, with non-infinity inputs mapped to infinity too.)
7211 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7213 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7214 handshake can force the use of weak keying material in OpenSSL
7215 SSL/TLS clients and servers.
7217 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7218 researching this issue. ([CVE-2014-0224])
7220 *KIKUCHI Masashi, Steve Henson*
7222 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7223 OpenSSL DTLS client the code can be made to recurse eventually crashing
7226 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7229 *Imre Rad, Steve Henson*
7231 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7232 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7233 client or server. This is potentially exploitable to run arbitrary
7234 code on a vulnerable client or server.
7236 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7238 *Jüri Aedla, Steve Henson*
7240 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7241 are subject to a denial of service attack.
7243 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7244 this issue. ([CVE-2014-3470])
7246 *Felix Gröbert, Ivan Fratric, Steve Henson*
7248 * Harmonize version and its documentation. -f flag is used to display
7251 *mancha <mancha1@zoho.com>*
7253 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7254 in i2d_ECPrivateKey.
7256 *mancha <mancha1@zoho.com>*
7258 * Fix some double frees. These are not thought to be exploitable.
7260 *mancha <mancha1@zoho.com>*
7262 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7264 * A missing bounds check in the handling of the TLS heartbeat extension
7265 can be used to reveal up to 64k of memory to a connected client or
7268 Thanks for Neel Mehta of Google Security for discovering this bug and to
7269 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7270 preparing the fix ([CVE-2014-0160])
7272 *Adam Langley, Bodo Moeller*
7274 * Fix for the attack described in the paper "Recovering OpenSSL
7275 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7276 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7277 <http://eprint.iacr.org/2014/140>
7279 Thanks to Yuval Yarom and Naomi Benger for discovering this
7280 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7282 *Yuval Yarom and Naomi Benger*
7284 * TLS pad extension: draft-agl-tls-padding-03
7286 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7287 TLS client Hello record length value would otherwise be > 255 and
7288 less that 512 pad with a dummy extension containing zeroes so it
7289 is at least 512 bytes long.
7291 *Adam Langley, Steve Henson*
7293 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7295 * Fix for TLS record tampering bug. A carefully crafted invalid
7296 handshake could crash OpenSSL with a NULL pointer exception.
7297 Thanks to Anton Johansson for reporting this issues.
7300 * Keep original DTLS digest and encryption contexts in retransmission
7301 structures so we can use the previous session parameters if they need
7302 to be resent. ([CVE-2013-6450])
7306 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7307 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7308 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7309 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7310 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7311 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7313 *Rob Stradling, Adam Langley*
7315 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7317 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7318 supporting platforms or when small records were transferred.
7320 *Andy Polyakov, Steve Henson*
7322 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7324 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7326 This addresses the flaw in CBC record processing discovered by
7327 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7328 at: <http://www.isg.rhul.ac.uk/tls/>
7330 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7331 Security Group at Royal Holloway, University of London
7332 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7333 Emilia Käsper for the initial patch.
7336 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7338 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7339 ciphersuites which can be exploited in a denial of service attack.
7340 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7341 and detecting this bug and to Wolfgang Ettlinger
7342 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7347 * Return an error when checking OCSP signatures when key is NULL.
7348 This fixes a DoS attack. ([CVE-2013-0166])
7352 * Make openssl verify return errors.
7354 *Chris Palmer <palmer@google.com> and Ben Laurie*
7356 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7357 the right response is stapled. Also change SSL_get_certificate()
7358 so it returns the certificate actually sent.
7359 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7361 *Rob Stradling <rob.stradling@comodo.com>*
7363 * Fix possible deadlock when decoding public keys.
7367 * Don't use TLS 1.0 record version number in initial client hello
7372 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7374 * Sanity check record length before skipping explicit IV in TLS
7375 1.2, 1.1 and DTLS to fix DoS attack.
7377 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7378 fuzzing as a service testing platform.
7383 * Initialise tkeylen properly when encrypting CMS messages.
7384 Thanks to Solar Designer of Openwall for reporting this issue.
7388 * In FIPS mode don't try to use composite ciphers as they are not
7393 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7395 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7396 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7397 mean any application compiled against OpenSSL 1.0.0 headers setting
7398 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7399 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7400 0x10000000L Any application which was previously compiled against
7401 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7402 will need to be recompiled as a result. Letting be results in
7403 inability to disable specifically TLS 1.1 and in client context,
7404 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7408 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7409 disable just protocol X, but all protocols above X *if* there are
7410 protocols *below* X still enabled. In more practical terms it means
7411 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7412 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7413 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7418 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7420 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7421 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7422 in CRYPTO_realloc_clean.
7424 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7425 issue and to Adam Langley <agl@chromium.org> for fixing it.
7428 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7430 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7434 * Workarounds for some broken servers that "hang" if a client hello
7435 record length exceeds 255 bytes.
7437 1. Do not use record version number > TLS 1.0 in initial client
7438 hello: some (but not all) hanging servers will now work.
7439 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7440 the number of ciphers sent in the client hello. This should be
7441 set to an even number, such as 50, for example by passing:
7442 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7443 Most broken servers should now work.
7444 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7445 TLS 1.2 client support entirely.
7449 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7453 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7455 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7456 STRING form instead of a DigestInfo.
7460 * The format used for MDC2 RSA signatures is inconsistent between EVP
7461 and the RSA_sign/RSA_verify functions. This was made more apparent when
7462 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7463 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7464 the correct format in RSA_verify so both forms transparently work.
7468 * Some servers which support TLS 1.0 can choke if we initially indicate
7469 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7470 encrypted premaster secret. As a workaround use the maximum permitted
7471 client version in client hello, this should keep such servers happy
7472 and still work with previous versions of OpenSSL.
7476 * Add support for TLS/DTLS heartbeats.
7478 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7480 * Add support for SCTP.
7482 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7484 * Improved PRNG seeding for VOS.
7486 *Paul Green <Paul.Green@stratus.com>*
7488 * Extensive assembler packs updates, most notably:
7490 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7491 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7492 - x86_64: bit-sliced AES implementation;
7493 - ARM: NEON support, contemporary platforms optimizations;
7494 - s390x: z196 support;
7495 - `*`: GHASH and GF(2^m) multiplication implementations;
7499 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7500 (removal of unnecessary code)
7502 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7504 * Add TLS key material exporter from RFC 5705.
7508 * Add DTLS-SRTP negotiation from RFC 5764.
7512 * Add Next Protocol Negotiation,
7513 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7514 disabled with a no-npn flag to config or Configure. Code donated
7517 *Adam Langley <agl@google.com> and Ben Laurie*
7519 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7520 NIST-P256, NIST-P521, with constant-time single point multiplication on
7521 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7522 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7523 Code made available under Apache License version 2.0.
7525 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7526 line to include this in your build of OpenSSL, and run "make depend" (or
7527 "make update"). This enables the following EC_METHODs:
7529 EC_GFp_nistp224_method()
7530 EC_GFp_nistp256_method()
7531 EC_GFp_nistp521_method()
7533 EC_GROUP_new_by_curve_name() will automatically use these (while
7534 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7537 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7539 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7540 all platforms. Move ssize_t definition from e_os.h to the public
7541 header file e_os2.h as it now appears in public header file cms.h
7545 * New -sigopt option to the ca, req and x509 utilities. Additional
7546 signature parameters can be passed using this option and in
7551 * Add RSA PSS signing function. This will generate and set the
7552 appropriate AlgorithmIdentifiers for PSS based on those in the
7553 corresponding EVP_MD_CTX structure. No application support yet.
7557 * Support for companion algorithm specific ASN1 signing routines.
7558 New function ASN1_item_sign_ctx() signs a pre-initialised
7559 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7560 the appropriate parameters.
7564 * Add new algorithm specific ASN1 verification initialisation function
7565 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7566 handling will be the same no matter what EVP_PKEY_METHOD is used.
7567 Add a PSS handler to support verification of PSS signatures: checked
7568 against a number of sample certificates.
7572 * Add signature printing for PSS. Add PSS OIDs.
7574 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7576 * Add algorithm specific signature printing. An individual ASN1 method
7577 can now print out signatures instead of the standard hex dump.
7579 More complex signatures (e.g. PSS) can print out more meaningful
7580 information. Include DSA version that prints out the signature
7585 * Password based recipient info support for CMS library: implementing
7590 * Split password based encryption into PBES2 and PBKDF2 functions. This
7591 neatly separates the code into cipher and PBE sections and is required
7592 for some algorithms that split PBES2 into separate pieces (such as
7593 password based CMS).
7597 * Session-handling fixes:
7598 - Fix handling of connections that are resuming with a session ID,
7599 but also support Session Tickets.
7600 - Fix a bug that suppressed issuing of a new ticket if the client
7601 presented a ticket with an expired session.
7602 - Try to set the ticket lifetime hint to something reasonable.
7603 - Make tickets shorter by excluding irrelevant information.
7604 - On the client side, don't ignore renewed tickets.
7606 *Adam Langley, Bodo Moeller (Google)*
7608 * Fix PSK session representation.
7612 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7614 This work was sponsored by Intel.
7618 * Add GCM support to TLS library. Some custom code is needed to split
7619 the IV between the fixed (from PRF) and explicit (from TLS record)
7620 portions. This adds all GCM ciphersuites supported by RFC5288 and
7621 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7622 add a special AESGCM string for GCM only.
7626 * Expand range of ctrls for AES GCM. Permit setting invocation
7627 field on decrypt and retrieval of invocation field only on encrypt.
7631 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7632 As required by RFC5289 these ciphersuites cannot be used if for
7633 versions of TLS earlier than 1.2.
7637 * For FIPS capable OpenSSL interpret a NULL default public key method
7638 as unset and return the appropriate default but do *not* set the default.
7639 This means we can return the appropriate method in applications that
7640 switch between FIPS and non-FIPS modes.
7644 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7645 ENGINE is used then we cannot handle that in the FIPS module so we
7646 keep original code iff non-FIPS operations are allowed.
7650 * Add -attime option to openssl utilities.
7652 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7654 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7658 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7659 FIPS EC methods unconditionally for now.
7663 * New build option no-ec2m to disable characteristic 2 code.
7667 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7668 all cases can be covered as some introduce binary incompatibilities.
7672 * Redirect RSA operations to FIPS module including keygen,
7673 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7677 * Add similar low-level API blocking to ciphers.
7681 * low-level digest APIs are not approved in FIPS mode: any attempt
7682 to use these will cause a fatal error. Applications that *really* want
7683 to use them can use the `private_*` version instead.
7687 * Redirect cipher operations to FIPS module for FIPS builds.
7691 * Redirect digest operations to FIPS module for FIPS builds.
7695 * Update build system to add "fips" flag which will link in fipscanister.o
7696 for static and shared library builds embedding a signature if needed.
7700 * Output TLS supported curves in preference order instead of numerical
7701 order. This is currently hardcoded for the highest order curves first.
7702 This should be configurable so applications can judge speed vs strength.
7706 * Add TLS v1.2 server support for client authentication.
7710 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7715 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7716 FIPS modules versions.
7720 * Add TLS v1.2 client side support for client authentication. Keep cache
7721 of handshake records longer as we don't know the hash algorithm to use
7722 until after the certificate request message is received.
7726 * Initial TLS v1.2 client support. Add a default signature algorithms
7727 extension including all the algorithms we support. Parse new signature
7728 format in client key exchange. Relax some ECC signing restrictions for
7729 TLS v1.2 as indicated in RFC5246.
7733 * Add server support for TLS v1.2 signature algorithms extension. Switch
7734 to new signature format when needed using client digest preference.
7735 All server ciphersuites should now work correctly in TLS v1.2. No client
7736 support yet and no support for client certificates.
7740 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7741 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7742 ciphersuites. At present only RSA key exchange ciphersuites work with
7743 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7744 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7745 and version checking.
7749 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7750 with this defined it will not be affected by any changes to ssl internal
7751 structures. Add several utility functions to allow openssl application
7752 to work with OPENSSL_NO_SSL_INTERN defined.
7756 * A long standing patch to add support for SRP from EdelWeb (Peter
7757 Sylvester and Christophe Renou) was integrated.
7758 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7759 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7762 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7766 * Permit abbreviated handshakes when renegotiating using the function
7767 SSL_renegotiate_abbreviated().
7769 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7771 * Add call to ENGINE_register_all_complete() to
7772 ENGINE_load_builtin_engines(), so some implementations get used
7773 automatically instead of needing explicit application support.
7777 * Add support for TLS key exporter as described in RFC5705.
7779 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7781 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7782 a few changes are required:
7784 Add SSL_OP_NO_TLSv1_1 flag.
7785 Add TLSv1_1 methods.
7786 Update version checking logic to handle version 1.1.
7787 Add explicit IV handling (ported from DTLS code).
7788 Add command line options to s_client/s_server.
7795 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7797 * X509_ATTRIBUTE memory leak
7799 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7800 memory. This structure is used by the PKCS#7 and CMS routines so any
7801 application which reads PKCS#7 or CMS data from untrusted sources is
7802 affected. SSL/TLS is not affected.
7804 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7810 * Race condition handling PSK identify hint
7812 If PSK identity hints are received by a multi-threaded client then
7813 the values are wrongly updated in the parent SSL_CTX structure. This can
7814 result in a race condition potentially leading to a double free of the
7820 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7822 * Malformed ECParameters causes infinite loop
7824 When processing an ECParameters structure OpenSSL enters an infinite loop
7825 if the curve specified is over a specially malformed binary polynomial
7828 This can be used to perform denial of service against any
7829 system which processes public keys, certificate requests or
7830 certificates. This includes TLS clients and TLS servers with
7831 client authentication enabled.
7833 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7838 * Exploitable out-of-bounds read in X509_cmp_time
7840 X509_cmp_time does not properly check the length of the ASN1_TIME
7841 string and can read a few bytes out of bounds. In addition,
7842 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7845 An attacker can use this to craft malformed certificates and CRLs of
7846 various sizes and potentially cause a segmentation fault, resulting in
7847 a DoS on applications that verify certificates or CRLs. TLS clients
7848 that verify CRLs are affected. TLS clients and servers with client
7849 authentication enabled may be affected if they use custom verification
7852 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7853 independently by Hanno Böck.
7858 * PKCS7 crash with missing EnvelopedContent
7860 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7861 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7862 with missing content and trigger a NULL pointer dereference on parsing.
7864 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7865 structures from untrusted sources are affected. OpenSSL clients and
7866 servers are not affected.
7868 This issue was reported to OpenSSL by Michal Zalewski (Google).
7873 * CMS verify infinite loop with unknown hash function
7875 When verifying a signedData message the CMS code can enter an infinite loop
7876 if presented with an unknown hash function OID. This can be used to perform
7877 denial of service against any system which verifies signedData messages using
7879 This issue was reported to OpenSSL by Johannes Bauer.
7884 * Race condition handling NewSessionTicket
7886 If a NewSessionTicket is received by a multi-threaded client when attempting to
7887 reuse a previous ticket then a race condition can occur potentially leading to
7888 a double free of the ticket data.
7893 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7895 * Segmentation fault in ASN1_TYPE_cmp fix
7897 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7898 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7899 certificate signature algorithm consistency this can be used to crash any
7900 certificate verification operation and exploited in a DoS attack. Any
7901 application which performs certificate verification is vulnerable including
7902 OpenSSL clients and servers which enable client authentication.
7907 * ASN.1 structure reuse memory corruption fix
7909 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7910 memory corruption via an invalid write. Such reuse is and has been
7911 strongly discouraged and is believed to be rare.
7913 Applications that parse structures containing CHOICE or ANY DEFINED BY
7914 components may be affected. Certificate parsing (d2i_X509 and related
7915 functions) are however not affected. OpenSSL clients and servers are
7921 * PKCS7 NULL pointer dereferences fix
7923 The PKCS#7 parsing code does not handle missing outer ContentInfo
7924 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7925 missing content and trigger a NULL pointer dereference on parsing.
7927 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7928 otherwise parse PKCS#7 structures from untrusted sources are
7929 affected. OpenSSL clients and servers are not affected.
7931 This issue was reported to OpenSSL by Michal Zalewski (Google).
7936 * DoS via reachable assert in SSLv2 servers fix
7938 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7939 servers that both support SSLv2 and enable export cipher suites by sending
7940 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7942 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7943 (OpenSSL development team).
7948 * Use After Free following d2i_ECPrivatekey error fix
7950 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7951 could cause a use after free condition. This, in turn, could cause a double
7952 free in several private key parsing functions (such as d2i_PrivateKey
7953 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7954 for applications that receive EC private keys from untrusted
7955 sources. This scenario is considered rare.
7957 This issue was discovered by the BoringSSL project and fixed in their
7963 * X509_to_X509_REQ NULL pointer deref fix
7965 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7966 the certificate key is invalid. This function is rarely used in practice.
7968 This issue was discovered by Brian Carpenter.
7973 * Removed the export ciphers from the DEFAULT ciphers
7977 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7979 * Build fixes for the Windows and OpenVMS platforms
7981 *Matt Caswell and Richard Levitte*
7983 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7985 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7986 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7987 dereference. This could lead to a Denial Of Service attack. Thanks to
7988 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7993 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7994 dtls1_buffer_record function under certain conditions. In particular this
7995 could occur if an attacker sent repeated DTLS records with the same
7996 sequence number but for the next epoch. The memory leak could be exploited
7997 by an attacker in a Denial of Service attack through memory exhaustion.
7998 Thanks to Chris Mueller for reporting this issue.
8003 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8004 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8005 method would be set to NULL which could later result in a NULL pointer
8006 dereference. Thanks to Frank Schmirler for reporting this issue.
8011 * Abort handshake if server key exchange message is omitted for ephemeral
8014 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8015 reporting this issue.
8020 * Remove non-export ephemeral RSA code on client and server. This code
8021 violated the TLS standard by allowing the use of temporary RSA keys in
8022 non-export ciphersuites and could be used by a server to effectively
8023 downgrade the RSA key length used to a value smaller than the server
8024 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8025 INRIA or reporting this issue.
8030 * Fixed issue where DH client certificates are accepted without verification.
8031 An OpenSSL server will accept a DH certificate for client authentication
8032 without the certificate verify message. This effectively allows a client to
8033 authenticate without the use of a private key. This only affects servers
8034 which trust a client certificate authority which issues certificates
8035 containing DH keys: these are extremely rare and hardly ever encountered.
8036 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8042 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8043 results on some platforms, including x86_64. This bug occurs at random
8044 with a very low probability, and is not known to be exploitable in any
8045 way, though its exact impact is difficult to determine. Thanks to Pieter
8046 Wuille (Blockstream) who reported this issue and also suggested an initial
8047 fix. Further analysis was conducted by the OpenSSL development team and
8048 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8049 the OpenSSL core team.
8054 * Fix various certificate fingerprint issues.
8056 By using non-DER or invalid encodings outside the signed portion of a
8057 certificate the fingerprint can be changed without breaking the signature.
8058 Although no details of the signed portion of the certificate can be changed
8059 this can cause problems with some applications: e.g. those using the
8060 certificate fingerprint for blacklists.
8062 1. Reject signatures with non zero unused bits.
8064 If the BIT STRING containing the signature has non zero unused bits reject
8065 the signature. All current signature algorithms require zero unused bits.
8067 2. Check certificate algorithm consistency.
8069 Check the AlgorithmIdentifier inside TBS matches the one in the
8070 certificate signature. NB: this will result in signature failure
8071 errors for some broken certificates.
8073 Thanks to Konrad Kraszewski from Google for reporting this issue.
8075 3. Check DSA/ECDSA signatures use DER.
8077 Reencode DSA/ECDSA signatures and compare with the original received
8078 signature. Return an error if there is a mismatch.
8080 This will reject various cases including garbage after signature
8081 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8082 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8083 (negative or with leading zeroes).
8085 Further analysis was conducted and fixes were developed by Stephen Henson
8086 of the OpenSSL core team.
8092 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8094 * Session Ticket Memory Leak.
8096 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8097 integrity of that ticket is first verified. In the event of a session
8098 ticket integrity check failing, OpenSSL will fail to free memory
8099 causing a memory leak. By sending a large number of invalid session
8100 tickets an attacker could exploit this issue in a Denial Of Service
8106 * Build option no-ssl3 is incomplete.
8108 When OpenSSL is configured with "no-ssl3" as a build option, servers
8109 could accept and complete a SSL 3.0 handshake, and clients could be
8110 configured to send them.
8113 *Akamai and the OpenSSL team*
8115 * Add support for TLS_FALLBACK_SCSV.
8116 Client applications doing fallback retries should call
8117 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8120 *Adam Langley, Bodo Moeller*
8122 * Add additional DigestInfo checks.
8124 Reencode DigestInto in DER and check against the original when
8125 verifying RSA signature: this will reject any improperly encoded
8126 DigestInfo structures.
8128 Note: this is a precautionary measure and no attacks are currently known.
8132 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8134 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8135 to a denial of service attack. A malicious server can crash the client
8136 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8137 ciphersuite and sending carefully crafted handshake messages.
8139 Thanks to Felix Gröbert (Google) for discovering and researching this
8145 * By sending carefully crafted DTLS packets an attacker could cause openssl
8146 to leak memory. This can be exploited through a Denial of Service attack.
8147 Thanks to Adam Langley for discovering and researching this issue.
8152 * An attacker can force openssl to consume large amounts of memory whilst
8153 processing DTLS handshake messages. This can be exploited through a
8154 Denial of Service attack.
8155 Thanks to Adam Langley for discovering and researching this issue.
8160 * An attacker can force an error condition which causes openssl to crash
8161 whilst processing DTLS packets due to memory being freed twice. This
8162 can be exploited through a Denial of Service attack.
8163 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8169 * If a multithreaded client connects to a malicious server using a resumed
8170 session and the server sends an ec point format extension it could write
8171 up to 255 bytes to freed memory.
8173 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8179 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8180 X509_name_oneline, X509_name_print_ex et al. to leak some information
8181 from the stack. Applications may be affected if they echo pretty printing
8182 output to the attacker.
8184 Thanks to Ivan Fratric (Google) for discovering this issue.
8187 *Emilia Käsper, and Steve Henson*
8189 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8190 for corner cases. (Certain input points at infinity could lead to
8191 bogus results, with non-infinity inputs mapped to infinity too.)
8195 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8197 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8198 handshake can force the use of weak keying material in OpenSSL
8199 SSL/TLS clients and servers.
8201 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8202 researching this issue. ([CVE-2014-0224])
8204 *KIKUCHI Masashi, Steve Henson*
8206 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8207 OpenSSL DTLS client the code can be made to recurse eventually crashing
8210 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8213 *Imre Rad, Steve Henson*
8215 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8216 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8217 client or server. This is potentially exploitable to run arbitrary
8218 code on a vulnerable client or server.
8220 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8222 *Jüri Aedla, Steve Henson*
8224 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8225 are subject to a denial of service attack.
8227 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8228 this issue. ([CVE-2014-3470])
8230 *Felix Gröbert, Ivan Fratric, Steve Henson*
8232 * Harmonize version and its documentation. -f flag is used to display
8235 *mancha <mancha1@zoho.com>*
8237 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8238 in i2d_ECPrivateKey.
8240 *mancha <mancha1@zoho.com>*
8242 * Fix some double frees. These are not thought to be exploitable.
8244 *mancha <mancha1@zoho.com>*
8246 * Fix for the attack described in the paper "Recovering OpenSSL
8247 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8248 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8249 <http://eprint.iacr.org/2014/140>
8251 Thanks to Yuval Yarom and Naomi Benger for discovering this
8252 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8254 *Yuval Yarom and Naomi Benger*
8256 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8258 * Keep original DTLS digest and encryption contexts in retransmission
8259 structures so we can use the previous session parameters if they need
8260 to be resent. ([CVE-2013-6450])
8264 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8265 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8266 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8267 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8268 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8269 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8271 *Rob Stradling, Adam Langley*
8273 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8275 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8277 This addresses the flaw in CBC record processing discovered by
8278 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8279 at: <http://www.isg.rhul.ac.uk/tls/>
8281 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8282 Security Group at Royal Holloway, University of London
8283 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8284 Emilia Käsper for the initial patch.
8287 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8289 * Return an error when checking OCSP signatures when key is NULL.
8290 This fixes a DoS attack. ([CVE-2013-0166])
8294 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8295 the right response is stapled. Also change SSL_get_certificate()
8296 so it returns the certificate actually sent.
8297 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8298 (This is a backport)
8300 *Rob Stradling <rob.stradling@comodo.com>*
8302 * Fix possible deadlock when decoding public keys.
8306 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8308 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8311 * Sanity check record length before skipping explicit IV in DTLS
8314 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8315 fuzzing as a service testing platform.
8320 * Initialise tkeylen properly when encrypting CMS messages.
8321 Thanks to Solar Designer of Openwall for reporting this issue.
8325 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8327 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8328 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8329 in CRYPTO_realloc_clean.
8331 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8332 issue and to Adam Langley <agl@chromium.org> for fixing it.
8335 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8337 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8339 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8340 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8341 content decryption and always return the same error. Note: this attack
8342 needs on average 2^20 messages so it only affects automated senders. The
8343 old behaviour can be re-enabled in the CMS code by setting the
8344 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8345 an MMA defence is not necessary.
8346 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8347 this issue. ([CVE-2012-0884])
8351 * Fix CVE-2011-4619: make sure we really are receiving a
8352 client hello before rejecting multiple SGC restarts. Thanks to
8353 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8357 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8359 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8360 Thanks to Antonio Martin, Enterprise Secure Access Research and
8361 Development, Cisco Systems, Inc. for discovering this bug and
8362 preparing a fix. ([CVE-2012-0050])
8366 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8368 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8369 of the Vaudenay padding oracle attack on CBC mode encryption
8370 which enables an efficient plaintext recovery attack against
8371 the OpenSSL implementation of DTLS. Their attack exploits timing
8372 differences arising during decryption processing. A research
8373 paper describing this attack can be found at:
8374 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8375 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8376 Security Group at Royal Holloway, University of London
8377 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8378 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8379 for preparing the fix. ([CVE-2011-4108])
8381 *Robin Seggelmann, Michael Tuexen*
8383 * Clear bytes used for block padding of SSL 3.0 records.
8386 *Adam Langley (Google)*
8388 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8389 Kadianakis <desnacked@gmail.com> for discovering this issue and
8390 Adam Langley for preparing the fix. ([CVE-2011-4619])
8392 *Adam Langley (Google)*
8394 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8396 *Andrey Kulikov <amdeich@gmail.com>*
8398 * Prevent malformed RFC3779 data triggering an assertion failure.
8399 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8400 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8402 *Rob Austein <sra@hactrn.net>*
8404 * Improved PRNG seeding for VOS.
8406 *Paul Green <Paul.Green@stratus.com>*
8408 * Fix ssl_ciph.c set-up race.
8410 *Adam Langley (Google)*
8412 * Fix spurious failures in ecdsatest.c.
8414 *Emilia Käsper (Google)*
8416 * Fix the BIO_f_buffer() implementation (which was mixing different
8417 interpretations of the `..._len` fields).
8419 *Adam Langley (Google)*
8421 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8422 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8423 threads won't reuse the same blinding coefficients.
8425 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8426 lock to call BN_BLINDING_invert_ex, and avoids one use of
8427 BN_BLINDING_update for each BN_BLINDING structure (previously,
8428 the last update always remained unused).
8430 *Emilia Käsper (Google)*
8432 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8434 *Bob Buckholz (Google)*
8436 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8438 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8439 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8441 *Kaspar Brand <ossl@velox.ch>*
8443 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8444 for multi-threaded use of ECDH. ([CVE-2011-3210])
8446 *Adam Langley (Google)*
8448 * Fix x509_name_ex_d2i memory leak on bad inputs.
8452 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8453 signature public key algorithm by using OID xref utilities instead.
8454 Before this you could only use some ECC ciphersuites with SHA1 only.
8458 * Add protection against ECDSA timing attacks as mentioned in the paper
8459 by Billy Bob Brumley and Nicola Tuveri, see:
8460 <http://eprint.iacr.org/2011/232.pdf>
8462 *Billy Bob Brumley and Nicola Tuveri*
8464 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8466 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8468 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8470 * Fix bug in string printing code: if *any* escaping is enabled we must
8471 escape the escape character (backslash) or the resulting string is
8476 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8478 * Disable code workaround for ancient and obsolete Netscape browsers
8479 and servers: an attacker can use it in a ciphersuite downgrade attack.
8480 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8484 * Fixed J-PAKE implementation error, originally discovered by
8485 Sebastien Martini, further info and confirmation from Stefan
8486 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8490 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8492 * Fix extension code to avoid race conditions which can result in a buffer
8493 overrun vulnerability: resumed sessions must not be modified as they can
8494 be shared by multiple threads. CVE-2010-3864
8498 * Fix WIN32 build system to correctly link an ENGINE directory into
8503 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8505 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8508 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8510 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8512 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8513 context. The operation can be customised via the ctrl mechanism in
8514 case ENGINEs want to include additional functionality.
8518 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8522 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8523 output hashes compatible with older versions of OpenSSL.
8525 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8527 * Fix compression algorithm handling: if resuming a session use the
8528 compression algorithm of the resumed session instead of determining
8529 it from client hello again. Don't allow server to change algorithm.
8533 * Add load_crls() function to commands tidying load_certs() too. Add option
8534 to verify utility to allow additional CRLs to be included.
8538 * Update OCSP request code to permit adding custom headers to the request:
8539 some responders need this.
8543 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8546 *Julia Lawall <julia@diku.dk>*
8548 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8549 needlessly dereferenced structures, used obsolete functions and
8550 didn't handle all updated verify codes correctly.
8554 * Disable MD2 in the default configuration.
8558 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8559 indicate the initial BIO being pushed or popped. This makes it possible
8560 to determine whether the BIO is the one explicitly called or as a result
8561 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8562 it handles reference counts correctly and doesn't zero out the I/O bio
8563 when it is not being explicitly popped. WARNING: applications which
8564 included workarounds for the old buggy behaviour will need to be modified
8565 or they could free up already freed BIOs.
8569 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8570 renaming to all platforms (within the 0.9.8 branch, this was
8571 done conditionally on Netware platforms to avoid a name clash).
8573 *Guenter <lists@gknw.net>*
8575 * Add ECDHE and PSK support to DTLS.
8577 *Michael Tuexen <tuexen@fh-muenster.de>*
8579 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8584 * Add "missing" function EVP_MD_flags() (without this the only way to
8585 retrieve a digest flags is by accessing the structure directly. Update
8586 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8587 or cipher is registered as in the "from" argument. Print out all
8588 registered digests in the dgst usage message instead of manually
8589 attempting to work them out.
8593 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8594 this allows the use of compression and extensions. Change default cipher
8595 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8596 by default unless an application cipher string requests it.
8600 * Alter match criteria in PKCS12_parse(). It used to try to use local
8601 key ids to find matching certificates and keys but some PKCS#12 files
8602 don't follow the (somewhat unwritten) rules and this strategy fails.
8603 Now just gather all certificates together and the first private key
8604 then look for the first certificate that matches the key.
8608 * Support use of registered digest and cipher names for dgst and cipher
8609 commands instead of having to add each one as a special case. So now
8616 openssl dgst -sha256 foo
8618 and this works for ENGINE based algorithms too.
8622 * Update Gost ENGINE to support parameter files.
8624 *Victor B. Wagner <vitus@cryptocom.ru>*
8626 * Support GeneralizedTime in ca utility.
8628 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8630 * Enhance the hash format used for certificate directory links. The new
8631 form uses the canonical encoding (meaning equivalent names will work
8632 even if they aren't identical) and uses SHA1 instead of MD5. This form
8633 is incompatible with the older format and as a result c_rehash should
8634 be used to rebuild symbolic links.
8638 * Make PKCS#8 the default write format for private keys, replacing the
8639 traditional format. This form is standardised, more secure and doesn't
8640 include an implicit MD5 dependency.
8644 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8645 committed to OpenSSL should pass this lot as a minimum.
8649 * Add session ticket override functionality for use by EAP-FAST.
8651 *Jouni Malinen <j@w1.fi>*
8653 * Modify HMAC functions to return a value. Since these can be implemented
8654 in an ENGINE errors can occur.
8658 * Type-checked OBJ_bsearch_ex.
8662 * Type-checked OBJ_bsearch. Also some constification necessitated
8663 by type-checking. Still to come: TXT_DB, bsearch(?),
8664 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8669 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8670 seconds to a tm structure directly, instead of going through OS
8671 specific date routines. This avoids any issues with OS routines such
8672 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8673 and X509_time_adj_ex() to cover the extended range. The existing
8674 X509_time_adj() is still usable and will no longer have any date issues.
8678 * Delta CRL support. New use deltas option which will attempt to locate
8679 and search any appropriate delta CRLs available.
8681 This work was sponsored by Google.
8685 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8686 code and add additional score elements. Validate alternate CRL paths
8687 as part of the CRL checking and indicate a new error "CRL path validation
8688 error" in this case. Applications wanting additional details can use
8689 the verify callback and check the new "parent" field. If this is not
8690 NULL CRL path validation is taking place. Existing applications won't
8691 see this because it requires extended CRL support which is off by
8694 This work was sponsored by Google.
8698 * Support for freshest CRL extension.
8700 This work was sponsored by Google.
8704 * Initial indirect CRL support. Currently only supported in the CRLs
8705 passed directly and not via lookup. Process certificate issuer
8706 CRL entry extension and lookup CRL entries by bother issuer name
8707 and serial number. Check and process CRL issuer entry in IDP extension.
8709 This work was sponsored by Google.
8713 * Add support for distinct certificate and CRL paths. The CRL issuer
8714 certificate is validated separately in this case. Only enabled if
8715 an extended CRL support flag is set: this flag will enable additional
8716 CRL functionality in future.
8718 This work was sponsored by Google.
8722 * Add support for policy mappings extension.
8724 This work was sponsored by Google.
8728 * Fixes to pathlength constraint, self issued certificate handling,
8729 policy processing to align with RFC3280 and PKITS tests.
8731 This work was sponsored by Google.
8735 * Support for name constraints certificate extension. DN, email, DNS
8736 and URI types are currently supported.
8738 This work was sponsored by Google.
8742 * To cater for systems that provide a pointer-based thread ID rather
8743 than numeric, deprecate the current numeric thread ID mechanism and
8744 replace it with a structure and associated callback type. This
8745 mechanism allows a numeric "hash" to be extracted from a thread ID in
8746 either case, and on platforms where pointers are larger than 'long',
8747 mixing is done to help ensure the numeric 'hash' is usable even if it
8748 can't be guaranteed unique. The default mechanism is to use "&errno"
8749 as a pointer-based thread ID to distinguish between threads.
8751 Applications that want to provide their own thread IDs should now use
8752 CRYPTO_THREADID_set_callback() to register a callback that will call
8753 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8755 Note that ERR_remove_state() is now deprecated, because it is tied
8756 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8757 to free the current thread's error state should be replaced by
8758 ERR_remove_thread_state(NULL).
8760 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8761 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8762 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8763 application was previously providing a numeric thread callback that
8764 was inappropriate for distinguishing threads, then uniqueness might
8765 have been obtained with &errno that happened immediately in the
8766 intermediate development versions of OpenSSL; this is no longer the
8767 case, the numeric thread callback will now override the automatic use
8770 *Geoff Thorpe, with help from Bodo Moeller*
8772 * Initial support for different CRL issuing certificates. This covers a
8773 simple case where the self issued certificates in the chain exist and
8774 the real CRL issuer is higher in the existing chain.
8776 This work was sponsored by Google.
8780 * Removed effectively defunct crypto/store from the build.
8784 * Revamp of STACK to provide stronger type-checking. Still to come:
8785 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8786 ASN1_STRING, CONF_VALUE.
8790 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8791 RAM on SSL connections. This option can save about 34k per idle SSL.
8795 * Revamp of LHASH to provide stronger type-checking. Still to come:
8796 STACK, TXT_DB, bsearch, qsort.
8800 * Initial support for Cryptographic Message Syntax (aka CMS) based
8801 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8802 support for data, signedData, compressedData, digestedData and
8803 encryptedData, envelopedData types included. Scripts to check against
8804 RFC4134 examples draft and interop and consistency checks of many
8805 content types and variants.
8809 * Add options to enc utility to support use of zlib compression BIO.
8813 * Extend mk1mf to support importing of options and assembly language
8814 files from Configure script, currently only included in VC-WIN32.
8815 The assembly language rules can now optionally generate the source
8816 files from the associated perl scripts.
8820 * Implement remaining functionality needed to support GOST ciphersuites.
8821 Interop testing has been performed using CryptoPro implementations.
8823 *Victor B. Wagner <vitus@cryptocom.ru>*
8825 * s390x assembler pack.
8829 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8834 * Implement Opaque PRF Input TLS extension as specified in
8835 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8836 official specification yet and no extension type assignment by
8837 IANA exists, this extension (for now) will have to be explicitly
8838 enabled when building OpenSSL by providing the extension number
8839 to use. For example, specify an option
8841 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8843 to the "config" or "Configure" script to enable the extension,
8844 assuming extension number 0x9527 (which is a completely arbitrary
8845 and unofficial assignment based on the MD5 hash of the Internet
8846 Draft). Note that by doing so, you potentially lose
8847 interoperability with other TLS implementations since these might
8848 be using the same extension number for other purposes.
8850 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8851 opaque PRF input value to use in the handshake. This will create
8852 an internal copy of the length-'len' string at 'src', and will
8853 return non-zero for success.
8855 To get more control and flexibility, provide a callback function
8858 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8859 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8863 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8866 Callback function 'cb' will be called in handshakes, and is
8867 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8868 Argument 'arg' is for application purposes (the value as given to
8869 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8870 be provided to the callback function). The callback function
8871 has to return non-zero to report success: usually 1 to use opaque
8872 PRF input just if possible, or 2 to enforce use of the opaque PRF
8873 input. In the latter case, the library will abort the handshake
8874 if opaque PRF input is not successfully negotiated.
8876 Arguments 'peerinput' and 'len' given to the callback function
8877 will always be NULL and 0 in the case of a client. A server will
8878 see the client's opaque PRF input through these variables if
8879 available (NULL and 0 otherwise). Note that if the server
8880 provides an opaque PRF input, the length must be the same as the
8881 length of the client's opaque PRF input.
8883 Note that the callback function will only be called when creating
8884 a new session (session resumption can resume whatever was
8885 previously negotiated), and will not be called in SSL 2.0
8886 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8887 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8888 for applications that need to enforce opaque PRF input.
8892 * Update ssl code to support digests other than SHA1+MD5 for handshake
8895 *Victor B. Wagner <vitus@cryptocom.ru>*
8897 * Add RFC4507 support to OpenSSL. This includes the corrections in
8898 RFC4507bis. The encrypted ticket format is an encrypted encoded
8899 SSL_SESSION structure, that way new session features are automatically
8902 If a client application caches session in an SSL_SESSION structure
8903 support is transparent because tickets are now stored in the encoded
8906 The SSL_CTX structure automatically generates keys for ticket
8907 protection in servers so again support should be possible
8908 with no application modification.
8910 If a client or server wishes to disable RFC4507 support then the option
8911 SSL_OP_NO_TICKET can be set.
8913 Add a TLS extension debugging callback to allow the contents of any client
8914 or server extensions to be examined.
8916 This work was sponsored by Google.
8920 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8921 OpenSSL should now compile cleanly on gcc 4.2
8923 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8925 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8926 support including streaming MAC support: this is required for GOST
8927 ciphersuite support.
8929 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8931 * Add option -stream to use PKCS#7 streaming in smime utility. New
8932 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8933 to output in BER and PEM format.
8937 * Experimental support for use of HMAC via EVP_PKEY interface. This
8938 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8939 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8940 ENGINE support for HMAC keys which are unextractable. New -mac and
8941 -macopt options to dgst utility.
8945 * New option -sigopt to dgst utility. Update dgst to use
8946 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8947 alternative signing parameters such as X9.31 or PSS in the dgst
8952 * Change ssl_cipher_apply_rule(), the internal function that does
8953 the work each time a ciphersuite string requests enabling
8954 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8955 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8956 the order of disabled ciphersuites such that those ciphersuites
8957 that most recently went from enabled to disabled not only stay
8958 in order with respect to each other, but also have higher priority
8959 than other disabled ciphersuites the next time ciphersuites are
8962 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8963 the same ciphersuites as with "HIGH" alone, but in a specific
8964 order where the PSK ciphersuites come first (since they are the
8965 most recently disabled ciphersuites when "HIGH" is parsed).
8967 Also, change ssl_create_cipher_list() (using this new
8968 functionality) such that between otherwise identical
8969 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8974 * Change ssl_create_cipher_list() so that it automatically
8975 arranges the ciphersuites in reasonable order before starting
8976 to process the rule string. Thus, the definition for "DEFAULT"
8977 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8978 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8979 This makes it much easier to arrive at a reasonable default order
8980 in applications for which anonymous ciphers are OK (meaning
8981 that you can't actually use DEFAULT).
8983 *Bodo Moeller; suggested by Victor Duchovni*
8985 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8986 processing) into multiple integers instead of setting
8987 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8988 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8989 (These masks as well as the individual bit definitions are hidden
8990 away into the non-exported interface ssl/ssl_locl.h, so this
8991 change to the definition of the SSL_CIPHER structure shouldn't
8992 affect applications.) This give us more bits for each of these
8993 categories, so there is no longer a need to coagulate AES128 and
8994 AES256 into a single algorithm bit, and to coagulate Camellia128
8995 and Camellia256 into a single algorithm bit, which has led to all
8998 Thus, among other things, the kludge introduced in 0.9.7m and
8999 0.9.8e for masking out AES256 independently of AES128 or masking
9000 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9002 With the change, we also introduce new ciphersuite aliases that
9003 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9008 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9009 Use the leftmost N bytes of the signature input if the input is
9010 larger than the prime q (with N being the size in bytes of q).
9014 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9015 it yet and it is largely untested.
9019 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9023 * Initial incomplete changes to avoid need for function casts in OpenSSL
9024 some compilers (gcc 4.2 and later) reject their use. Safestack is
9025 reimplemented. Update ASN1 to avoid use of legacy functions.
9029 * Win32/64 targets are linked with Winsock2.
9033 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9034 to external functions. This can be used to increase CRL handling
9035 efficiency especially when CRLs are very large by (for example) storing
9036 the CRL revoked certificates in a database.
9040 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9041 new CRLs added to a directory can be used. New command line option
9042 -verify_return_error to s_client and s_server. This causes real errors
9043 to be returned by the verify callback instead of carrying on no matter
9044 what. This reflects the way a "real world" verify callback would behave.
9048 * GOST engine, supporting several GOST algorithms and public key formats.
9049 Kindly donated by Cryptocom.
9053 * Partial support for Issuing Distribution Point CRL extension. CRLs
9054 partitioned by DP are handled but no indirect CRL or reason partitioning
9055 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9056 selected via a scoring technique which handles IDP and AKID in CRLs.
9060 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9061 will ultimately be used for all verify operations: this will remove the
9062 X509_STORE dependency on certificate verification and allow alternative
9063 lookup methods. X509_STORE based implementations of these two callbacks.
9067 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9068 Modify get_crl() to find a valid (unexpired) CRL if possible.
9072 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9073 this would be called X509_CRL_cmp() but that name is already used by
9074 a function that just compares CRL issuer names. Cache several CRL
9075 extensions in X509_CRL structure and cache CRLDP in X509.
9079 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9080 this maps equivalent X509_NAME structures into a consistent structure.
9081 Name comparison can then be performed rapidly using memcmp().
9085 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9090 * Allow digests to supply their own micalg string for S/MIME type using
9091 the ctrl EVP_MD_CTRL_MICALG.
9095 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9096 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9097 ctrl. It can then customise the structure before and/or after signing
9102 * New function OBJ_add_sigid() to allow application defined signature OIDs
9103 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9104 to free up any added signature OIDs.
9108 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9109 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9110 digest and cipher tables. New options added to openssl utility:
9111 list-message-digest-algorithms and list-cipher-algorithms.
9115 * Change the array representation of binary polynomials: the list
9116 of degrees of non-zero coefficients is now terminated with -1.
9117 Previously it was terminated with 0, which was also part of the
9118 value; thus, the array representation was not applicable to
9119 polynomials where t^0 has coefficient zero. This change makes
9120 the array representation useful in a more general context.
9124 * Various modifications and fixes to SSL/TLS cipher string
9125 handling. For ECC, the code now distinguishes between fixed ECDH
9126 with RSA certificates on the one hand and with ECDSA certificates
9127 on the other hand, since these are separate ciphersuites. The
9128 unused code for Fortezza ciphersuites has been removed.
9130 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9131 (not "ECDHE"). For consistency with the code for DH
9132 certificates, use of ECDH certificates is now considered ECDH
9133 authentication, not RSA or ECDSA authentication (the latter is
9134 merely the CA's signing algorithm and not actively used in the
9137 The temporary ciphersuite alias "ECCdraft" is no longer
9138 available, and ECC ciphersuites are no longer excluded from "ALL"
9139 and "DEFAULT". The following aliases now exist for RFC 4492
9140 ciphersuites, most of these by analogy with the DH case:
9142 kECDHr - ECDH cert, signed with RSA
9143 kECDHe - ECDH cert, signed with ECDSA
9144 kECDH - ECDH cert (signed with either RSA or ECDSA)
9145 kEECDH - ephemeral ECDH
9146 ECDH - ECDH cert or ephemeral ECDH
9152 AECDH - anonymous ECDH
9153 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9157 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9158 Use correct micalg parameters depending on digest(s) in signed message.
9162 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9163 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9167 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9168 an engine to register a method. Add ENGINE lookups for methods and
9169 functional reference processing.
9173 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9174 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9179 * New -resign option to smime utility. This adds one or more signers
9180 to an existing PKCS#7 signedData structure. Also -md option to use an
9181 alternative message digest algorithm for signing.
9185 * Tidy up PKCS#7 routines and add new functions to make it easier to
9186 create PKCS7 structures containing multiple signers. Update smime
9187 application to support multiple signers.
9191 * New -macalg option to pkcs12 utility to allow setting of an alternative
9196 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9197 Reorganize PBE internals to lookup from a static table using NIDs,
9198 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9199 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9200 PRF which will be automatically used with PBES2.
9204 * Replace the algorithm specific calls to generate keys in "req" with the
9209 * Update PKCS#7 enveloped data routines to use new API. This is now
9210 supported by any public key method supporting the encrypt operation. A
9211 ctrl is added to allow the public key algorithm to examine or modify
9212 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9217 * Add a ctrl to asn1 method to allow a public key algorithm to express
9218 a default digest type to use. In most cases this will be SHA1 but some
9219 algorithms (such as GOST) need to specify an alternative digest. The
9220 return value indicates how strong the preference is 1 means optional and
9221 2 is mandatory (that is it is the only supported type). Modify
9222 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9223 use the default md. Update openssl utilities to use the default digest
9224 type for signing if it is not explicitly indicated.
9228 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9229 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9230 signing method from the key type. This effectively removes the link
9231 between digests and public key types.
9235 * Add an OID cross reference table and utility functions. Its purpose is to
9236 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9237 rsaEncryption. This will allow some of the algorithm specific hackery
9238 needed to use the correct OID to be removed.
9242 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9243 structures for PKCS7_sign(). They are now set up by the relevant public
9248 * Add provisional EC pkey method with support for ECDSA and ECDH.
9252 * Add support for key derivation (agreement) in the API, DH method and
9257 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9258 public and private key formats. As a side effect these add additional
9259 command line functionality not previously available: DSA signatures can be
9260 generated and verified using pkeyutl and DH key support and generation in
9267 *Oliver Tappe <zooey@hirschkaefer.de>*
9269 * New make target "install_html_docs" installs HTML renditions of the
9272 *Oliver Tappe <zooey@hirschkaefer.de>*
9274 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9275 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9276 support key and parameter generation and add initial key generation
9277 functionality for RSA.
9281 * Add functions for main EVP_PKEY_method operations. The undocumented
9282 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9283 `EVP_PKEY_{encrypt,decrypt}_old`.
9287 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9288 key API, doesn't do much yet.
9292 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9293 public key algorithms. New option to openssl utility:
9294 "list-public-key-algorithms" to print out info.
9298 * Implement the Supported Elliptic Curves Extension for
9299 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9303 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9304 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9308 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9309 utilities such as rsa, dsa, dsaparam etc except they process any key
9314 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9315 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9316 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9321 * Initial support for pluggable public key ASN1.
9322 De-spaghettify the public key ASN1 handling. Move public and private
9323 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9324 algorithm specific handling to a single module within the relevant
9325 algorithm directory. Add functions to allow (near) opaque processing
9326 of public and private key structures.
9330 * Implement the Supported Point Formats Extension for
9331 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9335 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9336 for the psk identity [hint] and the psk callback functions to the
9337 SSL_SESSION, SSL and SSL_CTX structure.
9340 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9344 SSL_CTX_use_psk_identity_hint
9345 SSL_get_psk_identity_hint
9346 SSL_get_psk_identity
9347 SSL_use_psk_identity_hint
9349 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9351 * Add RFC 3161 compliant time stamp request creation, response generation
9352 and response verification functionality.
9354 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9356 * Add initial support for TLS extensions, specifically for the server_name
9357 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9358 have new members for a host name. The SSL data structure has an
9359 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9360 stored in that context to allow for session resumption, even after the
9361 SSL has been switched to a new SSL_CTX in reaction to a client's
9362 server_name extension.
9364 New functions (subject to change):
9366 SSL_get_servername()
9367 SSL_get_servername_type()
9370 New CTRL codes and macros (subject to change):
9372 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9373 - SSL_CTX_set_tlsext_servername_callback()
9374 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9375 - SSL_CTX_set_tlsext_servername_arg()
9376 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9378 openssl s_client has a new '-servername ...' option.
9380 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9381 '-key2 ...', '-servername_fatal' (subject to change). This allows
9382 testing the HostName extension for a specific single host name ('-cert'
9383 and '-key' remain fallbacks for handshakes without HostName
9384 negotiation). If the unrecognized_name alert has to be sent, this by
9385 default is a warning; it becomes fatal with the '-servername_fatal'
9388 *Peter Sylvester, Remy Allais, Christophe Renou*
9390 * Whirlpool hash implementation is added.
9394 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9395 bn(64,32). Because of instruction set limitations it doesn't have
9396 any negative impact on performance. This was done mostly in order
9397 to make it possible to share assembler modules, such as bn_mul_mont
9398 implementations, between 32- and 64-bit builds without hassle.
9402 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9403 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9408 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9409 dedicated Montgomery multiplication procedure, is introduced.
9410 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9411 "64-bit" performance on certain 32-bit targets.
9415 * New option SSL_OP_NO_COMP to disable use of compression selectively
9416 in SSL structures. New SSL ctrl to set maximum send fragment size.
9417 Save memory by setting the I/O buffer sizes dynamically instead of
9418 using the maximum available value.
9422 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9423 in addition to the text details.
9427 * Very, very preliminary EXPERIMENTAL support for printing of general
9428 ASN1 structures. This currently produces rather ugly output and doesn't
9429 handle several customised structures at all.
9433 * Integrated support for PVK file format and some related formats such
9434 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9435 these in the 'rsa' and 'dsa' utilities.
9439 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9443 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9444 place for the (very old) "NETSCAPE" format certificates which are now
9445 handled using new ASN1 code equivalents.
9449 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9450 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9451 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9455 * Modify CRL distribution points extension code to print out previously
9456 unsupported fields. Enhance extension setting code to allow setting of
9461 * Add print and set support for Issuing Distribution Point CRL extension.
9465 * Change 'Configure' script to enable Camellia by default.
9472 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9474 * When rejecting SSL/TLS records due to an incorrect version number, never
9475 update s->server with a new major version number. As of
9476 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9477 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9478 the previous behavior could result in a read attempt at NULL when
9479 receiving specific incorrect SSL/TLS records once record payload
9480 protection is active. ([CVE-2010-0740])
9482 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9484 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9485 could be crashed if the relevant tables were not present (e.g. chrooted).
9487 *Tomas Hoger <thoger@redhat.com>*
9489 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9491 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9493 *Martin Olsson, Neel Mehta*
9495 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9496 accommodate for stack sorting, always a write lock!).
9500 * On some versions of WIN32 Heap32Next is very slow. This can cause
9501 excessive delays in the RAND_poll(): over a minute. As a workaround
9502 include a time check in the inner Heap32Next loop too.
9506 * The code that handled flushing of data in SSL/TLS originally used the
9507 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9508 the problem outlined in PR#1949. The fix suggested there however can
9509 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9510 of Apache). So instead simplify the code to flush unconditionally.
9511 This should be fine since flushing with no data to flush is a no op.
9515 * Handle TLS versions 2.0 and later properly and correctly use the
9516 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9517 off ancient servers have a habit of sticking around for a while...
9521 * Modify compression code so it frees up structures without using the
9522 ex_data callbacks. This works around a problem where some applications
9523 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9524 restarting) then use compression (e.g. SSL with compression) later.
9525 This results in significant per-connection memory leaks and
9526 has caused some security issues including CVE-2008-1678 and
9531 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9532 change when encrypting or decrypting.
9536 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9537 connect and renegotiate with servers which do not support RI.
9538 Until RI is more widely deployed this option is enabled by default.
9542 * Add "missing" ssl ctrls to clear options and mode.
9546 * If client attempts to renegotiate and doesn't support RI respond with
9547 a no_renegotiation alert as required by RFC5746. Some renegotiating
9548 TLS clients will continue a connection gracefully when they receive
9549 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9550 waiting for a server hello which it will never receive. Now we treat a
9551 received no_renegotiation alert as a fatal error. This is because
9552 applications requesting a renegotiation might well expect it to succeed
9553 and would have no code in place to handle the server denying it so the
9554 only safe thing to do is to terminate the connection.
9558 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9559 peer supports secure renegotiation and 0 otherwise. Print out peer
9560 renegotiation support in s_client/s_server.
9564 * Replace the highly broken and deprecated SPKAC certification method with
9565 the updated NID creation version. This should correctly handle UTF8.
9569 * Implement RFC5746. Re-enable renegotiation but require the extension
9570 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9571 turns out to be a bad idea. It has been replaced by
9572 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9573 SSL_CTX_set_options(). This is really not recommended unless you
9574 know what you are doing.
9576 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9578 * Fixes to stateless session resumption handling. Use initial_ctx when
9579 issuing and attempting to decrypt tickets in case it has changed during
9580 servername handling. Use a non-zero length session ID when attempting
9581 stateless session resumption: this makes it possible to determine if
9582 a resumption has occurred immediately after receiving server hello
9583 (several places in OpenSSL subtly assume this) instead of later in
9588 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9589 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9590 fixes for a few places where the return code is not checked
9593 *Julia Lawall <julia@diku.dk>*
9595 * Add --strict-warnings option to Configure script to include devteam
9596 warnings in other configurations.
9600 * Add support for --libdir option and LIBDIR variable in makefiles. This
9601 makes it possible to install openssl libraries in locations which
9602 have names other than "lib", for example "/usr/lib64" which some
9605 *Steve Henson, based on patch from Jeremy Utley*
9607 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9608 X690 8.9.12 and can produce some misleading textual output of OIDs.
9610 *Steve Henson, reported by Dan Kaminsky*
9612 * Delete MD2 from algorithm tables. This follows the recommendation in
9613 several standards that it is not used in new applications due to
9614 several cryptographic weaknesses. For binary compatibility reasons
9615 the MD2 API is still compiled in by default.
9619 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9624 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9625 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9628 *Guenter <lists@gknw.net>*
9630 * Fix the server certificate chain building code to use X509_verify_cert(),
9631 it used to have an ad-hoc builder which was unable to cope with anything
9632 other than a simple chain.
9634 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9636 * Don't check self signed certificate signatures in X509_verify_cert()
9637 by default (a flag can override this): it just wastes time without
9638 adding any security. As a useful side effect self signed root CAs
9639 with non-FIPS digests are now usable in FIPS mode.
9643 * In dtls1_process_out_of_seq_message() the check if the current message
9644 is already buffered was missing. For every new message was memory
9645 allocated, allowing an attacker to perform an denial of service attack
9646 with sending out of seq handshake messages until there is no memory
9647 left. Additionally every future message was buffered, even if the
9648 sequence number made no sense and would be part of another handshake.
9649 So only messages with sequence numbers less than 10 in advance will be
9650 buffered. ([CVE-2009-1378])
9652 *Robin Seggelmann, discovered by Daniel Mentz*
9654 * Records are buffered if they arrive with a future epoch to be
9655 processed after finishing the corresponding handshake. There is
9656 currently no limitation to this buffer allowing an attacker to perform
9657 a DOS attack with sending records with future epochs until there is no
9658 memory left. This patch adds the pqueue_size() function to determine
9659 the size of a buffer and limits the record buffer to 100 entries.
9662 *Robin Seggelmann, discovered by Daniel Mentz*
9664 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9665 parent structure is freed. ([CVE-2009-1379])
9669 * Handle non-blocking I/O properly in SSL_shutdown() call.
9671 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9673 * Add `2.5.4.*` OIDs
9675 *Ilya O. <vrghost@gmail.com>*
9677 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9679 * Disable renegotiation completely - this fixes a severe security
9680 problem ([CVE-2009-3555]) at the cost of breaking all
9681 renegotiation. Renegotiation can be re-enabled by setting
9682 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9683 run-time. This is really not recommended unless you know what
9688 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9690 * Don't set val to NULL when freeing up structures, it is freed up by
9691 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9692 zeroing past the valid field. ([CVE-2009-0789])
9694 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9696 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9697 checked correctly. This would allow some invalid signed attributes to
9698 appear to verify correctly. ([CVE-2009-0591])
9700 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9702 * Reject UniversalString and BMPString types with invalid lengths. This
9703 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9704 a legal length. ([CVE-2009-0590])
9708 * Set S/MIME signing as the default purpose rather than setting it
9709 unconditionally. This allows applications to override it at the store
9714 * Permit restricted recursion of ASN1 strings. This is needed in practice
9715 to handle some structures.
9719 * Improve efficiency of mem_gets: don't search whole buffer each time
9722 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9724 * New -hex option for openssl rand.
9728 * Print out UTF8String and NumericString when parsing ASN1.
9732 * Support NumericString type for name components.
9736 * Allow CC in the environment to override the automatically chosen
9737 compiler. Note that nothing is done to ensure flags work with the
9742 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9744 * Properly check EVP_VerifyFinal() and similar return values
9747 *Ben Laurie, Bodo Moeller, Google Security Team*
9749 * Enable TLS extensions by default.
9753 * Allow the CHIL engine to be loaded, whether the application is
9754 multithreaded or not. (This does not release the developer from the
9755 obligation to set up the dynamic locking callbacks.)
9757 *Sander Temme <sander@temme.net>*
9759 * Use correct exit code if there is an error in dgst command.
9761 *Steve Henson; problem pointed out by Roland Dirlewanger*
9763 * Tweak Configure so that you need to say "experimental-jpake" to enable
9764 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9768 * Add experimental JPAKE support, including demo authentication in
9769 s_client and s_server.
9773 * Set the comparison function in v3_addr_canonize().
9775 *Rob Austein <sra@hactrn.net>*
9777 * Add support for XMPP STARTTLS in s_client.
9779 *Philip Paeps <philip@freebsd.org>*
9781 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9782 to ensure that even with this option, only ciphersuites in the
9783 server's preference list will be accepted. (Note that the option
9784 applies only when resuming a session, so the earlier behavior was
9785 just about the algorithm choice for symmetric cryptography.)
9789 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9791 * Fix NULL pointer dereference if a DTLS server received
9792 ChangeCipherSpec as first record ([CVE-2009-1386]).
9796 * Fix a state transition in s3_srvr.c and d1_srvr.c
9797 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9801 * The fix in 0.9.8c that supposedly got rid of unsafe
9802 double-checked locking was incomplete for RSA blinding,
9803 addressing just one layer of what turns out to have been
9804 doubly unsafe triple-checked locking.
9806 So now fix this for real by retiring the MONT_HELPER macro
9807 in crypto/rsa/rsa_eay.c.
9809 *Bodo Moeller; problem pointed out by Marius Schilder*
9811 * Various precautionary measures:
9813 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9815 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9816 (NB: This would require knowledge of the secret session ticket key
9817 to exploit, in which case you'd be SOL either way.)
9819 - Change bn_nist.c so that it will properly handle input BIGNUMs
9820 outside the expected range.
9822 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9825 *Neel Mehta, Bodo Moeller*
9827 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9828 the load fails. Useful for distros.
9830 *Ben Laurie and the FreeBSD team*
9832 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9836 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9840 * Expand ENGINE to support engine supplied SSL client certificate functions.
9842 This work was sponsored by Logica.
9846 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9847 keystores. Support for SSL/TLS client authentication too.
9848 Not compiled unless enable-capieng specified to Configure.
9850 This work was sponsored by Logica.
9854 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9855 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9856 attribute creation routines such as certificate requests and PKCS#12
9861 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9863 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9864 handshake which could lead to a client crash as found using the
9865 Codenomicon TLS test suite ([CVE-2008-1672])
9867 *Steve Henson, Mark Cox*
9869 * Fix double free in TLS server name extensions which could lead to
9870 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9874 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9876 Clear the error queue to ensure that error entries left from
9877 older function calls do not interfere with the correct operation.
9879 *Lutz Jaenicke, Erik de Castro Lopo*
9881 * Remove root CA certificates of commercial CAs:
9883 The OpenSSL project does not recommend any specific CA and does not
9884 have any policy with respect to including or excluding any CA.
9885 Therefore it does not make any sense to ship an arbitrary selection
9886 of root CA certificates with the OpenSSL software.
9890 * RSA OAEP patches to fix two separate invalid memory reads.
9891 The first one involves inputs when 'lzero' is greater than
9892 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9893 before the beginning of from). The second one involves inputs where
9894 the 'db' section contains nothing but zeroes (there is a one-byte
9895 invalid read after the end of 'db').
9897 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9899 * Partial backport from 0.9.9-dev:
9901 Introduce bn_mul_mont (dedicated Montgomery multiplication
9902 procedure) as a candidate for BIGNUM assembler implementation.
9903 While 0.9.9-dev uses assembler for various architectures, only
9904 x86_64 is available by default here in the 0.9.8 branch, and
9905 32-bit x86 is available through a compile-time setting.
9907 To try the 32-bit x86 assembler implementation, use Configure
9908 option "enable-montasm" (which exists only for this backport).
9910 As "enable-montasm" for 32-bit x86 disclaims code stability
9911 anyway, in this constellation we activate additional code
9912 backported from 0.9.9-dev for further performance improvements,
9913 namely BN_from_montgomery_word. (To enable this otherwise,
9914 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9916 *Andy Polyakov (backport partially by Bodo Moeller)*
9918 * Add TLS session ticket callback. This allows an application to set
9919 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9920 values. This is useful for key rollover for example where several key
9921 sets may exist with different names.
9925 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9926 This was broken until now in 0.9.8 releases, such that the only way
9927 a registered ENGINE could be used (assuming it initialises
9928 successfully on the host) was to explicitly set it as the default
9929 for the relevant algorithms. This is in contradiction with 0.9.7
9930 behaviour and the documentation. With this fix, when an ENGINE is
9931 registered into a given algorithm's table of implementations, the
9932 'uptodate' flag is reset so that auto-discovery will be used next
9933 time a new context for that algorithm attempts to select an
9936 *Ian Lister (tweaked by Geoff Thorpe)*
9938 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9939 implementation in the following ways:
9941 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9944 Lack of BER streaming support means one pass streaming processing is
9945 only supported if data is detached: setting the streaming flag is
9946 ignored for embedded content.
9948 CMS support is disabled by default and must be explicitly enabled
9949 with the enable-cms configuration option.
9953 * Update the GMP engine glue to do direct copies between BIGNUM and
9954 mpz_t when openssl and GMP use the same limb size. Otherwise the
9955 existing "conversion via a text string export" trick is still used.
9957 *Paul Sheer <paulsheer@gmail.com>*
9959 * Zlib compression BIO. This is a filter BIO which compressed and
9960 uncompresses any data passed through it.
9964 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9965 RFC3394 compatible AES key wrapping.
9969 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9970 sets string data without copying. X509_ALGOR_set0() and
9971 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9972 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9973 from an X509_ATTRIBUTE structure optionally checking it occurs only
9974 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9979 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9980 to get the expected BN_FLG_CONSTTIME behavior.
9982 *Bodo Moeller (Google)*
9986 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9987 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9988 - added some more tests to do_tests.pl
9989 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9990 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9991 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9992 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9993 - various changes to netware.pl to enable gcc-cross builds on Win32
9995 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9996 - various changes to fix missing prototype warnings
9997 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9998 - added AES, WHIRLPOOL and CPUID assembler code to build files
9999 - added missing AES assembler make rules to mk1mf.pl
10000 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10002 *Guenter Knauf <eflash@gmx.net>*
10004 * Implement certificate status request TLS extension defined in RFC3546.
10005 A client can set the appropriate parameters and receive the encoded
10006 OCSP response via a callback. A server can query the supplied parameters
10007 and set the encoded OCSP response in the callback. Add simplified examples
10008 to s_client and s_server.
10012 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10014 * Fix various bugs:
10015 + Binary incompatibility of ssl_ctx_st structure
10016 + DTLS interoperation with non-compliant servers
10017 + Don't call get_session_cb() without proposed session
10018 + Fix ia64 assembler code
10020 *Andy Polyakov, Steve Henson*
10022 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10024 * DTLS Handshake overhaul. There were longstanding issues with
10025 OpenSSL DTLS implementation, which were making it impossible for
10026 RFC 4347 compliant client to communicate with OpenSSL server.
10027 Unfortunately just fixing these incompatibilities would "cut off"
10028 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10029 server keeps tolerating non RFC compliant syntax. The opposite is
10030 not true, 0.9.8f client can not communicate with earlier server.
10031 This update even addresses CVE-2007-4995.
10035 * Changes to avoid need for function casts in OpenSSL: some compilers
10036 (gcc 4.2 and later) reject their use.
10037 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10040 * Add RFC4507 support to OpenSSL. This includes the corrections in
10041 RFC4507bis. The encrypted ticket format is an encrypted encoded
10042 SSL_SESSION structure, that way new session features are automatically
10045 If a client application caches session in an SSL_SESSION structure
10046 support is transparent because tickets are now stored in the encoded
10049 The SSL_CTX structure automatically generates keys for ticket
10050 protection in servers so again support should be possible
10051 with no application modification.
10053 If a client or server wishes to disable RFC4507 support then the option
10054 SSL_OP_NO_TICKET can be set.
10056 Add a TLS extension debugging callback to allow the contents of any client
10057 or server extensions to be examined.
10059 This work was sponsored by Google.
10063 * Add initial support for TLS extensions, specifically for the server_name
10064 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10065 have new members for a host name. The SSL data structure has an
10066 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10067 stored in that context to allow for session resumption, even after the
10068 SSL has been switched to a new SSL_CTX in reaction to a client's
10069 server_name extension.
10071 New functions (subject to change):
10073 SSL_get_servername()
10074 SSL_get_servername_type()
10077 New CTRL codes and macros (subject to change):
10079 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10080 - SSL_CTX_set_tlsext_servername_callback()
10081 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10082 - SSL_CTX_set_tlsext_servername_arg()
10083 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10085 openssl s_client has a new '-servername ...' option.
10087 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10088 '-key2 ...', '-servername_fatal' (subject to change). This allows
10089 testing the HostName extension for a specific single host name ('-cert'
10090 and '-key' remain fallbacks for handshakes without HostName
10091 negotiation). If the unrecognized_name alert has to be sent, this by
10092 default is a warning; it becomes fatal with the '-servername_fatal'
10095 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10097 * Add AES and SSE2 assembly language support to VC++ build.
10101 * Mitigate attack on final subtraction in Montgomery reduction.
10105 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10106 (which previously caused an internal error).
10110 * Squeeze another 10% out of IGE mode when in != out.
10114 * AES IGE mode speedup.
10116 *Dean Gaudet (Google)*
10118 * Add the Korean symmetric 128-bit cipher SEED (see
10119 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10120 add SEED ciphersuites from RFC 4162:
10122 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10123 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10124 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10125 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10127 To minimize changes between patchlevels in the OpenSSL 0.9.8
10128 series, SEED remains excluded from compilation unless OpenSSL
10129 is configured with 'enable-seed'.
10131 *KISA, Bodo Moeller*
10133 * Mitigate branch prediction attacks, which can be practical if a
10134 single processor is shared, allowing a spy process to extract
10135 information. For detailed background information, see
10136 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10137 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10138 and Necessary Software Countermeasures"). The core of the change
10139 are new versions BN_div_no_branch() and
10140 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10141 respectively, which are slower, but avoid the security-relevant
10142 conditional branches. These are automatically called by BN_div()
10143 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10144 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10145 remove a conditional branch.
10147 BN_FLG_CONSTTIME is the new name for the previous
10148 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10149 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10150 in the exponent causes BN_mod_exp_mont() to use the alternative
10151 implementation in BN_mod_exp_mont_consttime().) The old name
10152 remains as a deprecated alias.
10154 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10155 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10156 constant-time implementations for more than just exponentiation.
10157 Here too the old name is kept as a deprecated alias.
10159 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10160 the BN_BLINDING structure gets an independent copy of the
10161 modulus. This means that the previous `BIGNUM *m` argument to
10162 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10163 essentially becomes `const BIGNUM *m`, although we can't actually
10164 change this in the header file before 0.9.9. It allows
10165 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10166 enable BN_FLG_CONSTTIME.
10168 *Matthew D Wood (Intel Corp)*
10170 * In the SSL/TLS server implementation, be strict about session ID
10171 context matching (which matters if an application uses a single
10172 external cache for different purposes). Previously,
10173 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10174 set. This did ensure strict client verification, but meant that,
10175 with applications using a single external cache for quite
10176 different requirements, clients could circumvent ciphersuite
10177 restrictions for a given session ID context by starting a session
10178 in a different context.
10182 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10183 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10184 authentication-only ciphersuites.
10188 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10189 not complete and could lead to a possible single byte overflow
10190 ([CVE-2007-5135]) [Ben Laurie]
10192 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10194 * Since AES128 and AES256 (and similarly Camellia128 and
10195 Camellia256) share a single mask bit in the logic of
10196 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10197 kludge to work properly if AES128 is available and AES256 isn't
10198 (or if Camellia128 is available and Camellia256 isn't).
10202 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10203 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10204 When a point or a seed is encoded in a BIT STRING, we need to
10205 prevent the removal of trailing zero bits to get the proper DER
10206 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10207 of a NamedBitList, for which trailing 0 bits need to be removed.)
10211 * Have SSL/TLS server implementation tolerate "mismatched" record
10212 protocol version while receiving ClientHello even if the
10213 ClientHello is fragmented. (The server can't insist on the
10214 particular protocol version it has chosen before the ServerHello
10215 message has informed the client about his choice.)
10219 * Add RFC 3779 support.
10221 *Rob Austein for ARIN, Ben Laurie*
10223 * Load error codes if they are not already present instead of using a
10224 static variable. This allows them to be cleanly unloaded and reloaded.
10225 Improve header file function name parsing.
10229 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10230 or CAPABILITY handshake as required by RFCs.
10232 *Goetz Babin-Ebell*
10234 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10236 * Introduce limits to prevent malicious keys being able to
10237 cause a denial of service. ([CVE-2006-2940])
10239 *Steve Henson, Bodo Moeller*
10241 * Fix ASN.1 parsing of certain invalid structures that can result
10242 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10244 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10245 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10247 * Fix SSL client code which could crash if connecting to a
10248 malicious SSLv2 server. ([CVE-2006-4343])
10250 *Tavis Ormandy and Will Drewry, Google Security Team*
10252 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10253 match only those. Before that, "AES256-SHA" would be interpreted
10254 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10255 the same strength classification in 0.9.7h) as we currently only
10256 have a single AES bit in the ciphersuite description bitmap.
10257 That change, however, also applied to ciphersuite strings such as
10258 "RC4-MD5" that intentionally matched multiple ciphersuites --
10259 namely, SSL 2.0 ciphersuites in addition to the more common ones
10260 from SSL 3.0/TLS 1.0.
10262 So we change the selection algorithm again: Naming an explicit
10263 ciphersuite selects this one ciphersuite, and any other similar
10264 ciphersuite (same bitmap) from *other* protocol versions.
10265 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10266 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10268 Since SSL 2.0 does not have any ciphersuites for which the
10269 128/256 bit distinction would be relevant, this works for now.
10270 The proper fix will be to use different bits for AES128 and
10271 AES256, which would have avoided the problems from the beginning;
10272 however, bits are scarce, so we can only do this in a new release
10273 (not just a patchlevel) when we can change the SSL_CIPHER
10274 definition to split the single 'unsigned long mask' bitmap into
10275 multiple values to extend the available space.
10279 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10281 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10282 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10284 * Add AES IGE and biIGE modes.
10288 * Change the Unix randomness entropy gathering to use poll() when
10289 possible instead of select(), since the latter has some
10290 undesirable limitations.
10292 *Darryl Miles via Richard Levitte and Bodo Moeller*
10294 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10295 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10296 cannot be implicitly activated as part of, e.g., the "AES" alias.
10297 However, please upgrade to OpenSSL 0.9.9[-dev] for
10298 non-experimental use of the ECC ciphersuites to get TLS extension
10299 support, which is required for curve and point format negotiation
10300 to avoid potential handshake problems.
10304 * Disable rogue ciphersuites:
10306 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10307 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10308 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10310 The latter two were purportedly from
10311 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10314 Also deactivate the remaining ciphersuites from
10315 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10316 unofficial, and the ID has long expired.
10320 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10321 dual-core machines) and other potential thread-safety issues.
10325 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10326 versions), which is now available for royalty-free use
10327 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10328 Also, add Camellia TLS ciphersuites from RFC 4132.
10330 To minimize changes between patchlevels in the OpenSSL 0.9.8
10331 series, Camellia remains excluded from compilation unless OpenSSL
10332 is configured with 'enable-camellia'.
10336 * Disable the padding bug check when compression is in use. The padding
10337 bug check assumes the first packet is of even length, this is not
10338 necessarily true if compression is enabled and can result in false
10339 positives causing handshake failure. The actual bug test is ancient
10340 code so it is hoped that implementations will either have fixed it by
10341 now or any which still have the bug do not support compression.
10345 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10347 * When applying a cipher rule check to see if string match is an explicit
10348 cipher suite and only match that one cipher suite if it is.
10352 * Link in manifests for VC++ if needed.
10354 *Austin Ziegler <halostatue@gmail.com>*
10356 * Update support for ECC-based TLS ciphersuites according to
10357 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10358 TLS extensions, which are supported starting with the 0.9.9
10359 branch, not in the OpenSSL 0.9.8 branch).
10363 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10364 opaque EVP_CIPHER_CTX handling.
10368 * Fixes and enhancements to zlib compression code. We now only use
10369 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10370 to conform with the standards mentioned here:
10371 <http://www.zlib.net/DLL_FAQ.txt>
10372 Static zlib linking now works on Windows and the new --with-zlib-include
10373 --with-zlib-lib options to Configure can be used to supply the location
10374 of the headers and library. Gracefully handle case where zlib library
10379 * Several fixes and enhancements to the OID generation code. The old code
10380 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10381 handle numbers larger than ULONG_MAX, truncated printing and had a
10382 non standard OBJ_obj2txt() behaviour.
10386 * Add support for building of engines under engine/ as shared libraries
10387 under VC++ build system.
10391 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10392 Hopefully, we will not see any false combination of paths any more.
10396 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10398 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10399 (part of SSL_OP_ALL). This option used to disable the
10400 countermeasure against man-in-the-middle protocol-version
10401 rollback in the SSL 2.0 server implementation, which is a bad
10402 idea. ([CVE-2005-2969])
10404 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10405 for Information Security, National Institute of Advanced Industrial
10406 Science and Technology [AIST], Japan)*
10408 * Add two function to clear and return the verify parameter flags.
10412 * Keep cipherlists sorted in the source instead of sorting them at
10413 runtime, thus removing the need for a lock.
10417 * Avoid some small subgroup attacks in Diffie-Hellman.
10419 *Nick Mathewson and Ben Laurie*
10421 * Add functions for well-known primes.
10425 * Extended Windows CE support.
10427 *Satoshi Nakamura and Andy Polyakov*
10429 * Initialize SSL_METHOD structures at compile time instead of during
10430 runtime, thus removing the need for a lock.
10434 * Make PKCS7_decrypt() work even if no certificate is supplied by
10435 attempting to decrypt each encrypted key in turn. Add support to
10440 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10442 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10445 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10449 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10450 key into the same file any more.
10454 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10458 * Add -utf8 command line and config file option to 'ca'.
10460 *Stefan <stf@udoma.org*
10462 * Removed the macro des_crypt(), as it seems to conflict with some
10463 libraries. Use DES_crypt().
10467 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10468 involves renaming the source and generated shared-libs for
10469 both. The engines will accept the corrected or legacy ids
10470 ('ncipher' and '4758_cca' respectively) when binding. NB,
10471 this only applies when building 'shared'.
10473 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10475 * Add attribute functions to EVP_PKEY structure. Modify
10476 PKCS12_create() to recognize a CSP name attribute and
10477 use it. Make -CSP option work again in pkcs12 utility.
10481 * Add new functionality to the bn blinding code:
10482 - automatic re-creation of the BN_BLINDING parameters after
10483 a fixed number of uses (currently 32)
10484 - add new function for parameter creation
10485 - introduce flags to control the update behaviour of the
10486 BN_BLINDING parameters
10487 - hide BN_BLINDING structure
10488 Add a second BN_BLINDING slot to the RSA structure to improve
10489 performance when a single RSA object is shared among several
10494 * Add support for DTLS.
10496 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10498 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10499 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10503 * Remove buggy and incomplete DH cert support from
10504 ssl/ssl_rsa.c and ssl/s3_both.c
10508 * Use SHA-1 instead of MD5 as the default digest algorithm for
10509 the `apps/openssl` commands.
10513 * Compile clean with "-Wall -Wmissing-prototypes
10514 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10515 DEBUG_SAFESTACK must also be set.
10519 * Change ./Configure so that certain algorithms can be disabled by default.
10520 The new counterpiece to "no-xxx" is "enable-xxx".
10522 The patented RC5 and MDC2 algorithms will now be disabled unless
10523 "enable-rc5" and "enable-mdc2", respectively, are specified.
10525 (IDEA remains enabled despite being patented. This is because IDEA
10526 is frequently required for interoperability, and there is no license
10527 fee for non-commercial use. As before, "no-idea" can be used to
10528 avoid this algorithm.)
10532 * Add processing of proxy certificates (see RFC 3820). This work was
10533 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10534 EGEE (Enabling Grids for E-science in Europe).
10538 * RC4 performance overhaul on modern architectures/implementations, such
10539 as Intel P4, IA-64 and AMD64.
10543 * New utility extract-section.pl. This can be used specify an alternative
10544 section number in a pod file instead of having to treat each file as
10545 a separate case in Makefile. This can be done by adding two lines to the
10548 =for comment openssl_section:XXX
10550 The blank line is mandatory.
10554 * New arguments -certform, -keyform and -pass for s_client and s_server
10555 to allow alternative format key and certificate files and passphrase
10560 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10561 update associated structures and add various utility functions.
10563 Add new policy related verify parameters, include policy checking in
10564 standard verify code. Enhance 'smime' application with extra parameters
10565 to support policy checking and print out.
10569 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10570 Nehemiah processors. These extensions support AES encryption in hardware
10571 as well as RNG (though RNG support is currently disabled).
10573 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10575 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10579 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10581 *Andy Polyakov and a number of other people*
10583 * Improved PowerPC platform support. Most notably BIGNUM assembler
10584 implementation contributed by IBM.
10586 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10588 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10589 exponent rather than 'unsigned long'. There is a corresponding change to
10590 the new 'rsa_keygen' element of the RSA_METHOD structure.
10592 *Jelte Jansen, Geoff Thorpe*
10594 * Functionality for creating the initial serial number file is now
10595 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10597 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10598 number file to 1, which is bound to cause problems. To avoid
10599 the problems while respecting compatibility between different 0.9.7
10600 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10601 CA.pl for serial number initialization. With the new release 0.9.8,
10602 we can fix the problem directly in the 'ca' utility.)
10606 * Reduced header interdependencies by declaring more opaque objects in
10607 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10608 give fewer recursive includes, which could break lazy source code - so
10609 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10610 developers should define this symbol when building and using openssl to
10611 ensure they track the recommended behaviour, interfaces, [etc], but
10612 backwards-compatible behaviour prevails when this isn't defined.
10616 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10620 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10621 This will generate a random key of the appropriate length based on the
10622 cipher context. The EVP_CIPHER can provide its own random key generation
10623 routine to support keys of a specific form. This is used in the des and
10624 3des routines to generate a key of the correct parity. Update S/MIME
10625 code to use new functions and hence generate correct parity DES keys.
10626 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10627 valid (weak or incorrect parity).
10631 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10632 as looking them up. This is useful when the verified structure may contain
10633 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10634 present unless the new PKCS7_NO_CRL flag is asserted.
10638 * Extend ASN1 oid configuration module. It now additionally accepts the
10641 shortName = some long name, 1.2.3.4
10645 * Reimplemented the BN_CTX implementation. There is now no more static
10646 limitation on the number of variables it can handle nor the depth of the
10647 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10648 information can now expand as required, and rather than having a single
10649 static array of bignums, BN_CTX now uses a linked-list of such arrays
10650 allowing it to expand on demand whilst maintaining the usefulness of
10651 BN_CTX's "bundling".
10655 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10656 to allow all RSA operations to function using a single BN_CTX.
10660 * Preliminary support for certificate policy evaluation and checking. This
10661 is initially intended to pass the tests outlined in "Conformance Testing
10662 of Relying Party Client Certificate Path Processing Logic" v1.07.
10666 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10667 remained unused and not that useful. A variety of other little bignum
10668 tweaks and fixes have also been made continuing on from the audit (see
10673 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10674 associated ASN1, EVP and SSL functions and old ASN1 macros.
10678 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10679 and this should never fail. So the return value from the use of
10680 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10681 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10685 * BN_CTX_get() should return zero-valued bignums, providing the same
10686 initialised value as BN_new().
10688 *Geoff Thorpe, suggested by Ulf Möller*
10690 * Support for inhibitAnyPolicy certificate extension.
10694 * An audit of the BIGNUM code is underway, for which debugging code is
10695 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10696 is considered valid when processing BIGNUMs, and causes execution to
10697 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10698 further steps are taken to deliberately pollute unused data in BIGNUM
10699 structures to try and expose faulty code further on. For now, openssl will
10700 (in its default mode of operation) continue to tolerate the inconsistent
10701 forms that it has tolerated in the past, but authors and packagers should
10702 consider trying openssl and their own applications when compiled with
10703 these debugging symbols defined. It will help highlight potential bugs in
10704 their own code, and will improve the test coverage for OpenSSL itself. At
10705 some point, these tighter rules will become openssl's default to improve
10706 maintainability, though the assert()s and other overheads will remain only
10707 in debugging configurations. See bn.h for more details.
10709 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10711 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10712 that can only be obtained through BN_CTX_new() (which implicitly
10713 initialises it). The presence of this function only made it possible
10714 to overwrite an existing structure (and cause memory leaks).
10718 * Because of the callback-based approach for implementing LHASH as a
10719 template type, lh_insert() adds opaque objects to hash-tables and
10720 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10721 to clean up those corresponding objects before destroying the hash table
10722 (and losing the object pointers). So some over-zealous constifications in
10723 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10724 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10725 prototyped to have "const" restrictions on the object pointers they are
10726 given (and so aren't required to cast them away any more).
10730 * The tmdiff.h API was so ugly and minimal that our own timing utility
10731 (speed) prefers to use its own implementation. The two implementations
10732 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10733 its object type properly exposed (MS_TM) instead of casting to/from
10734 `char *`. This may still change yet if someone realises MS_TM and
10736 aren't necessarily the greatest nomenclatures - but this is what was used
10737 internally to the implementation so I've used that for now.
10741 * Ensure that deprecated functions do not get compiled when
10742 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10743 the self-tests were still using deprecated key-generation functions so
10744 these have been updated also.
10748 * Reorganise PKCS#7 code to separate the digest location functionality
10749 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10750 New function PKCS7_set_digest() to set the digest type for PKCS#7
10751 digestedData type. Add additional code to correctly generate the
10752 digestedData type and add support for this type in PKCS7 initialization
10757 * New function PKCS7_set0_type_other() this initializes a PKCS7
10758 structure of type "other".
10762 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10763 sure the loop does correctly stop and breaking ("division by zero")
10764 modulus operations are not performed. The (pre-generated) prime
10765 table crypto/bn/bn_prime.h was already correct, but it could not be
10766 re-generated on some platforms because of the "division by zero"
10767 situation in the script.
10769 *Ralf S. Engelschall*
10771 * Update support for ECC-based TLS ciphersuites according to
10772 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10773 SHA-1 now is only used for "small" curves (where the
10774 representation of a field element takes up to 24 bytes); for
10775 larger curves, the field element resulting from ECDH is directly
10776 used as premaster secret.
10778 *Douglas Stebila (Sun Microsystems Laboratories)*
10780 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10781 curve secp160r1 to the tests.
10783 *Douglas Stebila (Sun Microsystems Laboratories)*
10785 * Add the possibility to load symbols globally with DSO.
10787 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10789 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10790 control of the error stack.
10794 * Add support for STORE in ENGINE.
10798 * Add the STORE type. The intention is to provide a common interface
10799 to certificate and key stores, be they simple file-based stores, or
10800 HSM-type store, or LDAP stores, or...
10801 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10805 * Add a generic structure called OPENSSL_ITEM. This can be used to
10806 pass a list of arguments to any function as well as provide a way
10807 for a function to pass data back to the caller.
10811 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10812 works like BUF_strdup() but can be used to duplicate a portion of
10813 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10818 * Add the function sk_find_ex() which works like sk_find(), but will
10819 return an index to an element even if an exact match couldn't be
10820 found. The index is guaranteed to point at the element where the
10821 searched-for key would be inserted to preserve sorting order.
10825 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10826 takes an extra flags argument for optional functionality. Currently,
10827 the following flags are defined:
10829 OBJ_BSEARCH_VALUE_ON_NOMATCH
10830 This one gets OBJ_bsearch_ex() to return a pointer to the first
10831 element where the comparing function returns a negative or zero
10834 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10835 This one gets OBJ_bsearch_ex() to return a pointer to the first
10836 element where the comparing function returns zero. This is useful
10837 if there are more than one element where the comparing function
10842 * Make it possible to create self-signed certificates with 'openssl ca'
10843 in such a way that the self-signed certificate becomes part of the
10844 CA database and uses the same mechanisms for serial number generation
10845 as all other certificate signing. The new flag '-selfsign' enables
10846 this functionality. Adapt CA.sh and CA.pl.in.
10850 * Add functionality to check the public key of a certificate request
10851 against a given private. This is useful to check that a certificate
10852 request can be signed by that key (self-signing).
10856 * Make it possible to have multiple active certificates with the same
10857 subject in the CA index file. This is done only if the keyword
10858 'unique_subject' is set to 'no' in the main CA section (default
10859 if 'CA_default') of the configuration file. The value is saved
10860 with the database itself in a separate index attribute file,
10861 named like the index file with '.attr' appended to the name.
10865 * Generate multi-valued AVAs using '+' notation in config files for
10870 * Support for nameConstraints certificate extension.
10874 * Support for policyConstraints certificate extension.
10878 * Support for policyMappings certificate extension.
10882 * Make sure the default DSA_METHOD implementation only uses its
10883 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10884 and change its own handlers to be NULL so as to remove unnecessary
10885 indirection. This lets alternative implementations fallback to the
10886 default implementation more easily.
10890 * Support for directoryName in GeneralName related extensions
10895 * Make it possible to link applications using Makefile.shared.
10896 Make that possible even when linking against static libraries!
10900 * Support for single pass processing for S/MIME signing. This now
10901 means that S/MIME signing can be done from a pipe, in addition
10902 cleartext signing (multipart/signed type) is effectively streaming
10903 and the signed data does not need to be all held in memory.
10905 This is done with a new flag PKCS7_STREAM. When this flag is set
10906 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10907 is done after the data is output (and digests calculated) in
10908 SMIME_write_PKCS7().
10912 * Add full support for -rpath/-R, both in shared libraries and
10913 applications, at least on the platforms where it's known how
10918 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10919 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10920 will now compute a table of multiples of the generator that
10921 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10922 faster (notably in the case of a single point multiplication,
10923 scalar * generator).
10925 *Nils Larsch, Bodo Moeller*
10927 * IPv6 support for certificate extensions. The various extensions
10928 which use the IP:a.b.c.d can now take IPv6 addresses using the
10929 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10934 * Added an ENGINE that implements RSA by performing private key
10935 exponentiations with the GMP library. The conversions to and from
10936 GMP's mpz_t format aren't optimised nor are any montgomery forms
10937 cached, and on x86 it appears OpenSSL's own performance has caught up.
10938 However there are likely to be other architectures where GMP could
10939 provide a boost. This ENGINE is not built in by default, but it can be
10940 specified at Configure time and should be accompanied by the necessary
10941 linker additions, eg;
10942 ./config -DOPENSSL_USE_GMP -lgmp
10946 * "openssl engine" will not display ENGINE/DSO load failure errors when
10947 testing availability of engines with "-t" - the old behaviour is
10948 produced by increasing the feature's verbosity with "-tt".
10952 * ECDSA routines: under certain error conditions uninitialized BN objects
10953 could be freed. Solution: make sure initialization is performed early
10954 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10959 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10960 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10961 software implementations. For DSA and DH, parameter generation can
10962 also be overridden by providing the appropriate method callbacks.
10966 * Change the "progress" mechanism used in key-generation and
10967 primality testing to functions that take a new BN_GENCB pointer in
10968 place of callback/argument pairs. The new API functions have `_ex`
10969 postfixes and the older functions are reimplemented as wrappers for
10970 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10971 declarations of the old functions to help (graceful) attempts to
10972 migrate to the new functions. Also, the new key-generation API
10973 functions operate on a caller-supplied key-structure and return
10974 success/failure rather than returning a key or NULL - this is to
10975 help make "keygen" another member function of RSA_METHOD etc.
10977 Example for using the new callback interface:
10979 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10980 void *my_arg = ...;
10983 BN_GENCB_set(&my_cb, my_callback, my_arg);
10985 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10986 /* For the meaning of a, b in calls to my_callback(), see the
10987 * documentation of the function that calls the callback.
10988 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10989 * my_callback should return 1 if it wants BN_is_prime_ex()
10990 * to continue, or 0 to stop.
10995 * Change the ZLIB compression method to be stateful, and make it
10996 available to TLS with the number defined in
10997 draft-ietf-tls-compression-04.txt.
11001 * Add the ASN.1 structures and functions for CertificatePair, which
11002 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11004 CertificatePair ::= SEQUENCE {
11005 forward [0] Certificate OPTIONAL,
11006 reverse [1] Certificate OPTIONAL,
11007 -- at least one of the pair shall be present -- }
11009 Also implement the PEM functions to read and write certificate
11010 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11012 This needed to be defined, mostly for the sake of the LDAP
11013 attribute crossCertificatePair, but may prove useful elsewhere as
11018 * Make it possible to inhibit symlinking of shared libraries in
11019 Makefile.shared, for Cygwin's sake.
11023 * Extend the BIGNUM API by creating a function
11024 void BN_set_negative(BIGNUM *a, int neg);
11025 and a macro that behave like
11026 int BN_is_negative(const BIGNUM *a);
11028 to avoid the need to access 'a->neg' directly in applications.
11032 * Implement fast modular reduction for pseudo-Mersenne primes
11033 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11034 EC_GROUP_new_curve_GFp() will now automatically use this
11037 *Nils Larsch <nla@trustcenter.de>*
11039 * Add new lock type (CRYPTO_LOCK_BN).
11043 * Change the ENGINE framework to automatically load engines
11044 dynamically from specific directories unless they could be
11045 found to already be built in or loaded. Move all the
11046 current engines except for the cryptodev one to a new
11047 directory engines/.
11048 The engines in engines/ are built as shared libraries if
11049 the "shared" options was given to ./Configure or ./config.
11050 Otherwise, they are inserted in libcrypto.a.
11051 /usr/local/ssl/engines is the default directory for dynamic
11052 engines, but that can be overridden at configure time through
11053 the usual use of --prefix and/or --openssldir, and at run
11054 time with the environment variable OPENSSL_ENGINES.
11056 *Geoff Thorpe and Richard Levitte*
11058 * Add Makefile.shared, a helper makefile to build shared
11059 libraries. Adapt Makefile.org.
11063 * Add version info to Win32 DLLs.
11065 *Peter 'Luna' Runestig" <peter@runestig.com>*
11067 * Add new 'medium level' PKCS#12 API. Certificates and keys
11068 can be added using this API to created arbitrary PKCS#12
11069 files while avoiding the low-level API.
11071 New options to PKCS12_create(), key or cert can be NULL and
11072 will then be omitted from the output file. The encryption
11073 algorithm NIDs can be set to -1 for no encryption, the mac
11074 iteration count can be set to 0 to omit the mac.
11076 Enhance pkcs12 utility by making the -nokeys and -nocerts
11077 options work when creating a PKCS#12 file. New option -nomac
11078 to omit the mac, NONE can be set for an encryption algorithm.
11079 New code is modified to use the enhanced PKCS12_create()
11080 instead of the low-level API.
11084 * Extend ASN1 encoder to support indefinite length constructed
11085 encoding. This can output sequences tags and octet strings in
11086 this form. Modify pk7_asn1.c to support indefinite length
11087 encoding. This is experimental and needs additional code to
11088 be useful, such as an ASN1 bio and some enhanced streaming
11091 Extend template encode functionality so that tagging is passed
11092 down to the template encoder.
11096 * Let 'openssl req' fail if an argument to '-newkey' is not
11097 recognized instead of using RSA as a default.
11101 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11102 As these are not official, they are not included in "ALL";
11103 the "ECCdraft" ciphersuite group alias can be used to select them.
11105 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11107 * Add ECDH engine support.
11109 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11111 * Add ECDH in new directory crypto/ecdh/.
11113 *Douglas Stebila (Sun Microsystems Laboratories)*
11115 * Let BN_rand_range() abort with an error after 100 iterations
11116 without success (which indicates a broken PRNG).
11120 * Change BN_mod_sqrt() so that it verifies that the input value
11121 is really the square of the return value. (Previously,
11122 BN_mod_sqrt would show GIGO behaviour.)
11126 * Add named elliptic curves over binary fields from X9.62, SECG,
11127 and WAP/WTLS; add OIDs that were still missing.
11129 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11131 * Extend the EC library for elliptic curves over binary fields
11132 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11135 EC_GF2m_simple_method
11139 EC_GROUP_new_curve_GF2m
11140 EC_GROUP_set_curve_GF2m
11141 EC_GROUP_get_curve_GF2m
11142 EC_POINT_set_affine_coordinates_GF2m
11143 EC_POINT_get_affine_coordinates_GF2m
11144 EC_POINT_set_compressed_coordinates_GF2m
11146 Point compression for binary fields is disabled by default for
11147 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11150 As binary polynomials are represented as BIGNUMs, various members
11151 of the EC_GROUP and EC_POINT data structures can be shared
11152 between the implementations for prime fields and binary fields;
11153 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11154 are essentially identical to their `..._GFp` counterparts.
11155 (For simplicity, the `..._GFp` prefix has been dropped from
11156 various internal method names.)
11158 An internal 'field_div' method (similar to 'field_mul' and
11159 'field_sqr') has been added; this is used only for binary fields.
11161 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11163 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11164 through methods ('mul', 'precompute_mult').
11166 The generic implementations (now internally called 'ec_wNAF_mul'
11167 and 'ec_wNAF_precomputed_mult') remain the default if these
11168 methods are undefined.
11170 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11172 * New function EC_GROUP_get_degree, which is defined through
11173 EC_METHOD. For curves over prime fields, this returns the bit
11174 length of the modulus.
11176 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11178 * New functions EC_GROUP_dup, EC_POINT_dup.
11179 (These simply call ..._new and ..._copy).
11181 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11183 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11184 Polynomials are represented as BIGNUMs (where the sign bit is not
11185 used) in the following functions [macros]:
11188 BN_GF2m_sub [= BN_GF2m_add]
11189 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11190 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11191 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11193 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11194 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11195 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11196 BN_GF2m_cmp [= BN_ucmp]
11198 (Note that only the 'mod' functions are actually for fields GF(2^m).
11199 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11201 For some functions, an the irreducible polynomial defining a
11202 field can be given as an 'unsigned int[]' with strictly
11203 decreasing elements giving the indices of those bits that are set;
11204 i.e., p[] represents the polynomial
11205 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11207 p[0] > p[1] > ... > p[k] = 0.
11208 This applies to the following functions:
11211 BN_GF2m_mod_mul_arr
11212 BN_GF2m_mod_sqr_arr
11213 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11214 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11215 BN_GF2m_mod_exp_arr
11216 BN_GF2m_mod_sqrt_arr
11217 BN_GF2m_mod_solve_quad_arr
11221 Conversion can be performed by the following functions:
11226 bntest.c has additional tests for binary polynomial arithmetic.
11228 Two implementations for BN_GF2m_mod_div() are available.
11229 The default algorithm simply uses BN_GF2m_mod_inv() and
11230 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11231 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11232 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11234 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11236 * Add new error code 'ERR_R_DISABLED' that can be used when some
11237 functionality is disabled at compile-time.
11239 *Douglas Stebila <douglas.stebila@sun.com>*
11241 * Change default behaviour of 'openssl asn1parse' so that more
11242 information is visible when viewing, e.g., a certificate:
11244 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11245 mode the content of non-printable OCTET STRINGs is output in a
11246 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11247 avoid the appearance of a printable string.
11249 *Nils Larsch <nla@trustcenter.de>*
11251 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11253 EC_GROUP_set_asn1_flag()
11254 EC_GROUP_get_asn1_flag()
11255 EC_GROUP_set_point_conversion_form()
11256 EC_GROUP_get_point_conversion_form()
11257 These control ASN1 encoding details:
11258 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11259 has been set to OPENSSL_EC_NAMED_CURVE.
11260 - Points are encoded in uncompressed form by default; options for
11261 asn1_for are as for point2oct, namely
11262 POINT_CONVERSION_COMPRESSED
11263 POINT_CONVERSION_UNCOMPRESSED
11264 POINT_CONVERSION_HYBRID
11266 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11268 EC_GROUP_set_seed()
11269 EC_GROUP_get0_seed()
11270 EC_GROUP_get_seed_len()
11271 This is used only for ASN1 purposes (so far).
11273 *Nils Larsch <nla@trustcenter.de>*
11275 * Add 'field_type' member to EC_METHOD, which holds the NID
11276 of the appropriate field type OID. The new function
11277 EC_METHOD_get_field_type() returns this value.
11279 *Nils Larsch <nla@trustcenter.de>*
11282 EC_POINT_point2bn()
11283 EC_POINT_bn2point()
11284 EC_POINT_point2hex()
11285 EC_POINT_hex2point()
11286 providing useful interfaces to EC_POINT_point2oct() and
11287 EC_POINT_oct2point().
11289 *Nils Larsch <nla@trustcenter.de>*
11291 * Change internals of the EC library so that the functions
11292 EC_GROUP_set_generator()
11293 EC_GROUP_get_generator()
11294 EC_GROUP_get_order()
11295 EC_GROUP_get_cofactor()
11296 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11297 to methods, which would lead to unnecessary code duplication when
11298 adding different types of curves.
11300 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11302 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11303 arithmetic, and such that modified wNAFs are generated
11304 (which avoid length expansion in many cases).
11308 * Add a function EC_GROUP_check_discriminant() (defined via
11309 EC_METHOD) that verifies that the curve discriminant is non-zero.
11311 Add a function EC_GROUP_check() that makes some sanity tests
11312 on a EC_GROUP, its generator and order. This includes
11313 EC_GROUP_check_discriminant().
11315 *Nils Larsch <nla@trustcenter.de>*
11317 * Add ECDSA in new directory crypto/ecdsa/.
11319 Add applications 'openssl ecparam' and 'openssl ecdsa'
11320 (these are based on 'openssl dsaparam' and 'openssl dsa').
11322 ECDSA support is also included in various other files across the
11323 library. Most notably,
11324 - 'openssl req' now has a '-newkey ecdsa:file' option;
11325 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11326 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11327 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11328 them suitable for ECDSA where domain parameters must be
11329 extracted before the specific public key;
11330 - ECDSA engine support has been added.
11332 *Nils Larsch <nla@trustcenter.de>*
11334 * Include some named elliptic curves, and add OIDs from X9.62,
11335 SECG, and WAP/WTLS. Each curve can be obtained from the new
11337 EC_GROUP_new_by_curve_name(),
11338 and the list of available named curves can be obtained with
11339 EC_get_builtin_curves().
11340 Also add a 'curve_name' member to EC_GROUP objects, which can be
11342 EC_GROUP_set_curve_name()
11343 EC_GROUP_get_curve_name()
11345 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11347 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11348 was actually never needed) and in BN_mul(). The removal in BN_mul()
11349 required a small change in bn_mul_part_recursive() and the addition
11350 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11351 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11352 bn_sub_words() and bn_add_words() except they take arrays with
11357 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11359 * Cleanse PEM buffers before freeing them since they may contain
11362 *Benjamin Bennett <ben@psc.edu>*
11364 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11365 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11366 authentication-only ciphersuites.
11370 * Since AES128 and AES256 share a single mask bit in the logic of
11371 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11372 kludge to work properly if AES128 is available and AES256 isn't.
11376 * Expand security boundary to match 1.1.1 module.
11380 * Remove redundant features: hash file source, editing of test vectors
11381 modify fipsld to use external fips_premain.c signature.
11385 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11386 run algorithm test programs.
11390 * Make algorithm test programs more tolerant of whitespace.
11394 * Have SSL/TLS server implementation tolerate "mismatched" record
11395 protocol version while receiving ClientHello even if the
11396 ClientHello is fragmented. (The server can't insist on the
11397 particular protocol version it has chosen before the ServerHello
11398 message has informed the client about his choice.)
11402 * Load error codes if they are not already present instead of using a
11403 static variable. This allows them to be cleanly unloaded and reloaded.
11407 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11409 * Introduce limits to prevent malicious keys being able to
11410 cause a denial of service. ([CVE-2006-2940])
11412 *Steve Henson, Bodo Moeller*
11414 * Fix ASN.1 parsing of certain invalid structures that can result
11415 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11417 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11418 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11420 * Fix SSL client code which could crash if connecting to a
11421 malicious SSLv2 server. ([CVE-2006-4343])
11423 *Tavis Ormandy and Will Drewry, Google Security Team*
11425 * Change ciphersuite string processing so that an explicit
11426 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11427 will no longer include "AES128-SHA"), and any other similar
11428 ciphersuite (same bitmap) from *other* protocol versions (so that
11429 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11430 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11431 changes from 0.9.8b and 0.9.8d.
11435 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11437 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11438 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11440 * Change the Unix randomness entropy gathering to use poll() when
11441 possible instead of select(), since the latter has some
11442 undesirable limitations.
11444 *Darryl Miles via Richard Levitte and Bodo Moeller*
11446 * Disable rogue ciphersuites:
11448 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11449 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11450 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11452 The latter two were purportedly from
11453 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11456 Also deactivate the remaining ciphersuites from
11457 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11458 unofficial, and the ID has long expired.
11462 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11463 dual-core machines) and other potential thread-safety issues.
11467 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11469 * Adapt fipsld and the build system to link against the validated FIPS
11470 module in FIPS mode.
11474 * Fixes for VC++ 2005 build under Windows.
11478 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11479 from a Windows bash shell such as MSYS. It is autodetected from the
11480 "config" script when run from a VC++ environment. Modify standard VC++
11481 build to use fipscanister.o from the GNU make build.
11485 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11487 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11488 The value now differs depending on if you build for FIPS or not.
11489 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11490 safely run with a non-FIPSed libcrypto, as it may crash because of
11491 the difference induced by this change.
11495 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11497 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11498 (part of SSL_OP_ALL). This option used to disable the
11499 countermeasure against man-in-the-middle protocol-version
11500 rollback in the SSL 2.0 server implementation, which is a bad
11501 idea. ([CVE-2005-2969])
11503 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11504 for Information Security, National Institute of Advanced Industrial
11505 Science and Technology [AIST, Japan)]*
11507 * Minimal support for X9.31 signatures and PSS padding modes. This is
11508 mainly for FIPS compliance and not fully integrated at this stage.
11512 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11513 the exponentiation using a fixed-length exponent. (Otherwise,
11514 the information leaked through timing could expose the secret key
11515 after many signatures; cf. Bleichenbacher's attack on DSA with
11520 * Make a new fixed-window mod_exp implementation the default for
11521 RSA, DSA, and DH private-key operations so that the sequence of
11522 squares and multiplies and the memory access pattern are
11523 independent of the particular secret key. This will mitigate
11524 cache-timing and potential related attacks.
11526 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11527 and this is automatically used by BN_mod_exp_mont() if the new flag
11528 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11529 will use this BN flag for private exponents unless the flag
11530 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11531 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11533 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11535 * Change the client implementation for SSLv23_method() and
11536 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11537 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11538 (Previously, the SSL 2.0 backwards compatible Client Hello
11539 message format would be used even with SSL_OP_NO_SSLv2.)
11543 * Add support for smime-type MIME parameter in S/MIME messages which some
11548 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11549 a threadsafe manner. Modify rsa code to use new function and add calls
11550 to dsa and dh code (which had race conditions before).
11554 * Include the fixed error library code in the C error file definitions
11555 instead of fixing them up at runtime. This keeps the error code
11556 structures constant.
11560 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11562 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11565 * Fixes for newer kerberos headers. NB: the casts are needed because
11566 the 'length' field is signed on one version and unsigned on another
11567 with no (?) obvious way to tell the difference, without these VC++
11568 complains. Also the "definition" of FAR (blank) is no longer included
11569 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11570 some needed definitions.
11574 * Undo Cygwin change.
11578 * Added support for proxy certificates according to RFC 3820.
11579 Because they may be a security thread to unaware applications,
11580 they must be explicitly allowed in run-time. See
11581 docs/HOWTO/proxy_certificates.txt for further information.
11585 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11587 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11588 server and client random values. Previously
11589 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11590 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11592 This change has negligible security impact because:
11594 1. Server and client random values still have 24 bytes of pseudo random
11597 2. Server and client random values are sent in the clear in the initial
11600 3. The master secret is derived using the premaster secret (48 bytes in
11601 size for static RSA ciphersuites) as well as client server and random
11604 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11607 *Stephen Henson, reported by UK NISCC*
11609 * Use Windows randomness collection on Cygwin.
11613 * Fix hang in EGD/PRNGD query when communication socket is closed
11614 prematurely by EGD/PRNGD.
11616 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11618 * Prompt for pass phrases when appropriate for PKCS12 input format.
11622 * Back-port of selected performance improvements from development
11623 branch, as well as improved support for PowerPC platforms.
11627 * Add lots of checks for memory allocation failure, error codes to indicate
11628 failure and freeing up memory if a failure occurs.
11630 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11632 * Add new -passin argument to dgst.
11636 * Perform some character comparisons of different types in X509_NAME_cmp:
11637 this is needed for some certificates that re-encode DNs into UTF8Strings
11638 (in violation of RFC3280) and can't or won't issue name rollover
11643 * Make an explicit check during certificate validation to see that
11644 the CA setting in each certificate on the chain is correct. As a
11645 side effect always do the following basic checks on extensions,
11646 not just when there's an associated purpose to the check:
11648 - if there is an unhandled critical extension (unless the user
11649 has chosen to ignore this fault)
11650 - if the path length has been exceeded (if one is set at all)
11651 - that certain extensions fit the associated purpose (if one has
11656 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11658 * Avoid a race condition when CRLs are checked in a multi threaded
11659 environment. This would happen due to the reordering of the revoked
11660 entries during signature checking and serial number lookup. Now the
11661 encoding is cached and the serial number sort performed under a lock.
11662 Add new STACK function sk_is_sorted().
11666 * Add Delta CRL to the extension code.
11670 * Various fixes to s3_pkt.c so alerts are sent properly.
11672 *David Holmes <d.holmes@f5.com>*
11674 * Reduce the chances of duplicate issuer name and serial numbers (in
11675 violation of RFC3280) using the OpenSSL certificate creation utilities.
11676 This is done by creating a random 64 bit value for the initial serial
11677 number when a serial number file is created or when a self signed
11678 certificate is created using 'openssl req -x509'. The initial serial
11679 number file is created using 'openssl x509 -next_serial' in CA.pl
11680 rather than being initialized to 1.
11684 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11686 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11687 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11689 *Joe Orton, Steve Henson*
11691 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11694 *Joe Orton, Steve Henson*
11696 * Make it possible to have multiple active certificates with the same
11697 subject in the CA index file. This is done only if the keyword
11698 'unique_subject' is set to 'no' in the main CA section (default
11699 if 'CA_default') of the configuration file. The value is saved
11700 with the database itself in a separate index attribute file,
11701 named like the index file with '.attr' appended to the name.
11705 * X509 verify fixes. Disable broken certificate workarounds when
11706 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11707 keyUsage extension present. Don't accept CRLs with unhandled critical
11708 extensions: since verify currently doesn't process CRL extensions this
11709 rejects a CRL with *any* critical extensions. Add new verify error codes
11714 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11715 A clarification of RFC2560 will require the use of OCTET STRINGs and
11716 some implementations cannot handle the current raw format. Since OpenSSL
11717 copies and compares OCSP nonces as opaque blobs without any attempt at
11718 parsing them this should not create any compatibility issues.
11722 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11723 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11724 this HMAC (and other) operations are several times slower than OpenSSL
11729 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11731 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11733 * Use the correct content when signing type "other".
11737 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11739 * Fix various bugs revealed by running the NISCC test suite:
11741 Stop out of bounds reads in the ASN1 code when presented with
11742 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11744 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11746 If verify callback ignores invalid public key errors don't try to check
11747 certificate signature with the NULL public key.
11751 * New -ignore_err option in ocsp application to stop the server
11752 exiting on the first error in a request.
11756 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11757 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11762 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11763 extra data after the compression methods not only for TLS 1.0
11764 but also for SSL 3.0 (as required by the specification).
11766 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11768 * Change X509_certificate_type() to mark the key as exported/exportable
11769 when it's 512 *bits* long, not 512 bytes.
11773 * Change AES_cbc_encrypt() so it outputs exact multiple of
11774 blocks during encryption.
11778 * Various fixes to base64 BIO and non blocking I/O. On write
11779 flushes were not handled properly if the BIO retried. On read
11780 data was not being buffered properly and had various logic bugs.
11781 This also affects blocking I/O when the data being decoded is a
11786 * Various S/MIME bugfixes and compatibility changes:
11787 output correct application/pkcs7 MIME type if
11788 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11789 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11790 of files as .eml work). Correctly handle very long lines in MIME
11795 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11797 * Countermeasure against the Klima-Pokorny-Rosa extension of
11798 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11799 a protocol version number mismatch like a decryption error
11800 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11804 * Turn on RSA blinding by default in the default implementation
11805 to avoid a timing attack. Applications that don't want it can call
11806 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11807 They would be ill-advised to do so in most cases.
11809 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11811 * Change RSA blinding code so that it works when the PRNG is not
11812 seeded (in this case, the secret RSA exponent is abused as
11813 an unpredictable seed -- if it is not unpredictable, there
11814 is no point in blinding anyway). Make RSA blinding thread-safe
11815 by remembering the creator's thread ID in rsa->blinding and
11816 having all other threads use local one-time blinding factors
11817 (this requires more computation than sharing rsa->blinding, but
11818 avoids excessive locking; and if an RSA object is not shared
11819 between threads, blinding will still be very fast).
11823 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11824 ENGINE as defaults for all supported algorithms irrespective of
11825 the 'flags' parameter. 'flags' is now honoured, so applications
11826 should make sure they are passing it correctly.
11830 * Target "mingw" now allows native Windows code to be generated in
11831 the Cygwin environment as well as with the MinGW compiler.
11835 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11837 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11838 via timing by performing a MAC computation even if incorrect
11839 block cipher padding has been found. This is a countermeasure
11840 against active attacks where the attacker has to distinguish
11841 between bad padding and a MAC verification error. ([CVE-2003-0078])
11843 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11844 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11845 Martin Vuagnoux (EPFL, Ilion)*
11847 * Make the no-err option work as intended. The intention with no-err
11848 is not to have the whole error stack handling routines removed from
11849 libcrypto, it's only intended to remove all the function name and
11850 reason texts, thereby removing some of the footprint that may not
11851 be interesting if those errors aren't displayed anyway.
11853 NOTE: it's still possible for any application or module to have its
11854 own set of error texts inserted. The routines are there, just not
11855 used by default when no-err is given.
11859 * Add support for FreeBSD on IA64.
11861 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11863 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11864 Kerberos function mit_des_cbc_cksum(). Before this change,
11865 the value returned by DES_cbc_cksum() was like the one from
11866 mit_des_cbc_cksum(), except the bytes were swapped.
11868 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11870 * Allow an application to disable the automatic SSL chain building.
11871 Before this a rather primitive chain build was always performed in
11872 ssl3_output_cert_chain(): an application had no way to send the
11873 correct chain if the automatic operation produced an incorrect result.
11875 Now the chain builder is disabled if either:
11877 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11879 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11881 The reasoning behind this is that an application would not want the
11882 auto chain building to take place if extra chain certificates are
11883 present and it might also want a means of sending no additional
11884 certificates (for example the chain has two certificates and the
11889 * Add the possibility to build without the ENGINE framework.
11891 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11893 * Under Win32 gmtime() can return NULL: check return value in
11894 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11898 * DSA routines: under certain error conditions uninitialized BN objects
11899 could be freed. Solution: make sure initialization is performed early
11900 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11901 Nils Larsch <nla@trustcenter.de> via PR#459)
11905 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11906 checked on reconnect on the client side, therefore session resumption
11907 could still fail with a "ssl session id is different" error. This
11908 behaviour is masked when SSL_OP_ALL is used due to
11909 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11910 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11911 followup to PR #377.
11915 * IA-32 assembler support enhancements: unified ELF targets, support
11916 for SCO/Caldera platforms, fix for Cygwin shared build.
11920 * Add support for FreeBSD on sparc64. As a consequence, support for
11921 FreeBSD on non-x86 processors is separate from x86 processors on
11922 the config script, much like the NetBSD support.
11924 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11926 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11928 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11931 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11932 code (06) was taken as the first octet of the session ID and the last
11933 octet was ignored consequently. As a result SSLv2 client side session
11934 caching could not have worked due to the session ID mismatch between
11936 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11941 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11942 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11947 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11948 seems that in spite of existing for more than a year, many application
11949 author have done nothing to provide the necessary callbacks, which
11950 means that this particular engine will not work properly anywhere.
11951 This is a very unfortunate situation which forces us, in the name
11952 of usability, to give the hw_ncipher.c a static lock, which is part
11954 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11955 appear in 0.9.8 or later. We EXPECT application authors to have
11956 dealt properly with this when 0.9.8 is released (unless we actually
11957 make such changes in the libcrypto locking code that changes will
11958 have to be made anyway).
11962 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11963 octets have been read, EOF or an error occurs. Without this change
11964 some truncated ASN1 structures will not produce an error.
11968 * Disable Heimdal support, since it hasn't been fully implemented.
11969 Still give the possibility to force the use of Heimdal, but with
11970 warnings and a request that patches get sent to openssl-dev.
11974 * Add the VC-CE target, introduce the WINCE sysname, and add
11975 INSTALL.WCE and appropriate conditionals to make it build.
11977 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11979 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11980 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11981 edit numbers of the version.
11983 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11985 * Introduce safe string copy and catenation functions
11986 (BUF_strlcpy() and BUF_strlcat()).
11988 *Ben Laurie (CHATS) and Richard Levitte*
11990 * Avoid using fixed-size buffers for one-line DNs.
11992 *Ben Laurie (CHATS)*
11994 * Add BUF_MEM_grow_clean() to avoid information leakage when
11995 resizing buffers containing secrets, and use where appropriate.
11997 *Ben Laurie (CHATS)*
11999 * Avoid using fixed size buffers for configuration file location.
12001 *Ben Laurie (CHATS)*
12003 * Avoid filename truncation for various CA files.
12005 *Ben Laurie (CHATS)*
12007 * Use sizeof in preference to magic numbers.
12009 *Ben Laurie (CHATS)*
12011 * Avoid filename truncation in cert requests.
12013 *Ben Laurie (CHATS)*
12015 * Add assertions to check for (supposedly impossible) buffer
12018 *Ben Laurie (CHATS)*
12020 * Don't cache truncated DNS entries in the local cache (this could
12021 potentially lead to a spoofing attack).
12023 *Ben Laurie (CHATS)*
12025 * Fix various buffers to be large enough for hex/decimal
12026 representations in a platform independent manner.
12028 *Ben Laurie (CHATS)*
12030 * Add CRYPTO_realloc_clean() to avoid information leakage when
12031 resizing buffers containing secrets, and use where appropriate.
12033 *Ben Laurie (CHATS)*
12035 * Add BIO_indent() to avoid much slightly worrying code to do
12038 *Ben Laurie (CHATS)*
12040 * Convert sprintf()/BIO_puts() to BIO_printf().
12042 *Ben Laurie (CHATS)*
12044 * buffer_gets() could terminate with the buffer only half
12047 *Ben Laurie (CHATS)*
12049 * Add assertions to prevent user-supplied crypto functions from
12050 overflowing internal buffers by having large block sizes, etc.
12052 *Ben Laurie (CHATS)*
12054 * New OPENSSL_assert() macro (similar to assert(), but enabled
12057 *Ben Laurie (CHATS)*
12059 * Eliminate unused copy of key in RC4.
12061 *Ben Laurie (CHATS)*
12063 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12065 *Ben Laurie (CHATS)*
12067 * Fix off-by-one error in EGD path.
12069 *Ben Laurie (CHATS)*
12071 * If RANDFILE path is too long, ignore instead of truncating.
12073 *Ben Laurie (CHATS)*
12075 * Eliminate unused and incorrectly sized X.509 structure
12078 *Ben Laurie (CHATS)*
12080 * Eliminate unused and dangerous function knumber().
12082 *Ben Laurie (CHATS)*
12084 * Eliminate unused and dangerous structure, KSSL_ERR.
12086 *Ben Laurie (CHATS)*
12088 * Protect against overlong session ID context length in an encoded
12089 session object. Since these are local, this does not appear to be
12092 *Ben Laurie (CHATS)*
12094 * Change from security patch (see 0.9.6e below) that did not affect
12095 the 0.9.6 release series:
12097 Remote buffer overflow in SSL3 protocol - an attacker could
12098 supply an oversized master key in Kerberos-enabled versions.
12101 *Ben Laurie (CHATS)*
12103 * Change the SSL kerb5 codes to match RFC 2712.
12107 * Make -nameopt work fully for req and add -reqopt switch.
12109 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12111 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12113 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12115 * Make sure tests can be performed even if the corresponding algorithms
12116 have been removed entirely. This was also the last step to make
12117 OpenSSL compilable with DJGPP under all reasonable conditions.
12119 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12121 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12122 to allow version independent disabling of normally unselected ciphers,
12123 which may be activated as a side-effect of selecting a single cipher.
12125 (E.g., cipher list string "RSA" enables ciphersuites that are left
12126 out of "ALL" because they do not provide symmetric encryption.
12127 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12129 *Lutz Jaenicke, Bodo Moeller*
12131 * Add appropriate support for separate platform-dependent build
12132 directories. The recommended way to make a platform-dependent
12133 build directory is the following (tested on Linux), maybe with
12136 # Place yourself outside of the OpenSSL source tree. In
12137 # this example, the environment variable OPENSSL_SOURCE
12138 # is assumed to contain the absolute OpenSSL source directory.
12139 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12140 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12141 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12142 mkdir -p `dirname $F`
12143 ln -s $OPENSSL_SOURCE/$F $F
12146 To be absolutely sure not to disturb the source tree, a "make clean"
12147 is a good thing. If it isn't successful, don't worry about it,
12148 it probably means the source directory is very clean.
12152 * Make sure any ENGINE control commands make local copies of string
12153 pointers passed to them whenever necessary. Otherwise it is possible
12154 the caller may have overwritten (or deallocated) the original string
12155 data when a later ENGINE operation tries to use the stored values.
12157 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12159 * Improve diagnostics in file reading and command-line digests.
12161 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12163 * Add AES modes CFB and OFB to the object database. Correct an
12164 error in AES-CFB decryption.
12168 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12169 allows existing EVP_CIPHER_CTX structures to be reused after
12170 calling `EVP_*Final()`. This behaviour is used by encryption
12171 BIOs and some applications. This has the side effect that
12172 applications must explicitly clean up cipher contexts with
12173 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12177 * Check the values of dna and dnb in bn_mul_recursive before calling
12178 bn_mul_comba (a non zero value means the a or b arrays do not contain
12179 n2 elements) and fallback to bn_mul_normal if either is not zero.
12183 * Fix escaping of non-ASCII characters when using the -subj option
12184 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12188 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12189 form for "surname", serialNumber has no short form.
12190 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12191 therefore remove "mail" short name for "internet 7".
12192 The OID for unique identifiers in X509 certificates is
12193 x500UniqueIdentifier, not uniqueIdentifier.
12194 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12198 * Add an "init" command to the ENGINE config module and auto initialize
12199 ENGINEs. Without any "init" command the ENGINE will be initialized
12200 after all ctrl commands have been executed on it. If init=1 the
12201 ENGINE is initialized at that point (ctrls before that point are run
12202 on the uninitialized ENGINE and after on the initialized one). If
12203 init=0 then the ENGINE will not be initialized at all.
12207 * Fix the 'app_verify_callback' interface so that the user-defined
12208 argument is actually passed to the callback: In the
12209 SSL_CTX_set_cert_verify_callback() prototype, the callback
12210 declaration has been changed from
12213 int (*cb)(X509_STORE_CTX *,void *);
12214 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12215 i=s->ctx->app_verify_callback(&ctx)
12216 has been changed into
12217 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12219 To update applications using SSL_CTX_set_cert_verify_callback(),
12220 a dummy argument can be added to their callback functions.
12222 *D. K. Smetters <smetters@parc.xerox.com>*
12224 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12226 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12228 * Add and OPENSSL_LOAD_CONF define which will cause
12229 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12230 This allows older applications to transparently support certain
12231 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12232 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12233 load the config file and OPENSSL_add_all_algorithms_conf() which will
12234 always load it have also been added.
12238 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12239 Adjust NIDs and EVP layer.
12241 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12243 * Config modules support in openssl utility.
12245 Most commands now load modules from the config file,
12246 though in a few (such as version) this isn't done
12247 because it couldn't be used for anything.
12249 In the case of ca and req the config file used is
12250 the same as the utility itself: that is the -config
12251 command line option can be used to specify an
12256 * Move default behaviour from OPENSSL_config(). If appname is NULL
12257 use "openssl_conf" if filename is NULL use default openssl config file.
12261 * Add an argument to OPENSSL_config() to allow the use of an alternative
12262 config section name. Add a new flag to tolerate a missing config file
12263 and move code to CONF_modules_load_file().
12267 * Support for crypto accelerator cards from Accelerated Encryption
12268 Processing, www.aep.ie. (Use engine 'aep')
12269 The support was copied from 0.9.6c [engine] and adapted/corrected
12270 to work with the new engine framework.
12272 *AEP Inc. and Richard Levitte*
12274 * Support for SureWare crypto accelerator cards from Baltimore
12275 Technologies. (Use engine 'sureware')
12276 The support was copied from 0.9.6c [engine] and adapted
12277 to work with the new engine framework.
12281 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12282 make the newer ENGINE framework commands for the CHIL engine work.
12284 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12286 * Make it possible to produce shared libraries on ReliantUNIX.
12288 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12290 * Add the configuration target debug-linux-ppro.
12291 Make 'openssl rsa' use the general key loading routines
12292 implemented in `apps.c`, and make those routines able to
12293 handle the key format FORMAT_NETSCAPE and the variant
12296 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12298 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12300 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12302 * Add -keyform to rsautl, and document -engine.
12304 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12306 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12307 BIO_R_NO_SUCH_FILE error code rather than the generic
12308 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12312 * Add new functions
12313 ERR_peek_last_error
12314 ERR_peek_last_error_line
12315 ERR_peek_last_error_line_data.
12316 These are similar to
12318 ERR_peek_error_line
12319 ERR_peek_error_line_data,
12320 but report on the latest error recorded rather than the first one
12321 still in the error queue.
12323 *Ben Laurie, Bodo Moeller*
12325 * default_algorithms option in ENGINE config module. This allows things
12327 default_algorithms = ALL
12328 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12332 * Preliminary ENGINE config module.
12336 * New experimental application configuration code.
12340 * Change the AES code to follow the same name structure as all other
12341 symmetric ciphers, and behave the same way. Move everything to
12342 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12344 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12346 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12348 *Ben Laurie and Theo de Raadt*
12350 * Add option to output public keys in req command.
12352 *Massimiliano Pala madwolf@openca.org*
12354 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12355 (up to about 10% better than before for P-192 and P-224).
12359 * New functions/macros
12361 SSL_CTX_set_msg_callback(ctx, cb)
12362 SSL_CTX_set_msg_callback_arg(ctx, arg)
12363 SSL_set_msg_callback(ssl, cb)
12364 SSL_set_msg_callback_arg(ssl, arg)
12366 to request calling a callback function
12368 void cb(int write_p, int version, int content_type,
12369 const void *buf, size_t len, SSL *ssl, void *arg)
12371 whenever a protocol message has been completely received
12372 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12373 protocol version according to which the SSL library interprets
12374 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12375 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12376 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12377 specification (change_cipher_spec(20), alert(21), handshake(22)).
12378 'buf' and 'len' point to the actual message, 'ssl' to the
12379 SSL object, and 'arg' is the application-defined value set by
12380 SSL[_CTX]_set_msg_callback_arg().
12382 'openssl s_client' and 'openssl s_server' have new '-msg' options
12383 to enable a callback that displays all protocol messages.
12387 * Change the shared library support so shared libraries are built as
12388 soon as the corresponding static library is finished, and thereby get
12389 openssl and the test programs linked against the shared library.
12390 This still only happens when the keyword "shard" has been given to
12391 the configuration scripts.
12393 NOTE: shared library support is still an experimental thing, and
12394 backward binary compatibility is still not guaranteed.
12396 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12398 * Add support for Subject Information Access extension.
12400 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12402 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12403 additional bytes when new memory had to be allocated, not just
12404 when reusing an existing buffer.
12408 * New command line and configuration option 'utf8' for the req command.
12409 This allows field values to be specified as UTF8 strings.
12413 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12414 runs for the former and machine-readable output for the latter.
12418 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12419 of the e-mail address in the DN (i.e., it will go into a certificate
12420 extension only). The new configuration file option 'email_in_dn = no'
12421 has the same effect.
12423 *Massimiliano Pala madwolf@openca.org*
12425 * Change all functions with names starting with `des_` to be starting
12426 with `DES_` instead. Add wrappers that are compatible with libdes,
12427 but are named `_ossl_old_des_*`. Finally, add macros that map the
12428 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12429 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12430 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12433 Since we provide two compatibility mappings, the user needs to
12434 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12435 compatibility is desired. The default (i.e., when that macro
12436 isn't defined) is OpenSSL 0.9.6c compatibility.
12438 There are also macros that enable and disable the support of old
12439 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12440 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12441 are defined, the default will apply: to support the old des routines.
12443 In either case, one must include openssl/des.h to get the correct
12444 definitions. Do not try to just include openssl/des_old.h, that
12447 NOTE: This is a major break of an old API into a new one. Software
12448 authors are encouraged to switch to the `DES_` style functions. Some
12449 time in the future, des_old.h and the libdes compatibility functions
12450 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12451 default), and then completely removed.
12455 * Test for certificates which contain unsupported critical extensions.
12456 If such a certificate is found during a verify operation it is
12457 rejected by default: this behaviour can be overridden by either
12458 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12459 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12460 X509_supported_extension() has also been added which returns 1 if a
12461 particular extension is supported.
12465 * Modify the behaviour of EVP cipher functions in similar way to digests
12466 to retain compatibility with existing code.
12470 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12471 compatibility with existing code. In particular the 'ctx' parameter does
12472 not have to be to be initialized before the call to EVP_DigestInit() and
12473 it is tidied up after a call to EVP_DigestFinal(). New function
12474 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12475 EVP_MD_CTX_copy() changed to not require the destination to be
12476 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12477 requires the destination to be valid.
12479 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12480 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12484 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12485 so that complete 'Handshake' protocol structures are kept in memory
12486 instead of overwriting 'msg_type' and 'length' with 'body' data.
12490 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12492 *Massimo Santin via Richard Levitte*
12494 * Major restructuring to the underlying ENGINE code. This includes
12495 reduction of linker bloat, separation of pure "ENGINE" manipulation
12496 (initialisation, etc) from functionality dealing with implementations
12497 of specific crypto interfaces. This change also introduces integrated
12498 support for symmetric ciphers and digest implementations - so ENGINEs
12499 can now accelerate these by providing EVP_CIPHER and EVP_MD
12500 implementations of their own. This is detailed in
12501 [crypto/engine/README.md](crypto/engine/README.md)
12502 as it couldn't be adequately described here. However, there are a few
12503 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12504 were changed in the original introduction of ENGINE code have now
12505 reverted back - the hooking from this code to ENGINE is now a good
12506 deal more passive and at run-time, operations deal directly with
12507 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12508 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12509 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12510 they were not being used by the framework as there is no concept of a
12511 BIGNUM_METHOD and they could not be generalised to the new
12512 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12513 ENGINE_cpy() has been removed as it cannot be consistently defined in
12518 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12522 * Change mkdef.pl to sort symbols that get the same entry number,
12523 and make sure the automatically generated functions `ERR_load_*`
12524 become part of libeay.num as well.
12528 * New function SSL_renegotiate_pending(). This returns true once
12529 renegotiation has been requested (either SSL_renegotiate() call
12530 or HelloRequest/ClientHello received from the peer) and becomes
12531 false once a handshake has been completed.
12532 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12533 sends a HelloRequest, but does not ensure that a handshake takes
12534 place. SSL_renegotiate_pending() is useful for checking if the
12535 client has followed the request.)
12539 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12540 By default, clients may request session resumption even during
12541 renegotiation (if session ID contexts permit); with this option,
12542 session resumption is possible only in the first handshake.
12544 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12545 more bits available for options that should not be part of
12546 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12550 * Add some demos for certificate and certificate request creation.
12554 * Make maximum certificate chain size accepted from the peer application
12555 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12556 "Douglas E. Engert" <deengert@anl.gov>.
12560 * Add support for shared libraries for Unixware-7
12561 (Boyd Lynn Gerber <gerberb@zenez.com>).
12565 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12566 be done prior to destruction. Use this to unload error strings from
12567 ENGINEs that load their own error strings. NB: This adds two new API
12568 functions to "get" and "set" this destroy handler in an ENGINE.
12572 * Alter all existing ENGINE implementations (except "openssl" and
12573 "openbsd") to dynamically instantiate their own error strings. This
12574 makes them more flexible to be built both as statically-linked ENGINEs
12575 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12576 Also, add stub code to each that makes building them as self-contained
12577 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12581 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12582 implementations into applications that are completely implemented in
12583 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12584 commands that can be used to configure what shared-library to load and
12585 to control aspects of the way it is handled. Also, made an update to
12586 the [README-Engine.md](README-Engine.md) file
12587 that brings its information up-to-date and
12588 provides some information and instructions on the "dynamic" ENGINE
12589 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12593 * Make it possible to unload ranges of ERR strings with a new
12594 "ERR_unload_strings" function.
12598 * Add a copy() function to EVP_MD.
12602 * Make EVP_MD routines take a context pointer instead of just the
12603 md_data void pointer.
12607 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12608 that the digest can only process a single chunk of data
12609 (typically because it is provided by a piece of
12610 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12611 is only going to provide a single chunk of data, and hence the
12612 framework needn't accumulate the data for oneshot drivers.
12616 * As with "ERR", make it possible to replace the underlying "ex_data"
12617 functions. This change also alters the storage and management of global
12618 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12619 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12620 index counters. The API functions that use this state have been changed
12621 to take a "class_index" rather than pointers to the class's local STACK
12622 and counter, and there is now an API function to dynamically create new
12623 classes. This centralisation allows us to (a) plug a lot of the
12624 thread-safety problems that existed, and (b) makes it possible to clean
12625 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12626 such data would previously have always leaked in application code and
12627 workarounds were in place to make the memory debugging turn a blind eye
12628 to it. Application code that doesn't use this new function will still
12629 leak as before, but their memory debugging output will announce it now
12630 rather than letting it slide.
12632 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12633 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12634 has a return value to indicate success or failure.
12638 * Make it possible to replace the underlying "ERR" functions such that the
12639 global state (2 LHASH tables and 2 locks) is only used by the "default"
12640 implementation. This change also adds two functions to "get" and "set"
12641 the implementation prior to it being automatically set the first time
12642 any other ERR function takes place. Ie. an application can call "get",
12643 pass the return value to a module it has just loaded, and that module
12644 can call its own "set" function using that value. This means the
12645 module's "ERR" operations will use (and modify) the error state in the
12646 application and not in its own statically linked copy of OpenSSL code.
12650 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12651 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12652 the operation, and provides a more encapsulated way for external code
12653 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12654 to use these functions rather than manually incrementing the counts.
12656 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12660 * Add EVP test program.
12664 * Add symmetric cipher support to ENGINE. Expect the API to change!
12668 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12669 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12670 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12671 These allow a CRL to be built without having to access X509_CRL fields
12672 directly. Modify 'ca' application to use new functions.
12676 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12677 bug workarounds. Rollback attack detection is a security feature.
12678 The problem will only arise on OpenSSL servers when TLSv1 is not
12679 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12680 Software authors not wanting to support TLSv1 will have special reasons
12681 for their choice and can explicitly enable this option.
12683 *Bodo Moeller, Lutz Jaenicke*
12685 * Rationalise EVP so it can be extended: don't include a union of
12686 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12687 (similar to those existing for EVP_CIPHER_CTX).
12692 EVP_MD_CTX_init(&md); /* new function call */
12693 EVP_DigestInit(&md, EVP_sha1());
12694 EVP_DigestUpdate(&md, in, len);
12695 EVP_DigestFinal(&md, out, NULL);
12696 EVP_MD_CTX_cleanup(&md); /* new function call */
12700 * Make DES key schedule conform to the usual scheme, as well as
12701 correcting its structure. This means that calls to DES functions
12702 now have to pass a pointer to a des_key_schedule instead of a
12703 plain des_key_schedule (which was actually always a pointer
12706 des_key_schedule ks;
12708 des_set_key_checked(..., &ks);
12709 des_ncbc_encrypt(..., &ks, ...);
12711 (Note that a later change renames 'des_...' into 'DES_...'.)
12715 * Initial reduction of linker bloat: the use of some functions, such as
12716 PEM causes large amounts of unused functions to be linked in due to
12717 poor organisation. For example pem_all.c contains every PEM function
12718 which has a knock on effect of linking in large amounts of (unused)
12719 ASN1 code. Grouping together similar functions and splitting unrelated
12720 functions prevents this.
12724 * Cleanup of EVP macros.
12728 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12729 correct `_ecb suffix`.
12733 * Add initial OCSP responder support to ocsp application. The
12734 revocation information is handled using the text based index
12735 use by the ca application. The responder can either handle
12736 requests generated internally, supplied in files (for example
12737 via a CGI script) or using an internal minimal server.
12741 * Add configuration choices to get zlib compression for TLS.
12745 * Changes to Kerberos SSL for RFC 2712 compliance:
12746 1. Implemented real KerberosWrapper, instead of just using
12747 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12748 2. Implemented optional authenticator field of KerberosWrapper.
12750 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12751 and authenticator structs; see crypto/krb5/.
12753 Generalized Kerberos calls to support multiple Kerberos libraries.
12754 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12755 via Richard Levitte*
12757 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12758 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12759 values for each of the key sizes rather than having just
12760 parameters (and 'speed' generating keys each time).
12764 * Speed up EVP routines.
12767 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12768 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12769 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12770 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12772 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12773 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12774 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12777 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12779 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12783 * Added the OS2-EMX target.
12785 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12787 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12788 New functions to support `NCONF` routines in extension code.
12789 New function `CONF_set_nconf()`
12790 to allow functions which take an `NCONF` to also handle the old `LHASH`
12791 structure: this means that the old `CONF` compatible routines can be
12792 retained (in particular w.rt. extensions) without having to duplicate the
12793 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12797 * Enhance the general user interface with mechanisms for inner control
12798 and with possibilities to have yes/no kind of prompts.
12802 * Change all calls to low-level digest routines in the library and
12803 applications to use EVP. Add missing calls to HMAC_cleanup() and
12804 don't assume HMAC_CTX can be copied using memcpy().
12806 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12808 * Add the possibility to control engines through control names but with
12809 arbitrary arguments instead of just a string.
12810 Change the key loaders to take a UI_METHOD instead of a callback
12811 function pointer. NOTE: this breaks binary compatibility with earlier
12812 versions of OpenSSL [engine].
12813 Adapt the nCipher code for these new conditions and add a card insertion
12818 * Enhance the general user interface with mechanisms to better support
12819 dialog box interfaces, application-defined prompts, the possibility
12820 to use defaults (for example default passwords from somewhere else)
12821 and interrupts/cancellations.
12825 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12826 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12830 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12831 tidy up some unnecessarily weird code in 'sk_new()').
12833 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12835 * Change the key loading routines for ENGINEs to use the same kind
12836 callback (pem_password_cb) as all other routines that need this
12841 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12842 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12843 than this minimum value is recommended.
12847 * New random seeder for OpenVMS, using the system process statistics
12848 that are easily reachable.
12852 * Windows apparently can't transparently handle global
12853 variables defined in DLLs. Initialisations such as:
12855 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12857 won't compile. This is used by the any applications that need to
12858 declare their own ASN1 modules. This was fixed by adding the option
12859 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12860 needed for static libraries under Win32.
12864 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12865 setting of purpose and trust fields. New X509_STORE trust and
12866 purpose functions and tidy up setting in other SSL functions.
12870 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12871 structure. These are inherited by X509_STORE_CTX when it is
12872 initialised. This allows various defaults to be set in the
12873 X509_STORE structure (such as flags for CRL checking and custom
12874 purpose or trust settings) for functions which only use X509_STORE_CTX
12875 internally such as S/MIME.
12877 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12878 trust settings if they are not set in X509_STORE. This allows X509_STORE
12879 purposes and trust (in S/MIME for example) to override any set by default.
12881 Add command line options for CRL checking to smime, s_client and s_server
12886 * Initial CRL based revocation checking. If the CRL checking flag(s)
12887 are set then the CRL is looked up in the X509_STORE structure and
12888 its validity and signature checked, then if the certificate is found
12889 in the CRL the verify fails with a revoked error.
12891 Various new CRL related callbacks added to X509_STORE_CTX structure.
12893 Command line options added to 'verify' application to support this.
12895 This needs some additional work, such as being able to handle multiple
12896 CRLs with different times, extension based lookup (rather than just
12897 by subject name) and ultimately more complete V2 CRL extension
12902 * Add a general user interface API (crypto/ui/). This is designed
12903 to replace things like des_read_password and friends (backward
12904 compatibility functions using this new API are provided).
12905 The purpose is to remove prompting functions from the DES code
12906 section as well as provide for prompting through dialog boxes in
12907 a window system and the like.
12911 * Add "ex_data" support to ENGINE so implementations can add state at a
12912 per-structure level rather than having to store it globally.
12916 * Make it possible for ENGINE structures to be copied when retrieved by
12917 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12918 This causes the "original" ENGINE structure to act like a template,
12919 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12920 operational state can be localised to each ENGINE structure, despite the
12921 fact they all share the same "methods". New ENGINE structures returned in
12922 this case have no functional references and the return value is the single
12923 structural reference. This matches the single structural reference returned
12924 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12929 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12930 needs to match any other type at all we need to manually clear the
12935 * Changes to the "openssl engine" utility to include;
12936 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12937 about an ENGINE's available control commands.
12938 - executing control commands from command line arguments using the
12939 '-pre' and '-post' switches. '-post' is only used if '-t' is
12940 specified and the ENGINE is successfully initialised. The syntax for
12941 the individual commands are colon-separated, for example;
12942 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12946 * New dynamic control command support for ENGINEs. ENGINEs can now
12947 declare their own commands (numbers), names (strings), descriptions,
12948 and input types for run-time discovery by calling applications. A
12949 subset of these commands are implicitly classed as "executable"
12950 depending on their input type, and only these can be invoked through
12951 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12952 can be based on user input, config files, etc). The distinction is
12953 that "executable" commands cannot return anything other than a boolean
12954 result and can only support numeric or string input, whereas some
12955 discoverable commands may only be for direct use through
12956 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12957 pointers, or other custom uses. The "executable" commands are to
12958 support parameterisations of ENGINE behaviour that can be
12959 unambiguously defined by ENGINEs and used consistently across any
12960 OpenSSL-based application. Commands have been added to all the
12961 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12962 control over shared-library paths without source code alterations.
12966 * Changed all ENGINE implementations to dynamically allocate their
12967 ENGINEs rather than declaring them statically. Apart from this being
12968 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12969 this also allows the implementations to compile without using the
12970 internal engine_int.h header.
12974 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12975 'const' value. Any code that should be able to modify a RAND_METHOD
12976 should already have non-const pointers to it (ie. they should only
12977 modify their own ones).
12981 * Made a variety of little tweaks to the ENGINE code.
12982 - "atalla" and "ubsec" string definitions were moved from header files
12983 to C code. "nuron" string definitions were placed in variables
12984 rather than hard-coded - allowing parameterisation of these values
12985 later on via ctrl() commands.
12986 - Removed unused "#if 0"'d code.
12987 - Fixed engine list iteration code so it uses ENGINE_free() to release
12988 structural references.
12989 - Constified the RAND_METHOD element of ENGINE structures.
12990 - Constified various get/set functions as appropriate and added
12991 missing functions (including a catch-all ENGINE_cpy that duplicates
12992 all ENGINE values onto a new ENGINE except reference counts/state).
12993 - Removed NULL parameter checks in get/set functions. Setting a method
12994 or function to NULL is a way of cancelling out a previously set
12995 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12996 and doesn't justify the extra error symbols and code.
12997 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12998 flags from engine_int.h to engine.h.
12999 - Changed prototypes for ENGINE handler functions (init(), finish(),
13000 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13004 * Implement binary inversion algorithm for BN_mod_inverse in addition
13005 to the algorithm using long division. The binary algorithm can be
13006 used only if the modulus is odd. On 32-bit systems, it is faster
13007 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13008 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13009 up to 450 bits. In 64-bit environments, the binary algorithm
13010 appears to be advantageous for much longer moduli; here we use it
13011 for moduli up to 2048 bits.
13015 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13016 could not support the combine flag in choice fields.
13020 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13021 extensions from a certificate request to the certificate.
13025 * Allow multiple 'certopt' and 'nameopt' options to be separated
13026 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13027 file: this allows the display of the certificate about to be
13028 signed to be customised, to allow certain fields to be included
13029 or excluded and extension details. The old system didn't display
13030 multicharacter strings properly, omitted fields not in the policy
13031 and couldn't display additional details such as extensions.
13035 * Function EC_POINTs_mul for multiple scalar multiplication
13036 of an arbitrary number of elliptic curve points
13037 \sum scalars[i]*points[i],
13038 optionally including the generator defined for the EC_GROUP:
13039 scalar*generator + \sum scalars[i]*points[i].
13041 EC_POINT_mul is a simple wrapper function for the typical case
13042 that the point list has just one item (besides the optional
13047 * First EC_METHODs for curves over GF(p):
13049 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13050 operations and provides various method functions that can also
13051 operate with faster implementations of modular arithmetic.
13053 EC_GFp_mont_method() reuses most functions that are part of
13054 EC_GFp_simple_method, but uses Montgomery arithmetic.
13056 *Bodo Moeller; point addition and point doubling
13057 implementation directly derived from source code provided by
13058 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13060 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13061 crypto/ec/ec_lib.c):
13063 Curves are EC_GROUP objects (with an optional group generator)
13064 based on EC_METHODs that are built into the library.
13066 Points are EC_POINT objects based on EC_GROUP objects.
13068 Most of the framework would be able to handle curves over arbitrary
13069 finite fields, but as there are no obvious types for fields other
13070 than GF(p), some functions are limited to that for now.
13074 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13075 that the file contains a complete HTTP response.
13079 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13080 change the def and num file printf format specifier from "%-40sXXX"
13081 to "%-39s XXX". The latter will always guarantee a space after the
13082 field while the former will cause them to run together if the field
13083 is 40 of more characters long.
13087 * Constify the cipher and digest 'method' functions and structures
13088 and modify related functions to take constant EVP_MD and EVP_CIPHER
13093 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13094 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13098 * Modify `EVP_Digest*()` routines so they now return values. Although the
13099 internal software routines can never fail additional hardware versions
13104 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13106 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13107 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13110 ERR_R_NESTED_ASN1_ERROR
13112 ERR_R_MISSING_ASN1_EOS
13113 were 4 .. 9, conflicting with
13114 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13116 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13117 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13119 Add new error code 'ERR_R_INTERNAL_ERROR'.
13123 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13128 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13129 sets the subject name for a new request or supersedes the
13130 subject name in a given request. Formats that can be parsed are
13131 'CN=Some Name, OU=myOU, C=IT'
13133 'CN=Some Name/OU=myOU/C=IT'.
13135 Add options '-batch' and '-verbose' to 'openssl req'.
13137 *Massimiliano Pala <madwolf@hackmasters.net>*
13139 * Introduce the possibility to access global variables through
13140 functions on platform were that's the best way to handle exporting
13141 global variables in shared libraries. To enable this functionality,
13142 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13143 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13144 is normally done by Configure or something similar).
13146 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13147 in the source file (foo.c) like this:
13149 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13150 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13152 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13153 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13155 OPENSSL_DECLARE_GLOBAL(int,foo);
13156 #define foo OPENSSL_GLOBAL_REF(foo)
13157 OPENSSL_DECLARE_GLOBAL(double,bar);
13158 #define bar OPENSSL_GLOBAL_REF(bar)
13160 The #defines are very important, and therefore so is including the
13161 header file everywhere where the defined globals are used.
13163 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13164 of ASN.1 items, but that structure is a bit different.
13166 The largest change is in util/mkdef.pl which has been enhanced with
13167 better and easier to understand logic to choose which symbols should
13168 go into the Windows .def files as well as a number of fixes and code
13169 cleanup (among others, algorithm keywords are now sorted
13170 lexicographically to avoid constant rewrites).
13174 * In BN_div() keep a copy of the sign of 'num' before writing the
13175 result to 'rm' because if rm==num the value will be overwritten
13176 and produce the wrong result if 'num' is negative: this caused
13177 problems with BN_mod() and BN_nnmod().
13181 * Function OCSP_request_verify(). This checks the signature on an
13182 OCSP request and verifies the signer certificate. The signer
13183 certificate is just checked for a generic purpose and OCSP request
13188 * Add OCSP_check_validity() function to check the validity of OCSP
13189 responses. OCSP responses are prepared in real time and may only
13190 be a few seconds old. Simply checking that the current time lies
13191 between thisUpdate and nextUpdate max reject otherwise valid responses
13192 caused by either OCSP responder or client clock inaccuracy. Instead
13193 we allow thisUpdate and nextUpdate to fall within a certain period of
13194 the current time. The age of the response can also optionally be
13195 checked. Two new options -validity_period and -status_age added to
13200 * If signature or public key algorithm is unrecognized print out its
13201 OID rather that just UNKNOWN.
13205 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13206 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13207 ID to be generated from the issuer certificate alone which can then be
13208 passed to OCSP_id_issuer_cmp().
13212 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13213 ASN1 modules to export functions returning ASN1_ITEM pointers
13214 instead of the ASN1_ITEM structures themselves. This adds several
13215 new macros which allow the underlying ASN1 function/structure to
13216 be accessed transparently. As a result code should not use ASN1_ITEM
13217 references directly (such as &X509_it) but instead use the relevant
13218 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13219 use of the new ASN1 code on platforms where exporting structures
13220 is problematical (for example in shared libraries) but exporting
13221 functions returning pointers to structures is not.
13225 * Add support for overriding the generation of SSL/TLS session IDs.
13226 These callbacks can be registered either in an SSL_CTX or per SSL.
13227 The purpose of this is to allow applications to control, if they wish,
13228 the arbitrary values chosen for use as session IDs, particularly as it
13229 can be useful for session caching in multiple-server environments. A
13230 command-line switch for testing this (and any client code that wishes
13231 to use such a feature) has been added to "s_server".
13233 *Geoff Thorpe, Lutz Jaenicke*
13235 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13236 of the form `#if defined(...) || defined(...) || ...` and
13237 `#if !defined(...) && !defined(...) && ...`. This also avoids
13238 the growing number of special cases it was previously handling.
13242 * Make all configuration macros available for application by making
13243 sure they are available in opensslconf.h, by giving them names starting
13244 with `OPENSSL_` to avoid conflicts with other packages and by making
13245 sure e_os2.h will cover all platform-specific cases together with
13247 Additionally, it is now possible to define configuration/platform-
13248 specific names (called "system identities"). In the C code, these
13249 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13250 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13251 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13256 * New option -set_serial to 'req' and 'x509' this allows the serial
13257 number to use to be specified on the command line. Previously self
13258 signed certificates were hard coded with serial number 0 and the
13259 CA options of 'x509' had to use a serial number in a file which was
13264 * New options to 'ca' utility to support V2 CRL entry extensions.
13265 Currently CRL reason, invalidity date and hold instruction are
13266 supported. Add new CRL extensions to V3 code and some new objects.
13270 * New function EVP_CIPHER_CTX_set_padding() this is used to
13271 disable standard block padding (aka PKCS#5 padding) in the EVP
13272 API, which was previously mandatory. This means that the data is
13273 not padded in any way and so the total length much be a multiple
13274 of the block size, otherwise an error occurs.
13278 * Initial (incomplete) OCSP SSL support.
13282 * New function OCSP_parse_url(). This splits up a URL into its host,
13283 port and path components: primarily to parse OCSP URLs. New -url
13284 option to ocsp utility.
13288 * New nonce behavior. The return value of OCSP_check_nonce() now
13289 reflects the various checks performed. Applications can decide
13290 whether to tolerate certain situations such as an absent nonce
13291 in a response when one was present in a request: the ocsp application
13292 just prints out a warning. New function OCSP_add1_basic_nonce()
13293 this is to allow responders to include a nonce in a response even if
13294 the request is nonce-less.
13298 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13299 skipped when using openssl x509 multiple times on a single input file,
13300 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13304 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13305 set string type: to handle setting ASN1_TIME structures. Fix ca
13306 utility to correctly initialize revocation date of CRLs.
13310 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13311 the clients preferred ciphersuites and rather use its own preferences.
13312 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13313 Internet Explorer by ensuring unchanged hash method during stepup.
13314 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13318 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13319 to aes and add a new 'exist' option to print out symbols that don't
13324 * Additional options to ocsp utility to allow flags to be set and
13325 additional certificates supplied.
13329 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13330 OCSP client a number of certificate to only verify the response
13335 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13336 handle the new API. Currently only ECB, CBC modes supported. Add new
13339 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13340 Encryption Standard (AES) Ciphersuites for Transport Layer
13341 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13342 not enabled by default and were not part of the "ALL" ciphersuite
13343 alias because they were not yet official; they could be
13344 explicitly requested by specifying the "AESdraft" ciphersuite
13345 group alias. In the final release of OpenSSL 0.9.7, the group
13346 alias is called "AES" and is part of "ALL".)
13348 *Ben Laurie, Steve Henson, Bodo Moeller*
13350 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13351 request to response.
13355 * Functions for OCSP responders. OCSP_request_onereq_count(),
13356 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13357 extract information from a certificate request. OCSP_response_create()
13358 creates a response and optionally adds a basic response structure.
13359 OCSP_basic_add1_status() adds a complete single response to a basic
13360 response and returns the OCSP_SINGLERESP structure just added (to allow
13361 extensions to be included for example). OCSP_basic_add1_cert() adds a
13362 certificate to a basic response and OCSP_basic_sign() signs a basic
13363 response with various flags. New helper functions ASN1_TIME_check()
13364 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13365 (converts ASN1_TIME to GeneralizedTime).
13369 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13370 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13371 structure from a certificate. X509_pubkey_digest() digests the public_key
13372 contents: this is used in various key identifiers.
13376 * Make sk_sort() tolerate a NULL argument.
13378 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13380 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13381 passed by the function are trusted implicitly. If any of them signed the
13382 response then it is assumed to be valid and is not verified.
13386 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13387 to data. This was previously part of the PKCS7 ASN1 code. This
13388 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13389 *Steve Henson, reported by Kenneth R. Robinette
13390 <support@securenetterm.com>*
13392 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13393 routines: without these tracing memory leaks is very painful.
13394 Fix leaks in PKCS12 and PKCS7 routines.
13398 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13399 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13400 effectively meant GeneralizedTime would never be used. Now it
13401 is initialised to -1 but X509_time_adj() now has to check the value
13402 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13403 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13404 *Steve Henson, reported by Kenneth R. Robinette
13405 <support@securenetterm.com>*
13407 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13408 result in a zero length in the ASN1_INTEGER structure which was
13409 not consistent with the structure when d2i_ASN1_INTEGER() was used
13410 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13411 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13412 where it did not print out a minus for negative ASN1_INTEGER.
13416 * Add summary printout to ocsp utility. The various functions which
13417 convert status values to strings have been renamed to:
13418 OCSP_response_status_str(), OCSP_cert_status_str() and
13419 OCSP_crl_reason_str() and are no longer static. New options
13420 to verify nonce values and to disable verification. OCSP response
13421 printout format cleaned up.
13425 * Add additional OCSP certificate checks. These are those specified
13426 in RFC2560. This consists of two separate checks: the CA of the
13427 certificate being checked must either be the OCSP signer certificate
13428 or the issuer of the OCSP signer certificate. In the latter case the
13429 OCSP signer certificate must contain the OCSP signing extended key
13430 usage. This check is performed by attempting to match the OCSP
13431 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13432 in the OCSP_CERTID structures of the response.
13436 * Initial OCSP certificate verification added to OCSP_basic_verify()
13437 and related routines. This uses the standard OpenSSL certificate
13438 verify routines to perform initial checks (just CA validity) and
13439 to obtain the certificate chain. Then additional checks will be
13440 performed on the chain. Currently the root CA is checked to see
13441 if it is explicitly trusted for OCSP signing. This is used to set
13442 a root CA as a global signing root: that is any certificate that
13443 chains to that CA is an acceptable OCSP signing certificate.
13447 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13448 extensions from a separate configuration file.
13449 As when reading extensions from the main configuration file,
13450 the '-extensions ...' option may be used for specifying the
13453 *Massimiliano Pala <madwolf@comune.modena.it>*
13455 * New OCSP utility. Allows OCSP requests to be generated or
13456 read. The request can be sent to a responder and the output
13457 parsed, outputted or printed in text form. Not complete yet:
13458 still needs to check the OCSP response validity.
13462 * New subcommands for 'openssl ca':
13463 `openssl ca -status <serial>` prints the status of the cert with
13464 the given serial number (according to the index file).
13465 `openssl ca -updatedb` updates the expiry status of certificates
13468 *Massimiliano Pala <madwolf@comune.modena.it>*
13470 * New '-newreq-nodes' command option to CA.pl. This is like
13471 '-newreq', but calls 'openssl req' with the '-nodes' option
13472 so that the resulting key is not encrypted.
13474 *Damien Miller <djm@mindrot.org>*
13476 * New configuration for the GNU Hurd.
13478 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13480 * Initial code to implement OCSP basic response verify. This
13481 is currently incomplete. Currently just finds the signer's
13482 certificate and verifies the signature on the response.
13486 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13487 value of OPENSSLDIR. This is available via the new '-d' option
13488 to 'openssl version', and is also included in 'openssl version -a'.
13492 * Allowing defining memory allocation callbacks that will be given
13493 file name and line number information in additional arguments
13494 (a `const char*` and an int). The basic functionality remains, as
13495 well as the original possibility to just replace malloc(),
13496 realloc() and free() by functions that do not know about these
13497 additional arguments. To register and find out the current
13498 settings for extended allocation functions, the following
13499 functions are provided:
13501 CRYPTO_set_mem_ex_functions
13502 CRYPTO_set_locked_mem_ex_functions
13503 CRYPTO_get_mem_ex_functions
13504 CRYPTO_get_locked_mem_ex_functions
13506 These work the same way as CRYPTO_set_mem_functions and friends.
13507 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13508 extended allocation function is enabled.
13509 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13510 a conventional allocation function is enabled.
13512 *Richard Levitte, Bodo Moeller*
13514 * Finish off removing the remaining LHASH function pointer casts.
13515 There should no longer be any prototype-casting required when using
13516 the LHASH abstraction, and any casts that remain are "bugs". See
13517 the callback types and macros at the head of lhash.h for details
13518 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13522 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13523 If /dev/[u]random devices are not available or do not return enough
13524 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13526 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13527 /etc/entropy will be queried once each in this sequence, querying stops
13528 when enough entropy was collected without querying more sockets.
13532 * Change the Unix RAND_poll() variant to be able to poll several
13533 random devices, as specified by DEVRANDOM, until a sufficient amount
13534 of data has been collected. We spend at most 10 ms on each file
13535 (select timeout) and read in non-blocking mode. DEVRANDOM now
13536 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13537 (previously it was just the string "/dev/urandom"), so on typical
13538 platforms the 10 ms delay will never occur.
13539 Also separate out the Unix variant to its own file, rand_unix.c.
13540 For VMS, there's a currently-empty rand_vms.c.
13544 * Move OCSP client related routines to ocsp_cl.c. These
13545 provide utility functions which an application needing
13546 to issue a request to an OCSP responder and analyse the
13547 response will typically need: as opposed to those which an
13548 OCSP responder itself would need which will be added later.
13550 OCSP_request_sign() signs an OCSP request with an API similar
13551 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13552 response. OCSP_response_get1_basic() extracts basic response
13553 from response. OCSP_resp_find_status(): finds and extracts status
13554 information from an OCSP_CERTID structure (which will be created
13555 when the request structure is built). These are built from lower
13556 level functions which work on OCSP_SINGLERESP structures but
13557 won't normally be used unless the application wishes to examine
13558 extensions in the OCSP response for example.
13560 Replace nonce routines with a pair of functions.
13561 OCSP_request_add1_nonce() adds a nonce value and optionally
13562 generates a random value. OCSP_check_nonce() checks the
13563 validity of the nonce in an OCSP response.
13567 * Change function OCSP_request_add() to OCSP_request_add0_id().
13568 This doesn't copy the supplied OCSP_CERTID and avoids the
13569 need to free up the newly created id. Change return type
13570 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13571 This can then be used to add extensions to the request.
13572 Deleted OCSP_request_new(), since most of its functionality
13573 is now in OCSP_REQUEST_new() (and the case insensitive name
13574 clash) apart from the ability to set the request name which
13575 will be added elsewhere.
13579 * Update OCSP API. Remove obsolete extensions argument from
13580 various functions. Extensions are now handled using the new
13581 OCSP extension code. New simple OCSP HTTP function which
13582 can be used to send requests and parse the response.
13586 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13587 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13588 uses the special reorder version of SET OF to sort the attributes
13589 and reorder them to match the encoded order. This resolves a long
13590 standing problem: a verify on a PKCS7 structure just after signing
13591 it used to fail because the attribute order did not match the
13592 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13593 it uses the received order. This is necessary to tolerate some broken
13594 software that does not order SET OF. This is handled by encoding
13595 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13596 to produce the required SET OF.
13600 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13601 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13602 files to get correct declarations of the ASN.1 item variables.
13606 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13607 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13608 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13609 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13610 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13611 ASN1_ITEM and no wrapper functions.
13615 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13616 replace the old function pointer based I/O routines. Change most of
13617 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13621 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13622 lines, recognize more "algorithms" that can be deselected, and make
13623 it complain about algorithm deselection that isn't recognised.
13627 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13628 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13629 to use new functions. Add NO_ASN1_OLD which can be set to remove
13630 some old style ASN1 functions: this can be used to determine if old
13631 code will still work when these eventually go away.
13635 * New extension functions for OCSP structures, these follow the
13636 same conventions as certificates and CRLs.
13640 * New function X509V3_add1_i2d(). This automatically encodes and
13641 adds an extension. Its behaviour can be customised with various
13642 flags to append, replace or delete. Various wrappers added for
13643 certificates and CRLs.
13647 * Fix to avoid calling the underlying ASN1 print routine when
13648 an extension cannot be parsed. Correct a typo in the
13649 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13653 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13654 entries for variables.
13658 * Add functionality to `apps/openssl.c` for detecting locking
13659 problems: As the program is single-threaded, all we have
13660 to do is register a locking callback using an array for
13661 storing which locks are currently held by the program.
13665 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13666 SSL_get_ex_data_X509_STORE_idx(), which is used in
13667 ssl_verify_cert_chain() and thus can be called at any time
13668 during TLS/SSL handshakes so that thread-safety is essential.
13669 Unfortunately, the ex_data design is not at all suited
13670 for multi-threaded use, so it probably should be abolished.
13674 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13676 *Broadcom, tweaked and integrated by Geoff Thorpe*
13678 * Move common extension printing code to new function
13679 X509V3_print_extensions(). Reorganise OCSP print routines and
13680 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13684 * New function X509_signature_print() to remove duplication in some
13689 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13690 set (this was treated exactly the same as SET OF previously). This
13691 is used to reorder the STACK representing the structure to match the
13692 encoding. This will be used to get round a problem where a PKCS7
13693 structure which was signed could not be verified because the STACK
13694 order did not reflect the encoded order.
13698 * Reimplement the OCSP ASN1 module using the new code.
13702 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13703 for its ASN1 operations. The old style function pointers still exist
13704 for now but they will eventually go away.
13708 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13709 completely replaces the old ASN1 functionality with a table driven
13710 encoder and decoder which interprets an ASN1_ITEM structure describing
13711 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13712 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13713 has also been converted to the new form.
13717 * Change BN_mod_exp_recp so that negative moduli are tolerated
13718 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13719 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13720 for negative moduli.
13724 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13725 of not touching the result's sign bit.
13729 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13734 * Changed the LHASH code to use prototypes for callbacks, and created
13735 macros to declare and implement thin (optionally static) functions
13736 that provide type-safety and avoid function pointer casting for the
13737 type-specific callbacks.
13741 * Added Kerberos Cipher Suites to be used with TLS, as written in
13743 *Veers Staats <staatsvr@asc.hpc.mil>,
13744 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13746 * Reformat the FAQ so the different questions and answers can be divided
13747 in sections depending on the subject.
13751 * Have the zlib compression code load ZLIB.DLL dynamically under
13756 * New function BN_mod_sqrt for computing square roots modulo a prime
13757 (using the probabilistic Tonelli-Shanks algorithm unless
13758 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13759 be handled deterministically).
13761 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13763 * Make BN_mod_inverse faster by explicitly handling small quotients
13764 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13765 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13769 * New function BN_kronecker.
13773 * Fix BN_gcd so that it works on negative inputs; the result is
13774 positive unless both parameters are zero.
13775 Previously something reasonably close to an infinite loop was
13776 possible because numbers could be growing instead of shrinking
13777 in the implementation of Euclid's algorithm.
13781 * Fix BN_is_word() and BN_is_one() macros to take into account the
13782 sign of the number in question.
13784 Fix BN_is_word(a,w) to work correctly for w == 0.
13786 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13787 because its test if the absolute value of 'a' equals 'w'.
13788 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13789 it exists mostly for use in the implementations of BN_is_zero(),
13790 BN_is_one(), and BN_is_word().
13794 * New function BN_swap.
13798 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13799 the exponentiation functions are more likely to produce reasonable
13800 results on negative inputs.
13804 * Change BN_mod_mul so that the result is always non-negative.
13805 Previously, it could be negative if one of the factors was negative;
13806 I don't think anyone really wanted that behaviour.
13810 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13811 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13812 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13813 and add new functions:
13822 BN_mod_lshift1_quick
13824 BN_mod_lshift_quick
13826 These functions always generate non-negative results.
13828 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13829 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13831 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13832 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13833 be reduced modulo `m`.
13835 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13838 The following entry accidentally appeared in the CHANGES file
13839 distributed with OpenSSL 0.9.7. The modifications described in
13840 it do *not* apply to OpenSSL 0.9.7.
13842 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13843 was actually never needed) and in BN_mul(). The removal in BN_mul()
13844 required a small change in bn_mul_part_recursive() and the addition
13845 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13846 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13847 bn_sub_words() and bn_add_words() except they take arrays with
13853 * In 'openssl passwd', verify passwords read from the terminal
13854 unless the '-salt' option is used (which usually means that
13855 verification would just waste user's time since the resulting
13856 hash is going to be compared with some given password hash)
13857 or the new '-noverify' option is used.
13859 This is an incompatible change, but it does not affect
13860 non-interactive use of 'openssl passwd' (passwords on the command
13861 line, '-stdin' option, '-in ...' option) and thus should not
13862 cause any problems.
13866 * Remove all references to RSAref, since there's no more need for it.
13870 * Make DSO load along a path given through an environment variable
13871 (SHLIB_PATH) with shl_load().
13875 * Constify the ENGINE code as a result of BIGNUM constification.
13876 Also constify the RSA code and most things related to it. In a
13877 few places, most notable in the depth of the ASN.1 code, ugly
13878 casts back to non-const were required (to be solved at a later
13883 * Make it so the openssl application has all engines loaded by default.
13887 * Constify the BIGNUM routines a little more.
13891 * Add the following functions:
13893 ENGINE_load_cswift()
13895 ENGINE_load_atalla()
13896 ENGINE_load_nuron()
13897 ENGINE_load_builtin_engines()
13899 That way, an application can itself choose if external engines that
13900 are built-in in OpenSSL shall ever be used or not. The benefit is
13901 that applications won't have to be linked with libdl or other dso
13902 libraries unless it's really needed.
13904 Changed 'openssl engine' to load all engines on demand.
13905 Changed the engine header files to avoid the duplication of some
13906 declarations (they differed!).
13910 * 'openssl engine' can now list capabilities.
13914 * Better error reporting in 'openssl engine'.
13918 * Never call load_dh_param(NULL) in s_server.
13922 * Add engine application. It can currently list engines by name and
13923 identity, and test if they are actually available.
13927 * Improve RPM specification file by forcing symbolic linking and making
13928 sure the installed documentation is also owned by root.root.
13930 *Damien Miller <djm@mindrot.org>*
13932 * Give the OpenSSL applications more possibilities to make use of
13933 keys (public as well as private) handled by engines.
13937 * Add OCSP code that comes from CertCo.
13941 * Add VMS support for the Rijndael code.
13945 * Added untested support for Nuron crypto accelerator.
13949 * Add support for external cryptographic devices. This code was
13950 previously distributed separately as the "engine" branch.
13952 *Geoff Thorpe, Richard Levitte*
13954 * Rework the filename-translation in the DSO code. It is now possible to
13955 have far greater control over how a "name" is turned into a filename
13956 depending on the operating environment and any oddities about the
13957 different shared library filenames on each system.
13961 * Support threads on FreeBSD-elf in Configure.
13965 * Fix for SHA1 assembly problem with MASM: it produces
13966 warnings about corrupt line number information when assembling
13967 with debugging information. This is caused by the overlapping
13970 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13973 NCONF_get_number() has no error checking at all. As a replacement,
13974 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13975 promoted strongly. The old NCONF_get_number is kept around for
13976 binary backward compatibility.
13977 Make it possible for methods to load from something other than a BIO,
13978 by providing a function pointer that is given a name instead of a BIO.
13979 For example, this could be used to load configuration data from an
13984 * Fix for non blocking accept BIOs. Added new I/O special reason
13985 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13986 with non blocking I/O was not possible because no retry code was
13987 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13992 * Added the beginnings of Rijndael support.
13996 * Fix for bug in DirectoryString mask setting. Add support for
13997 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13998 to allow certificate printing to more controllable, additional
13999 'certopt' option to 'x509' to allow new printing options to be
14004 * Clean old EAY MD5 hack from e_os.h.
14008 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14010 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14011 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14013 *Joe Orton, Steve Henson*
14015 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14017 * Fix additional bug revealed by the NISCC test suite:
14019 Stop bug triggering large recursion when presented with
14020 certain ASN.1 tags ([CVE-2003-0851])
14024 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14026 * Fix various bugs revealed by running the NISCC test suite:
14028 Stop out of bounds reads in the ASN1 code when presented with
14029 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14031 If verify callback ignores invalid public key errors don't try to check
14032 certificate signature with the NULL public key.
14036 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14037 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14042 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14043 extra data after the compression methods not only for TLS 1.0
14044 but also for SSL 3.0 (as required by the specification).
14046 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14048 * Change X509_certificate_type() to mark the key as exported/exportable
14049 when it's 512 *bits* long, not 512 bytes.
14053 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14055 * Countermeasure against the Klima-Pokorny-Rosa extension of
14056 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14057 a protocol version number mismatch like a decryption error
14058 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14062 * Turn on RSA blinding by default in the default implementation
14063 to avoid a timing attack. Applications that don't want it can call
14064 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14065 They would be ill-advised to do so in most cases.
14067 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14069 * Change RSA blinding code so that it works when the PRNG is not
14070 seeded (in this case, the secret RSA exponent is abused as
14071 an unpredictable seed -- if it is not unpredictable, there
14072 is no point in blinding anyway). Make RSA blinding thread-safe
14073 by remembering the creator's thread ID in rsa->blinding and
14074 having all other threads use local one-time blinding factors
14075 (this requires more computation than sharing rsa->blinding, but
14076 avoids excessive locking; and if an RSA object is not shared
14077 between threads, blinding will still be very fast).
14081 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14083 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14084 via timing by performing a MAC computation even if incorrect
14085 block cipher padding has been found. This is a countermeasure
14086 against active attacks where the attacker has to distinguish
14087 between bad padding and a MAC verification error. ([CVE-2003-0078])
14089 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14090 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14091 Martin Vuagnoux (EPFL, Ilion)*
14093 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14095 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14096 memory from its contents. This is done with a counter that will
14097 place alternating values in each byte. This can be used to solve
14098 two issues: 1) the removal of calls to memset() by highly optimizing
14099 compilers, and 2) cleansing with other values than 0, since those can
14100 be read through on certain media, for example a swap space on disk.
14104 * Bugfix: client side session caching did not work with external caching,
14105 because the session->cipher setting was not restored when reloading
14106 from the external cache. This problem was masked, when
14107 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14108 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14112 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14113 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14115 *Zeev Lieber <zeev-l@yahoo.com>*
14117 * Undo an undocumented change introduced in 0.9.6e which caused
14118 repeated calls to OpenSSL_add_all_ciphers() and
14119 OpenSSL_add_all_digests() to be ignored, even after calling
14124 * Change the default configuration reader to deal with last line not
14125 being properly terminated.
14129 * Change X509_NAME_cmp() so it applies the special rules on handling
14130 DN values that are of type PrintableString, as well as RDNs of type
14131 emailAddress where the value has the type ia5String.
14133 *stefank@valicert.com via Richard Levitte*
14135 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14136 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14137 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14138 the bitwise-OR of the two for use by the majority of applications
14139 wanting this behaviour, and update the docs. The documented
14140 behaviour and actual behaviour were inconsistent and had been
14141 changing anyway, so this is more a bug-fix than a behavioural
14144 *Geoff Thorpe, diagnosed by Nadav Har'El*
14146 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14147 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14151 * Fix initialization code race conditions in
14152 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14153 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14154 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14155 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14156 ssl2_get_cipher_by_char(),
14157 ssl3_get_cipher_by_char().
14159 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14161 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14162 the cached sessions are flushed, as the remove_cb() might use ex_data
14163 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14164 (see [openssl.org #212]).
14166 *Geoff Thorpe, Lutz Jaenicke*
14168 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14169 length, instead of the encoding length to d2i_ASN1_OBJECT.
14173 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14175 * [In 0.9.6g-engine release:]
14176 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14178 *Lynn Gazis <lgazis@rainbow.com>*
14180 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14182 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14183 and get fix the header length calculation.
14184 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14185 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14187 * Use proper error handling instead of 'assertions' in buffer
14188 overflow checks added in 0.9.6e. This prevents DoS (the
14189 assertions could call abort()).
14191 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14193 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14195 * Add various sanity checks to asn1_get_length() to reject
14196 the ASN1 length bytes if they exceed sizeof(long), will appear
14197 negative or the content length exceeds the length of the
14200 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14202 * Fix cipher selection routines: ciphers without encryption had no flags
14203 for the cipher strength set and where therefore not handled correctly
14204 by the selection routines (PR #130).
14208 * Fix EVP_dsa_sha macro.
14213 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14214 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14215 that was added in OpenSSL 0.9.6d.
14217 As the countermeasure turned out to be incompatible with some
14218 broken SSL implementations, the new option is part of SSL_OP_ALL.
14219 SSL_OP_ALL is usually employed when compatibility with weird SSL
14220 implementations is desired (e.g. '-bugs' option to 's_client' and
14221 's_server'), so the new option is automatically set in many
14226 * Changes in security patch:
14228 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14229 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14230 Air Force Materiel Command, USAF, under agreement number
14233 * Add various sanity checks to asn1_get_length() to reject
14234 the ASN1 length bytes if they exceed sizeof(long), will appear
14235 negative or the content length exceeds the length of the
14236 supplied buffer. ([CVE-2002-0659])
14238 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14240 * Assertions for various potential buffer overflows, not known to
14241 happen in practice.
14243 *Ben Laurie (CHATS)*
14245 * Various temporary buffers to hold ASCII versions of integers were
14246 too small for 64 bit platforms. ([CVE-2002-0655])
14247 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14249 * Remote buffer overflow in SSL3 protocol - an attacker could
14250 supply an oversized session ID to a client. ([CVE-2002-0656])
14252 *Ben Laurie (CHATS)*
14254 * Remote buffer overflow in SSL2 protocol - an attacker could
14255 supply an oversized client master key. ([CVE-2002-0656])
14257 *Ben Laurie (CHATS)*
14259 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14261 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14262 encoded as NULL) with id-dsa-with-sha1.
14264 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14266 * Check various `X509_...()` return values in `apps/req.c`.
14268 *Nils Larsch <nla@trustcenter.de>*
14270 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14271 an end-of-file condition would erroneously be flagged, when the CRLF
14272 was just at the end of a processed block. The bug was discovered when
14273 processing data through a buffering memory BIO handing the data to a
14274 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14275 <ptsekov@syntrex.com> and Nedelcho Stanev.
14279 * Implement a countermeasure against a vulnerability recently found
14280 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14281 before application data chunks to avoid the use of known IVs
14282 with data potentially chosen by the attacker.
14286 * Fix length checks in ssl3_get_client_hello().
14290 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14291 to prevent ssl3_read_internal() from incorrectly assuming that
14292 ssl3_read_bytes() found application data while handshake
14293 processing was enabled when in fact s->s3->in_read_app_data was
14294 merely automatically cleared during the initial handshake.
14296 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14298 * Fix object definitions for Private and Enterprise: they were not
14299 recognized in their shortname (=lowercase) representation. Extend
14300 obj_dat.pl to issue an error when using undefined keywords instead
14301 of silently ignoring the problem (Svenning Sorensen
14302 <sss@sss.dnsalias.net>).
14306 * Fix DH_generate_parameters() so that it works for 'non-standard'
14307 generators, i.e. generators other than 2 and 5. (Previously, the
14308 code did not properly initialise the 'add' and 'rem' values to
14309 BN_generate_prime().)
14311 In the new general case, we do not insist that 'generator' is
14312 actually a primitive root: This requirement is rather pointless;
14313 a generator of the order-q subgroup is just as good, if not
14318 * Map new X509 verification errors to alerts. Discovered and submitted by
14319 Tom Wu <tom@arcot.com>.
14323 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14324 returning non-zero before the data has been completely received
14325 when using non-blocking I/O.
14327 *Bodo Moeller; problem pointed out by John Hughes*
14329 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14331 *Ben Laurie, Lutz Jaenicke*
14333 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14334 Yoram Zahavi <YoramZ@gilian.com>).
14338 * Add information about CygWin 1.3 and on, and preserve proper
14339 configuration for the versions before that.
14341 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14343 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14344 check whether we deal with a copy of a session and do not delete from
14345 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14346 <izhar@checkpoint.com>.
14350 * Do not store session data into the internal session cache, if it
14351 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14352 flag is set). Proposed by Aslam <aslam@funk.com>.
14356 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14361 * [In 0.9.6d-engine release:]
14362 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14364 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14366 * Add the configuration target linux-s390x.
14368 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14370 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14371 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14372 variable as an indication that a ClientHello message has been
14373 received. As the flag value will be lost between multiple
14374 invocations of ssl3_accept when using non-blocking I/O, the
14375 function may not be aware that a handshake has actually taken
14376 place, thus preventing a new session from being added to the
14379 To avoid this problem, we now set s->new_session to 2 instead of
14380 using a local variable.
14382 *Lutz Jaenicke, Bodo Moeller*
14384 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14385 if the SSL_R_LENGTH_MISMATCH error is detected.
14387 *Geoff Thorpe, Bodo Moeller*
14389 * New 'shared_ldflag' column in Configure platform table.
14393 * Fix EVP_CIPHER_mode macro.
14395 *"Dan S. Camper" <dan@bti.net>*
14397 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14398 type, we must throw them away by setting rr->length to 0.
14400 *D P Chang <dpc@qualys.com>*
14402 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14404 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14405 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14406 worked incorrectly for those cases where range = `10..._2` and
14407 `3*range` is two bits longer than range.)
14411 * Only add signing time to PKCS7 structures if it is not already
14416 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14417 OBJ_ld_ce should be OBJ_id_ce.
14418 Also some ip-pda OIDs in crypto/objects/objects.txt were
14419 incorrect (cf. RFC 3039).
14421 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14423 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14424 returns early because it has nothing to do.
14426 *Andy Schneider <andy.schneider@bjss.co.uk>*
14428 * [In 0.9.6c-engine release:]
14429 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14431 *Andy Schneider <andy.schneider@bjss.co.uk>*
14433 * [In 0.9.6c-engine release:]
14434 Add support for Cryptographic Appliance's keyserver technology.
14435 (Use engine 'keyclient')
14437 *Cryptographic Appliances and Geoff Thorpe*
14439 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14440 is called via tools/c89.sh because arguments have to be
14441 rearranged (all '-L' options must appear before the first object
14444 *Richard Shapiro <rshapiro@abinitio.com>*
14446 * [In 0.9.6c-engine release:]
14447 Add support for Broadcom crypto accelerator cards, backported
14450 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14452 * [In 0.9.6c-engine release:]
14453 Add support for SureWare crypto accelerator cards from
14454 Baltimore Technologies. (Use engine 'sureware')
14456 *Baltimore Technologies and Mark Cox*
14458 * [In 0.9.6c-engine release:]
14459 Add support for crypto accelerator cards from Accelerated
14460 Encryption Processing, www.aep.ie. (Use engine 'aep')
14462 *AEP Inc. and Mark Cox*
14464 * Add a configuration entry for gcc on UnixWare.
14466 *Gary Benson <gbenson@redhat.com>*
14468 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14469 messages are stored in a single piece (fixed-length part and
14470 variable-length part combined) and fix various bugs found on the way.
14474 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14475 instead. BIO_gethostbyname() does not know what timeouts are
14476 appropriate, so entries would stay in cache even when they have
14478 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14480 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14481 faced with a pathologically small ClientHello fragment that does
14482 not contain client_version: Instead of aborting with an error,
14483 simply choose the highest available protocol version (i.e.,
14484 TLS 1.0 unless it is disabled). In practice, ClientHello
14485 messages are never sent like this, but this change gives us
14486 strictly correct behaviour at least for TLS.
14490 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14491 never resets s->method to s->ctx->method when called from within
14492 one of the SSL handshake functions.
14494 *Bodo Moeller; problem pointed out by Niko Baric*
14496 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14497 (sent using the client's version number) if client_version is
14498 smaller than the protocol version in use. Also change
14499 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14500 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14501 the client will at least see that alert.
14505 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14510 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14511 client receives HelloRequest while in a handshake.
14513 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14515 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14516 should end in 'break', not 'goto end' which circumvents various
14517 cleanups done in state SSL_ST_OK. But session related stuff
14518 must be disabled for SSL_ST_OK in the case that we just sent a
14521 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14522 before just sending a HelloRequest.
14524 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14526 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14527 reveal whether illegal block cipher padding was found or a MAC
14528 verification error occurred. (Neither SSLerr() codes nor alerts
14529 are directly visible to potential attackers, but the information
14530 may leak via logfiles.)
14532 Similar changes are not required for the SSL 2.0 implementation
14533 because the number of padding bytes is sent in clear for SSL 2.0,
14534 and the extra bytes are just ignored. However ssl/s2_pkt.c
14535 failed to verify that the purported number of padding bytes is in
14540 * Add OpenUNIX-8 support including shared libraries
14541 (Boyd Lynn Gerber <gerberb@zenez.com>).
14545 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14546 'wristwatch attack' using huge encoding parameters (cf.
14547 James H. Manger's CRYPTO 2001 paper). Note that the
14548 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14549 encoding parameters and hence was not vulnerable.
14553 * BN_sqr() bug fix.
14555 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14557 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14558 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14559 followed by modular reduction.
14561 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14563 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14564 equivalent based on BN_pseudo_rand() instead of BN_rand().
14568 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14569 This function was broken, as the check for a new client hello message
14570 to handle SGC did not allow these large messages.
14571 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14575 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14579 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14580 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14584 * Rework the configuration and shared library support for Tru64 Unix.
14585 The configuration part makes use of modern compiler features and
14586 still retains old compiler behavior for those that run older versions
14587 of the OS. The shared library support part includes a variant that
14588 uses the RPATH feature, and is available through the special
14589 configuration target "alpha-cc-rpath", which will never be selected
14592 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14594 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14595 with the same message size as in ssl3_get_certificate_request().
14596 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14597 messages might inadvertently be reject as too long.
14599 *Petr Lampa <lampa@fee.vutbr.cz>*
14601 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14605 * Modified SSL library such that the verify_callback that has been set
14606 specifically for an SSL object with SSL_set_verify() is actually being
14607 used. Before the change, a verify_callback set with this function was
14608 ignored and the verify_callback() set in the SSL_CTX at the time of
14609 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14610 to allow the necessary settings.
14614 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14615 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14616 done automatically (in contradiction to the requirements of the C
14617 standard). This made problems when used from OpenSSH.
14621 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14622 dh->length and always used
14624 BN_rand_range(priv_key, dh->p).
14626 BN_rand_range() is not necessary for Diffie-Hellman, and this
14627 specific range makes Diffie-Hellman unnecessarily inefficient if
14628 dh->length (recommended exponent length) is much smaller than the
14629 length of dh->p. We could use BN_rand_range() if the order of
14630 the subgroup was stored in the DH structure, but we only have
14635 BN_rand(priv_key, l, ...)
14637 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14644 RSA_eay_public_encrypt
14645 RSA_eay_private_decrypt
14646 RSA_eay_private_encrypt (signing)
14647 RSA_eay_public_decrypt (signature verification)
14649 (default implementations for RSA_public_encrypt,
14650 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14651 always reject numbers >= n.
14655 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14656 to synchronize access to 'locking_thread'. This is necessary on
14657 systems where access to 'locking_thread' (an 'unsigned long'
14658 variable) is not atomic.
14662 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14663 *before* setting the 'crypto_lock_rand' flag. The previous code had
14664 a race condition if 0 is a valid thread ID.
14666 *Travis Vitek <vitek@roguewave.com>*
14668 * Add support for shared libraries under Irix.
14670 *Albert Chin-A-Young <china@thewrittenword.com>*
14672 * Add configuration option to build on Linux on both big-endian and
14673 little-endian MIPS.
14675 *Ralf Baechle <ralf@uni-koblenz.de>*
14677 * Add the possibility to create shared libraries on HP-UX.
14681 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14683 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14684 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14685 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14686 PRNG state recovery was possible based on the output of
14687 one PRNG request appropriately sized to gain knowledge on
14688 'md' followed by enough consecutive 1-byte PRNG requests
14689 to traverse all of 'state'.
14691 1. When updating 'md_local' (the current thread's copy of 'md')
14692 during PRNG output generation, hash all of the previous
14693 'md_local' value, not just the half used for PRNG output.
14695 2. Make the number of bytes from 'state' included into the hash
14696 independent from the number of PRNG bytes requested.
14698 The first measure alone would be sufficient to avoid
14699 Markku-Juhani's attack. (Actually it had never occurred
14700 to me that the half of 'md_local' used for chaining was the
14701 half from which PRNG output bytes were taken -- I had always
14702 assumed that the secret half would be used.) The second
14703 measure makes sure that additional data from 'state' is never
14704 mixed into 'md_local' in small portions; this heuristically
14705 further strengthens the PRNG.
14709 * Fix crypto/bn/asm/mips3.s.
14713 * When only the key is given to "enc", the IV is undefined. Print out
14714 an error message in this case.
14718 * Handle special case when X509_NAME is empty in X509 printing routines.
14722 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14723 positive and less than q.
14727 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14728 used: it isn't thread safe and the add_lock_callback should handle
14731 *Paul Rose <Paul.Rose@bridge.com>*
14733 * Verify that incoming data obeys the block size in
14734 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14740 *Ulf Möller, Bodo Möller*
14742 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14743 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14744 when fixing the server behaviour for backwards-compatible 'client
14745 hello' messages. (Note that the attack is impractical against
14746 SSL 3.0 and TLS 1.0 anyway because length and version checking
14747 means that the probability of guessing a valid ciphertext is
14748 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14751 Before 0.9.5, the countermeasure (hide the error by generating a
14752 random 'decryption result') did not work properly because
14753 ERR_clear_error() was missing, meaning that SSL_get_error() would
14754 detect the supposedly ignored error.
14756 Both problems are now fixed.
14760 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14761 (previously it was 1024).
14765 * Fix for compatibility mode trust settings: ignore trust settings
14766 unless some valid trust or reject settings are present.
14770 * Fix for blowfish EVP: its a variable length cipher.
14774 * Fix various bugs related to DSA S/MIME verification. Handle missing
14775 parameters in DSA public key structures and return an error in the
14776 DSA routines if parameters are absent.
14780 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14781 in the current directory if neither $RANDFILE nor $HOME was set.
14782 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14783 caused some confusion to Windows users who haven't defined $HOME.
14784 Thus RAND_file_name() is changed again: e_os.h can define a
14785 DEFAULT_HOME, which will be used if $HOME is not set.
14786 For Windows, we use "C:"; on other platforms, we still require
14787 environment variables.
14789 * Move 'if (!initialized) RAND_poll()' into regions protected by
14790 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14791 having multiple threads call RAND_poll() concurrently.
14795 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14796 combination of a flag and a thread ID variable.
14797 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14798 flag), *other* threads can enter ssleay_add_bytes without obeying
14799 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14800 that they do not hold after the first thread unsets add_do_not_lock).
14804 * Change bctest again: '-x' expressions are not available in all
14805 versions of 'test'.
14809 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14811 * Fix a couple of memory leaks in PKCS7_dataDecode()
14813 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14815 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14816 the default extension for executables, if any. Also, make the perl
14817 scripts that use symlink() to test if it really exists and use "cp"
14818 if it doesn't. All this made OpenSSL compilable and installable in
14823 * Fix for asn1_GetSequence() for indefinite length constructed data.
14824 If SEQUENCE is length is indefinite just set c->slen to the total
14825 amount of data available.
14827 *Steve Henson, reported by shige@FreeBSD.org*
14829 *This change does not apply to 0.9.7.*
14831 * Change bctest to avoid here-documents inside command substitution
14832 (workaround for FreeBSD /bin/sh bug).
14833 For compatibility with Ultrix, avoid shell functions (introduced
14834 in the bctest version that searches along $PATH).
14838 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14839 with des_encrypt() defined on some operating systems, like Solaris
14844 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14845 On the Importance of Eliminating Errors in Cryptographic
14846 Computations, J. Cryptology 14 (2001) 2, 101-119,
14847 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14851 * MIPS assembler BIGNUM division bug fix.
14855 * Disabled incorrect Alpha assembler code.
14859 * Fix PKCS#7 decode routines so they correctly update the length
14860 after reading an EOC for the EXPLICIT tag.
14864 *This change does not apply to 0.9.7.*
14866 * Fix bug in PKCS#12 key generation routines. This was triggered
14867 if a 3DES key was generated with a 0 initial byte. Include
14868 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14869 (but broken) behaviour.
14873 * Enhance bctest to search for a working bc along $PATH and print
14876 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14878 * Fix memory leaks in err.c: free err_data string if necessary;
14879 don't write to the wrong index in ERR_set_error_data.
14883 * Implement ssl23_peek (analogous to ssl23_read), which previously
14888 * Replace rdtsc with `_emit` statements for VC++ version 5.
14890 *Jeremy Cooper <jeremy@baymoo.org>*
14892 * Make it possible to reuse SSLv2 sessions.
14896 * In copy_email() check for >= 0 as a return value for
14897 X509_NAME_get_index_by_NID() since 0 is a valid index.
14899 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14901 * Avoid coredump with unsupported or invalid public keys by checking if
14902 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14903 PKCS7_verify() fails with non detached data.
14907 * Don't use getenv in library functions when run as setuid/setgid.
14908 New function OPENSSL_issetugid().
14912 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14913 due to incorrect handling of multi-threading:
14915 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14917 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14919 3. Count how many times MemCheck_off() has been called so that
14920 nested use can be treated correctly. This also avoids
14921 inband-signalling in the previous code (which relied on the
14922 assumption that thread ID 0 is impossible).
14926 * Add "-rand" option also to s_client and s_server.
14930 * Fix CPU detection on Irix 6.x.
14931 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14932 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14934 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14939 *This change does not apply to 0.9.7.*
14941 * Use the cached encoding of an X509_NAME structure rather than
14942 copying it. This is apparently the reason for the libsafe "errors"
14943 but the code is actually correct.
14947 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14948 Bleichenbacher's DSA attack.
14949 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14950 to be set and top=0 forces the highest bit to be set; top=-1 is new
14951 and leaves the highest bit random.
14953 *Ulf Moeller, Bodo Moeller*
14955 * In the `NCONF_...`-based implementations for `CONF_...` queries
14956 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14957 a temporary CONF structure with the data component set to NULL
14958 (which gives segmentation faults in lh_retrieve).
14959 Instead, use NULL for the CONF pointer in CONF_get_string and
14960 CONF_get_number (which may use environment variables) and directly
14961 return NULL from CONF_get_section.
14965 * Fix potential buffer overrun for EBCDIC.
14969 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14970 keyUsage if basicConstraints absent for a CA.
14974 * Make SMIME_write_PKCS7() write mail header values with a format that
14975 is more generally accepted (no spaces before the semicolon), since
14976 some programs can't parse those values properly otherwise. Also make
14977 sure BIO's that break lines after each write do not create invalid
14982 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14983 macros previously used would not encode an empty SEQUENCE OF
14984 and break the signature.
14988 *This change does not apply to 0.9.7.*
14990 * Zero the premaster secret after deriving the master secret in
14995 * Add some EVP_add_digest_alias registrations (as found in
14996 OpenSSL_add_all_digests()) to SSL_library_init()
14997 aka OpenSSL_add_ssl_algorithms(). This provides improved
14998 compatibility with peers using X.509 certificates
14999 with unconventional AlgorithmIdentifier OIDs.
15003 * Fix for Irix with NO_ASM.
15005 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15007 * ./config script fixes.
15009 *Ulf Moeller, Richard Levitte*
15011 * Fix 'openssl passwd -1'.
15015 * Change PKCS12_key_gen_asc() so it can cope with non null
15016 terminated strings whose length is passed in the passlen
15017 parameter, for example from PEM callbacks. This was done
15018 by adding an extra length parameter to asc2uni().
15020 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15022 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15023 call failed, free the DSA structure.
15027 * Fix to uni2asc() to cope with zero length Unicode strings.
15028 These are present in some PKCS#12 files.
15032 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15033 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15034 when writing a 32767 byte record.
15036 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15038 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15039 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15041 (RSA objects have a reference count access to which is protected
15042 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15043 so they are meant to be shared between threads.)
15044 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15045 "Reddie, Steven" <Steven.Reddie@ca.com>*
15047 * Fix a deadlock in CRYPTO_mem_leaks().
15051 * Use better test patterns in bntest.
15055 * rand_win.c fix for Borland C.
15059 * BN_rshift bugfix for n == 0.
15063 * Add a 'bctest' script that checks for some known 'bc' bugs
15064 so that 'make test' does not abort just because 'bc' is broken.
15068 * Store verify_result within SSL_SESSION also for client side to
15069 avoid potential security hole. (Re-used sessions on the client side
15070 always resulted in verify_result==X509_V_OK, not using the original
15071 result of the server certificate verification.)
15075 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15076 SSL3_RT_APPLICATION_DATA, return 0.
15077 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15082 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15083 releases, have been re-implemented by renaming the previous
15084 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15085 and ssl3_read_internal, respectively, and adding 'peek' parameters
15086 to them. The new ssl[23]_{read,peek} functions are calls to
15087 ssl[23]_read_internal with the 'peek' flag set appropriately.
15088 A 'peek' parameter has also been added to ssl3_read_bytes, which
15089 does the actual work for ssl3_read_internal.
15093 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15094 the method-specific "init()" handler. Also clean up ex_data after
15095 calling the method-specific "finish()" handler. Previously, this was
15096 happening the other way round.
15100 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15101 The previous value, 12, was not always sufficient for BN_mod_exp().
15105 * Make sure that shared libraries get the internal name engine with
15106 the full version number and not just 0. This should mark the
15107 shared libraries as not backward compatible. Of course, this should
15108 be changed again when we can guarantee backward binary compatibility.
15112 * Fix typo in get_cert_by_subject() in by_dir.c
15114 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15116 * Rework the system to generate shared libraries:
15118 - Make note of the expected extension for the shared libraries and
15119 if there is a need for symbolic links from for example libcrypto.so.0
15120 to libcrypto.so.0.9.7. There is extended info in Configure for
15123 - Make as few rebuilds of the shared libraries as possible.
15125 - Still avoid linking the OpenSSL programs with the shared libraries.
15127 - When installing, install the shared libraries separately from the
15132 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15134 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15135 and not in SSL_clear because the latter is also used by the
15136 accept/connect functions; previously, the settings made by
15137 SSL_set_read_ahead would be lost during the handshake.
15139 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15141 * Correct util/mkdef.pl to be selective about disabled algorithms.
15142 Previously, it would create entries for disabled algorithms no
15147 * Added several new manual pages for SSL_* function.
15151 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15153 * In ssl23_get_client_hello, generate an error message when faced
15154 with an initial SSL 3.0/TLS record that is too small to contain the
15155 first two bytes of the ClientHello message, i.e. client_version.
15156 (Note that this is a pathologic case that probably has never happened
15157 in real life.) The previous approach was to use the version number
15158 from the record header as a substitute; but our protocol choice
15159 should not depend on that one because it is not authenticated
15160 by the Finished messages.
15164 * More robust randomness gathering functions for Windows.
15166 *Jeffrey Altman <jaltman@columbia.edu>*
15168 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15169 not set then we don't setup the error code for issuer check errors
15170 to avoid possibly overwriting other errors which the callback does
15171 handle. If an application does set the flag then we assume it knows
15172 what it is doing and can handle the new informational codes
15177 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15178 a general "ANY" type, as such it should be able to decode anything
15179 including tagged types. However it didn't check the class so it would
15180 wrongly interpret tagged types in the same way as their universal
15181 counterpart and unknown types were just rejected. Changed so that the
15182 tagged and unknown types are handled in the same way as a SEQUENCE:
15183 that is the encoding is stored intact. There is also a new type
15184 "V_ASN1_OTHER" which is used when the class is not universal, in this
15185 case we have no idea what the actual type is so we just lump them all
15190 * On VMS, stdout may very well lead to a file that is written to
15191 in a record-oriented fashion. That means that every write() will
15192 write a separate record, which will be read separately by the
15193 programs trying to read from it. This can be very confusing.
15195 The solution is to put a BIO filter in the way that will buffer
15196 text until a linefeed is reached, and then write everything a
15197 line at a time, so every record written will be an actual line,
15198 not chunks of lines and not (usually doesn't happen, but I've
15199 seen it once) several lines in one record. BIO_f_linebuffer() is
15202 Currently, it's a VMS-only method, because that's where it has
15203 been tested well enough.
15207 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15208 it can return incorrect results.
15209 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15210 but it was in 0.9.6-beta[12].)
15214 * Disable the check for content being present when verifying detached
15215 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15216 include zero length content when signing messages.
15220 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15221 BIO_ctrl (for BIO pairs).
15225 * Add DSO method for VMS.
15229 * Bug fix: Montgomery multiplication could produce results with the
15234 * Add RPM specification openssl.spec and modify it to build three
15235 packages. The default package contains applications, application
15236 documentation and run-time libraries. The devel package contains
15237 include files, static libraries and function documentation. The
15238 doc package contains the contents of the doc directory. The original
15239 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15243 * Add a large number of documentation files for many SSL routines.
15245 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15247 * Add a configuration entry for Sony News 4.
15249 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15251 * Don't set the two most significant bits to one when generating a
15252 random number < q in the DSA library.
15256 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15257 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15258 the underlying transport is blocking) if a handshake took place.
15259 (The default behaviour is needed by applications such as s_client
15260 and s_server that use select() to determine when to use SSL_read;
15261 but for applications that know in advance when to expect data, it
15262 just makes things more complicated.)
15266 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15271 * Add a few more EBCDIC conditionals that make `req` and `x509`
15272 work better on such systems.
15274 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15276 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15277 Update PKCS12_parse() so it copies the friendlyName and the
15278 keyid to the certificates aux info.
15282 * Fix bug in PKCS7_verify() which caused an infinite loop
15283 if there was more than one signature.
15285 *Sven Uszpelkat <su@celocom.de>*
15287 * Major change in util/mkdef.pl to include extra information
15288 about each symbol, as well as presenting variables as well
15289 as functions. This change means that there's n more need
15290 to rebuild the .num files when some algorithms are excluded.
15294 * Allow the verify time to be set by an application,
15295 rather than always using the current time.
15299 * Phase 2 verify code reorganisation. The certificate
15300 verify code now looks up an issuer certificate by a
15301 number of criteria: subject name, authority key id
15302 and key usage. It also verifies self signed certificates
15303 by the same criteria. The main comparison function is
15304 X509_check_issued() which performs these checks.
15306 Lot of changes were necessary in order to support this
15307 without completely rewriting the lookup code.
15309 Authority and subject key identifier are now cached.
15311 The LHASH 'certs' is X509_STORE has now been replaced
15312 by a STACK_OF(X509_OBJECT). This is mainly because an
15313 LHASH can't store or retrieve multiple objects with
15314 the same hash value.
15316 As a result various functions (which were all internal
15317 use only) have changed to handle the new X509_STORE
15318 structure. This will break anything that messed round
15319 with X509_STORE internally.
15321 The functions X509_STORE_add_cert() now checks for an
15322 exact match, rather than just subject name.
15324 The X509_STORE API doesn't directly support the retrieval
15325 of multiple certificates matching a given criteria, however
15326 this can be worked round by performing a lookup first
15327 (which will fill the cache with candidate certificates)
15328 and then examining the cache for matches. This is probably
15329 the best we can do without throwing out X509_LOOKUP
15330 entirely (maybe later...).
15332 The X509_VERIFY_CTX structure has been enhanced considerably.
15334 All certificate lookup operations now go via a get_issuer()
15335 callback. Although this currently uses an X509_STORE it
15336 can be replaced by custom lookups. This is a simple way
15337 to bypass the X509_STORE hackery necessary to make this
15338 work and makes it possible to use more efficient techniques
15339 in future. A very simple version which uses a simple
15340 STACK for its trusted certificate store is also provided
15341 using X509_STORE_CTX_trusted_stack().
15343 The verify_cb() and verify() callbacks now have equivalents
15344 in the X509_STORE_CTX structure.
15346 X509_STORE_CTX also has a 'flags' field which can be used
15347 to customise the verify behaviour.
15351 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15352 excludes S/MIME capabilities.
15356 * When a certificate request is read in keep a copy of the
15357 original encoding of the signed data and use it when outputting
15358 again. Signatures then use the original encoding rather than
15359 a decoded, encoded version which may cause problems if the
15360 request is improperly encoded.
15364 * For consistency with other BIO_puts implementations, call
15365 buffer_write(b, ...) directly in buffer_puts instead of calling
15368 In BIO_puts, increment b->num_write as in BIO_write.
15370 *Peter.Sylvester@EdelWeb.fr*
15372 * Fix BN_mul_word for the case where the word is 0. (We have to use
15373 BN_zero, we may not return a BIGNUM with an array consisting of
15374 words set to zero.)
15378 * Avoid calling abort() from within the library when problems are
15379 detected, except if preprocessor symbols have been defined
15380 (such as REF_CHECK, BN_DEBUG etc.).
15384 * New openssl application 'rsautl'. This utility can be
15385 used for low-level RSA operations. DER public key
15386 BIO/fp routines also added.
15390 * New Configure entry and patches for compiling on QNX 4.
15392 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15394 * A demo state-machine implementation was sponsored by
15395 Nuron (<http://www.nuron.com/>) and is now available in
15396 demos/state_machine.
15400 * New options added to the 'dgst' utility for signature
15401 generation and verification.
15405 * Unrecognized PKCS#7 content types are now handled via a
15406 catch all ASN1_TYPE structure. This allows unsupported
15407 types to be stored as a "blob" and an application can
15408 encode and decode it manually.
15412 * Fix various signed/unsigned issues to make a_strex.c
15413 compile under VC++.
15415 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15417 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15418 length if passed a buffer. ASN1_INTEGER_to_BN failed
15419 if passed a NULL BN and its argument was negative.
15421 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15423 * Modification to PKCS#7 encoding routines to output definite
15424 length encoding. Since currently the whole structures are in
15425 memory there's not real point in using indefinite length
15426 constructed encoding. However if OpenSSL is compiled with
15427 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15431 * Added BIO_vprintf() and BIO_vsnprintf().
15435 * Added more prefixes to parse for in the strings written
15436 through a logging bio, to cover all the levels that are available
15437 through syslog. The prefixes are now:
15439 PANIC, EMERG, EMR => LOG_EMERG
15440 ALERT, ALR => LOG_ALERT
15441 CRIT, CRI => LOG_CRIT
15442 ERROR, ERR => LOG_ERR
15443 WARNING, WARN, WAR => LOG_WARNING
15444 NOTICE, NOTE, NOT => LOG_NOTICE
15445 INFO, INF => LOG_INFO
15446 DEBUG, DBG => LOG_DEBUG
15448 and as before, if none of those prefixes are present at the
15449 beginning of the string, LOG_ERR is chosen.
15451 On Win32, the `LOG_*` levels are mapped according to this:
15453 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15454 LOG_WARNING => EVENTLOG_WARNING_TYPE
15455 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15459 * Made it possible to reconfigure with just the configuration
15460 argument "reconf" or "reconfigure". The command line arguments
15461 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15462 and are retrieved from there when reconfiguring.
15468 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15470 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15474 * The obj_dat.pl script was messing up the sorting of object
15475 names. The reason was that it compared the quoted version
15476 of strings as a result "OCSP" > "OCSP Signing" because
15477 " > SPACE. Changed script to store unquoted versions of
15478 names and add quotes on output. It was also omitting some
15479 names from the lookup table if they were given a default
15480 value (that is if SN is missing it is given the same
15481 value as LN and vice versa), these are now added on the
15482 grounds that if an object has a name we should be able to
15483 look it up. Finally added warning output when duplicate
15484 short or long names are found.
15488 * Changes needed for Tandem NSK.
15490 *Scott Uroff <scott@xypro.com>*
15492 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15493 RSA_padding_check_SSLv23(), special padding was never detected
15494 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15495 version rollback attacks was not effective.
15497 In s23_clnt.c, don't use special rollback-attack detection padding
15498 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15499 client; similarly, in s23_srvr.c, don't do the rollback check if
15500 SSL 2.0 is the only protocol enabled in the server.
15504 * Make it possible to get hexdumps of unprintable data with 'openssl
15505 asn1parse'. By implication, the functions ASN1_parse_dump() and
15506 BIO_dump_indent() are added.
15510 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15511 these print out strings and name structures based on various
15512 flags including RFC2253 support and proper handling of
15513 multibyte characters. Added options to the 'x509' utility
15514 to allow the various flags to be set.
15518 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15519 Also change the functions X509_cmp_current_time() and
15520 X509_gmtime_adj() work with an ASN1_TIME structure,
15521 this will enable certificates using GeneralizedTime in validity
15522 dates to be checked.
15526 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15527 negative public key encodings) on by default,
15528 NO_NEG_PUBKEY_BUG can be set to disable it.
15532 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15533 content octets. An i2c_ASN1_OBJECT is unnecessary because
15534 the encoding can be trivially obtained from the structure.
15538 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15539 not read locks (`CRYPTO_r_[un]lock`).
15543 * A first attempt at creating official support for shared
15544 libraries through configuration. I've kept it so the
15545 default is static libraries only, and the OpenSSL programs
15546 are always statically linked for now, but there are
15547 preparations for dynamic linking in place.
15548 This has been tested on Linux and Tru64.
15552 * Randomness polling function for Win9x, as described in:
15553 Peter Gutmann, Software Generation of Practically Strong
15558 * Fix so PRNG is seeded in req if using an already existing
15563 * New options to smime application. -inform and -outform
15564 allow alternative formats for the S/MIME message including
15565 PEM and DER. The -content option allows the content to be
15566 specified separately. This should allow things like Netscape
15567 form signing output easier to verify.
15571 * Fix the ASN1 encoding of tags using the 'long form'.
15575 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15576 STRING types. These convert content octets to and from the
15577 underlying type. The actual tag and length octets are
15578 already assumed to have been read in and checked. These
15579 are needed because all other string types have virtually
15580 identical handling apart from the tag. By having versions
15581 of the ASN1 functions that just operate on content octets
15582 IMPLICIT tagging can be handled properly. It also allows
15583 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15584 and ASN1_INTEGER are identical apart from the tag.
15588 * Change the handling of OID objects as follows:
15590 - New object identifiers are inserted in objects.txt, following
15591 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15592 - objects.pl is used to process obj_mac.num and create a new
15594 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15597 This is currently kind of a hack, and the perl code in objects.pl
15598 isn't very elegant, but it works as I intended. The simplest way
15599 to check that it worked correctly is to look in obj_dat.h and
15600 check the array nid_objs and make sure the objects haven't moved
15601 around (this is important!). Additions are OK, as well as
15602 consistent name changes.
15606 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15610 * Addition of the command line parameter '-rand file' to 'openssl req'.
15611 The given file adds to whatever has already been seeded into the
15612 random pool through the RANDFILE configuration file option or
15613 environment variable, or the default random state file.
15617 * mkstack.pl now sorts each macro group into lexical order.
15618 Previously the output order depended on the order the files
15619 appeared in the directory, resulting in needless rewriting
15624 * Patches to make OpenSSL compile under Win32 again. Mostly
15625 work arounds for the VC++ problem that it treats func() as
15626 func(void). Also stripped out the parts of mkdef.pl that
15627 added extra typesafe functions: these no longer exist.
15631 * Reorganisation of the stack code. The macros are now all
15632 collected in safestack.h . Each macro is defined in terms of
15633 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15634 DEBUG_SAFESTACK is now handled in terms of function casts,
15635 this has the advantage of retaining type safety without the
15636 use of additional functions. If DEBUG_SAFESTACK is not defined
15637 then the non typesafe macros are used instead. Also modified the
15638 mkstack.pl script to handle the new form. Needs testing to see
15639 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15640 the default if no major problems. Similar behaviour for ASN1_SET_OF
15641 and PKCS12_STACK_OF.
15645 * When some versions of IIS use the 'NET' form of private key the
15646 key derivation algorithm is different. Normally MD5(password) is
15647 used as a 128 bit RC4 key. In the modified case
15648 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15649 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15650 as the old Netscape_RSA functions except they have an additional
15651 'sgckey' parameter which uses the modified algorithm. Also added
15652 an -sgckey command line option to the rsa utility. Thanks to
15653 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15654 algorithm to openssl-dev.
15658 * The evp_local.h macros were using 'c.##kname' which resulted in
15659 invalid expansion on some systems (SCO 5.0.5 for example).
15660 Corrected to 'c.kname'.
15662 *Phillip Porch <root@theporch.com>*
15664 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15665 a STACK of email addresses from a certificate or request, these look
15666 in the subject name and the subject alternative name extensions and
15667 omit any duplicate addresses.
15671 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15672 This makes DSA verification about 2 % faster.
15676 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15677 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15678 plus overhead for 1024 bit moduli).
15679 This makes exponentiations about 0.5 % faster for 1024 bit
15680 exponents (as measured by "openssl speed rsa2048").
15684 * Rename memory handling macros to avoid conflicts with other
15686 Malloc => OPENSSL_malloc
15687 Malloc_locked => OPENSSL_malloc_locked
15688 Realloc => OPENSSL_realloc
15689 Free => OPENSSL_free
15693 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15694 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15698 * CygWin32 support.
15700 *John Jarvie <jjarvie@newsguy.com>*
15702 * The type-safe stack code has been rejigged. It is now only compiled
15703 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15704 by default all type-specific stack functions are "#define"d back to
15705 standard stack functions. This results in more streamlined output
15706 but retains the type-safety checking possibilities of the original
15711 * The STACK code has been cleaned up, and certain type declarations
15712 that didn't make a lot of sense have been brought in line. This has
15713 also involved a cleanup of sorts in safestack.h to more correctly
15714 map type-safe stack functions onto their plain stack counterparts.
15715 This work has also resulted in a variety of "const"ifications of
15716 lots of the code, especially `_cmp` operations which should normally
15717 be prototyped with "const" parameters anyway.
15721 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15722 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15723 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15724 where all of 'md' is used each time the PRNG is used, but 'state'
15725 is used only indexed by a cyclic counter. As entropy may not be
15726 well distributed from the beginning, 'md' is important as a
15727 chaining variable. However, the output function chains only half
15728 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15729 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15730 in all of 'state' being rewritten, with the new values depending
15731 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15735 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15736 the handshake is continued after ssl_verify_cert_chain();
15737 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15738 can lead to 'unexplainable' connection aborts later.
15740 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15742 * Major EVP API cipher revision.
15743 Add hooks for extra EVP features. This allows various cipher
15744 parameters to be set in the EVP interface. Support added for variable
15745 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15746 setting of RC2 and RC5 parameters.
15748 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15751 Remove lots of duplicated code from the EVP library. For example *every*
15752 cipher init() function handles the 'iv' in the same way according to the
15753 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15754 for CFB and OFB modes they zero ctx->num.
15756 New functionality allows removal of S/MIME code RC2 hack.
15758 Most of the routines have the same form and so can be declared in terms
15761 By shifting this to the top level EVP_CipherInit() it can be removed from
15762 all individual ciphers. If the cipher wants to handle IVs or keys
15763 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15766 Change lots of functions like EVP_EncryptUpdate() to now return a
15767 value: although software versions of the algorithms cannot fail
15768 any installed hardware versions can.
15772 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15773 this option is set, tolerate broken clients that send the negotiated
15774 protocol version number instead of the requested protocol version
15779 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15780 i.e. non-zero for export ciphersuites, zero otherwise.
15781 Previous versions had this flag inverted, inconsistent with
15782 rsa_tmp_cb (..._TMP_RSA_CB).
15784 *Bodo Moeller; problem reported by Amit Chopra*
15786 * Add missing DSA library text string. Work around for some IIS
15787 key files with invalid SEQUENCE encoding.
15791 * Add a document (doc/standards.txt) that list all kinds of standards
15792 and so on that are implemented in OpenSSL.
15796 * Enhance c_rehash script. Old version would mishandle certificates
15797 with the same subject name hash and wouldn't handle CRLs at all.
15798 Added -fingerprint option to crl utility, to support new c_rehash
15803 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15807 * Fix for SSL server purpose checking. Server checking was
15808 rejecting certificates which had extended key usage present
15809 but no ssl client purpose.
15811 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15813 * Make PKCS#12 code work with no password. The PKCS#12 spec
15814 is a little unclear about how a blank password is handled.
15815 Since the password in encoded as a BMPString with terminating
15816 double NULL a zero length password would end up as just the
15817 double NULL. However no password at all is different and is
15818 handled differently in the PKCS#12 key generation code. NS
15819 treats a blank password as zero length. MSIE treats it as no
15820 password on export: but it will try both on import. We now do
15821 the same: PKCS12_parse() tries zero length and no password if
15822 the password is set to "" or NULL (NULL is now a valid password:
15823 it wasn't before) as does the pkcs12 application.
15827 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15828 perror when PEM_read_bio_X509_REQ fails, the error message must
15829 be obtained from the error queue.
15833 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15834 it in ERR_remove_state if appropriate, and change ERR_get_state
15835 accordingly to avoid race conditions (this is necessary because
15836 thread_hash is no longer constant once set).
15840 * Bugfix for linux-elf makefile.one.
15844 * RSA_get_default_method() will now cause a default
15845 RSA_METHOD to be chosen if one doesn't exist already.
15846 Previously this was only set during a call to RSA_new()
15847 or RSA_new_method(NULL) meaning it was possible for
15848 RSA_get_default_method() to return NULL.
15852 * Added native name translation to the existing DSO code
15853 that will convert (if the flag to do so is set) filenames
15854 that are sufficiently small and have no path information
15855 into a canonical native form. Eg. "blah" converted to
15856 "libblah.so" or "blah.dll" etc.
15860 * New function ERR_error_string_n(e, buf, len) which is like
15861 ERR_error_string(e, buf), but writes at most 'len' bytes
15862 including the 0 terminator. For ERR_error_string_n, 'buf'
15865 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15867 * CONF library reworked to become more general. A new CONF
15868 configuration file reader "class" is implemented as well as a
15869 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15870 old `CONF_*` functions are still there, but are reimplemented to
15871 work in terms of the new functions. Also, a set of functions
15872 to handle the internal storage of the configuration data is
15873 provided to make it easier to write new configuration file
15874 reader "classes" (I can definitely see something reading a
15875 configuration file in XML format, for example), called `_CONF_*`,
15876 or "the configuration storage API"...
15878 The new configuration file reading functions are:
15880 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15881 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15883 NCONF_default, NCONF_WIN32
15885 NCONF_dump_fp, NCONF_dump_bio
15887 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15888 NCONF_new creates a new CONF object. This works in the same way
15889 as other interfaces in OpenSSL, like the BIO interface.
15890 `NCONF_dump_*` dump the internal storage of the configuration file,
15891 which is useful for debugging. All other functions take the same
15892 arguments as the old `CONF_*` functions with the exception of the
15893 first that must be a `CONF *` instead of a `LHASH *`.
15895 To make it easier to use the new classes with the old `CONF_*` functions,
15896 the function CONF_set_default_method is provided.
15900 * Add '-tls1' option to 'openssl ciphers', which was already
15901 mentioned in the documentation but had not been implemented.
15902 (This option is not yet really useful because even the additional
15903 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15907 * Initial DSO code added into libcrypto for letting OpenSSL (and
15908 OpenSSL-based applications) load shared libraries and bind to
15909 them in a portable way.
15911 *Geoff Thorpe, with contributions from Richard Levitte*
15913 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15915 * Make sure _lrotl and _lrotr are only used with MSVC.
15917 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15918 (the default implementation of RAND_status).
15920 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15921 to '-clrext' (= clear extensions), as intended and documented.
15922 *Bodo Moeller; inconsistency pointed out by Michael Attili
15923 <attili@amaxo.com>*
15925 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15926 was larger than the MD block size.
15928 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15930 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15931 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15932 using the passed key: if the passed key was a private key the result
15933 of X509_print(), for example, would be to print out all the private key
15938 * des_quad_cksum() byte order bug fix.
15939 *Ulf Möller, using the problem description in krb4-0.9.7, where
15940 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15942 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15945 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15947 * For easily testing in shell scripts whether some command
15948 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15949 returns with exit code 0 iff no command of the given name is available.
15950 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15951 the output goes to stdout and nothing is printed to stderr.
15952 Additional arguments are always ignored.
15954 Since for each cipher there is a command of the same name,
15955 the 'no-cipher' compilation switches can be tested this way.
15957 ('openssl no-XXX' is not able to detect pseudo-commands such
15958 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15962 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15966 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15967 is set; it will be thrown away anyway because each handshake creates
15969 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15970 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15971 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15972 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15976 * New s_client option -ign_eof: EOF at stdin is ignored, and
15977 'Q' and 'R' lose their special meanings (quit/renegotiate).
15978 This is part of what -quiet does; unlike -quiet, -ign_eof
15979 does not suppress any output.
15983 * Add compatibility options to the purpose and trust code. The
15984 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15985 accepts a certificate or CA, this was the previous behaviour,
15986 with all the associated security issues.
15988 X509_TRUST_COMPAT is the old trust behaviour: only and
15989 automatically trust self signed roots in certificate store. A
15990 new trust setting X509_TRUST_DEFAULT is used to specify that
15991 a purpose has no associated trust setting and it should instead
15992 use the value in the default purpose.
15996 * Fix the PKCS#8 DSA private key code so it decodes keys again
15997 and fix a memory leak.
16001 * In util/mkerr.pl (which implements 'make errors'), preserve
16002 reason strings from the previous version of the .c file, as
16003 the default to have only downcase letters (and digits) in
16004 automatically generated reasons codes is not always appropriate.
16008 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16009 using strerror. Previously, ERR_reason_error_string() returned
16010 library names as reason strings for SYSerr; but SYSerr is a special
16011 case where small numbers are errno values, not library numbers.
16015 * Add '-dsaparam' option to 'openssl dhparam' application. This
16016 converts DSA parameters into DH parameters. (When creating parameters,
16017 DSA_generate_parameters is used.)
16021 * Include 'length' (recommended exponent length) in C code generated
16022 by 'openssl dhparam -C'.
16026 * The second argument to set_label in perlasm was already being used
16027 so couldn't be used as a "file scope" flag. Moved to third argument
16032 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16033 instead of RAND_bytes for encryption IVs and salts.
16037 * Include RAND_status() into RAND_METHOD instead of implementing
16038 it only for md_rand.c Otherwise replacing the PRNG by calling
16039 RAND_set_rand_method would be impossible.
16043 * Don't let DSA_generate_key() enter an infinite loop if the random
16044 number generation fails.
16048 * New 'rand' application for creating pseudo-random output.
16052 * Added configuration support for Linux/IA64
16054 *Rolf Haberrecker <rolf@suse.de>*
16056 * Assembler module support for Mingw32.
16060 * Shared library support for HPUX (in shlib/).
16062 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16064 * Shared library support for Solaris gcc.
16066 *Lutz Behnke <behnke@trustcenter.de>*
16068 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16070 * PKCS7_encrypt() was adding text MIME headers twice because they
16071 were added manually and by SMIME_crlf_copy().
16075 * In bntest.c don't call BN_rand with zero bits argument.
16077 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16079 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16080 case was implemented. This caused BN_div_recp() to fail occasionally.
16084 * Add an optional second argument to the set_label() in the perl
16085 assembly language builder. If this argument exists and is set
16086 to 1 it signals that the assembler should use a symbol whose
16087 scope is the entire file, not just the current function. This
16088 is needed with MASM which uses the format label:: for this scope.
16090 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16092 * Change the ASN1 types so they are typedefs by default. Before
16093 almost all types were #define'd to ASN1_STRING which was causing
16094 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16099 * Change names of new functions to the new get1/get0 naming
16100 convention: After 'get1', the caller owns a reference count
16101 and has to call `..._free`; 'get0' returns a pointer to some
16102 data structure without incrementing reference counters.
16103 (Some of the existing 'get' functions increment a reference
16104 counter, some don't.)
16105 Similarly, 'set1' and 'add1' functions increase reference
16106 counters or duplicate objects.
16110 * Allow for the possibility of temp RSA key generation failure:
16111 the code used to assume it always worked and crashed on failure.
16115 * Fix potential buffer overrun problem in BIO_printf().
16116 *Ulf Möller, using public domain code by Patrick Powell; problem
16117 pointed out by David Sacerdote <das33@cornell.edu>*
16119 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16120 RAND_egd() and RAND_status(). In the command line application,
16121 the EGD socket can be specified like a seed file using RANDFILE
16126 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16127 Some CAs (e.g. Verisign) distribute certificates in this form.
16131 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16132 list to exclude them. This means that no special compilation option
16133 is needed to use anonymous DH: it just needs to be included in the
16138 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16139 EVP_MD_type. The old functionality is available in a new macro called
16140 EVP_MD_md(). Change code that uses it and update docs.
16144 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16145 where the `void *` argument is replaced by a function pointer argument.
16146 Previously `void *` was abused to point to functions, which works on
16147 many platforms, but is not correct. As these functions are usually
16148 called by macros defined in OpenSSL header files, most source code
16149 should work without changes.
16153 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16154 sections with information on -D... compiler switches used for
16155 compiling the library so that applications can see them. To enable
16156 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16157 must be defined. E.g.,
16158 #define OPENSSL_ALGORITHM_DEFINES
16159 #include <openssl/opensslconf.h>
16160 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16162 *Richard Levitte, Ulf and Bodo Möller*
16164 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16169 * Change the 'other' type in certificate aux info to a STACK_OF
16170 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16171 the required ASN1 format: arbitrary types determined by an OID.
16175 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16176 argument to 'req'. This is not because the function is newer or
16177 better than others it just uses the work 'NEW' in the certificate
16178 request header lines. Some software needs this.
16182 * Reorganise password command line arguments: now passwords can be
16183 obtained from various sources. Delete the PEM_cb function and make
16184 it the default behaviour: i.e. if the callback is NULL and the
16185 usrdata argument is not NULL interpret it as a null terminated pass
16186 phrase. If usrdata and the callback are NULL then the pass phrase
16187 is prompted for as usual.
16191 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16192 the support is automatically enabled. The resulting binaries will
16193 autodetect the card and use it if present.
16195 *Ben Laurie and Compaq Inc.*
16197 * Work around for Netscape hang bug. This sends certificate request
16198 and server done in one record. Since this is perfectly legal in the
16199 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16200 the bugs/SSLv3 entry for more info.
16204 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16208 * Add -rand argument to smime and pkcs12 applications and read/write
16213 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16217 * Add command line password options to the remaining applications.
16221 * Bug fix for BN_div_recp() for numerators with an even number of
16226 * More tests in bntest.c, and changed test_bn output.
16230 * ./config recognizes MacOS X now.
16234 * Bug fix for BN_div() when the first words of num and divisor are
16235 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16239 * Add support for various broken PKCS#8 formats, and command line
16240 options to produce them.
16244 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16245 get temporary BIGNUMs from a BN_CTX.
16249 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16254 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16255 include a #define from the old name to the new. The original intent
16256 was that statically linked binaries could for example just call
16257 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16258 link with digests. This never worked because SSLeay_add_all_digests()
16259 and SSLeay_add_all_ciphers() were in the same source file so calling
16260 one would link with the other. They are now in separate source files.
16264 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16268 * Use a less unusual form of the Miller-Rabin primality test (it used
16269 a binary algorithm for exponentiation integrated into the Miller-Rabin
16270 loop, our standard modexp algorithms are faster).
16274 * Support for the EBCDIC character set completed.
16276 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16278 * Source code cleanups: use const where appropriate, eliminate casts,
16279 use `void *` instead of `char *` in lhash.
16283 * Bugfix: ssl3_send_server_key_exchange was not restartable
16284 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16285 this the server could overwrite ephemeral keys that the client
16290 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16291 using 50 iterations of the Rabin-Miller test.
16293 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16294 iterations of the Rabin-Miller test as required by the appendix
16295 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16296 As BN_is_prime_fasttest includes trial division, DSA parameter
16297 generation becomes much faster.
16299 This implies a change for the callback functions in DSA_is_prime
16300 and DSA_generate_parameters: The callback function is called once
16301 for each positive witness in the Rabin-Miller test, not just
16302 occasionally in the inner loop; and the parameters to the
16303 callback function now provide an iteration count for the outer
16304 loop rather than for the current invocation of the inner loop.
16305 DSA_generate_parameters additionally can call the callback
16306 function with an 'iteration count' of -1, meaning that a
16307 candidate has passed the trial division test (when q is generated
16308 from an application-provided seed, trial division is skipped).
16312 * New function BN_is_prime_fasttest that optionally does trial
16313 division before starting the Rabin-Miller test and has
16314 an additional BN_CTX * argument (whereas BN_is_prime always
16315 has to allocate at least one BN_CTX).
16316 'callback(1, -1, cb_arg)' is called when a number has passed the
16317 trial division stage.
16321 * Fix for bug in CRL encoding. The validity dates weren't being handled
16326 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16330 * New function BN_pseudo_rand().
16334 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16335 bignum version of BN_from_montgomery() with the working code from
16336 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16341 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16342 made it impossible to use the same SSL_SESSION data structure in
16343 SSL2 clients in multiple threads.
16347 * The return value of RAND_load_file() no longer counts bytes obtained
16348 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16349 to seed the PRNG (previously an explicit byte count was required).
16351 *Ulf Möller, Bodo Möller*
16353 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16354 used `char *` instead of `void *` and had casts all over the place.
16358 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16362 * Retain source code compatibility for BN_prime_checks macro:
16363 BN_is_prime(..., BN_prime_checks, ...) now uses
16364 BN_prime_checks_for_size to determine the appropriate number of
16365 Rabin-Miller iterations.
16369 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16370 DH_CHECK_P_NOT_SAFE_PRIME.
16371 (Check if this is true? OpenPGP calls them "strong".)
16375 * Merge the functionality of "dh" and "gendh" programs into a new program
16376 "dhparam". The old programs are retained for now but will handle DH keys
16377 (instead of parameters) in future.
16381 * Make the ciphers, s_server and s_client programs check the return values
16382 when a new cipher list is set.
16386 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16387 ciphers. Before when the 56bit ciphers were enabled the sorting was
16390 The syntax for the cipher sorting has been extended to support sorting by
16391 cipher-strength (using the strength_bits hard coded in the tables).
16392 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16394 Fix a bug in the cipher-command parser: when supplying a cipher command
16395 string with an "undefined" symbol (neither command nor alphanumeric
16396 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16397 an error is flagged.
16399 Due to the strength-sorting extension, the code of the
16400 ssl_create_cipher_list() function was completely rearranged. I hope that
16401 the readability was also increased :-)
16403 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16405 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16406 for the first serial number and places 2 in the serial number file. This
16407 avoids problems when the root CA is created with serial number zero and
16408 the first user certificate has the same issuer name and serial number
16413 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16414 the new code. Add documentation for this stuff.
16418 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16419 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16420 structures and behave in an analogous way to the X509v3 functions:
16421 they shouldn't be called directly but wrapper functions should be used
16424 So we also now have some wrapper functions that call the X509at functions
16425 when passed certificate requests. (TO DO: similar things can be done with
16426 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16427 things. Some of these need some d2i or i2d and print functionality
16428 because they handle more complex structures.)
16432 * Add missing #ifndefs that caused missing symbols when building libssl
16433 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16434 NO_RSA in `ssl/s2*.c`.
16436 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16438 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16439 has a return value which indicates the quality of the random data
16440 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16441 error queue. New function RAND_pseudo_bytes() generates output that is
16442 guaranteed to be unique but not unpredictable. RAND_add is like
16443 RAND_seed, but takes an extra argument for an entropy estimate
16444 (RAND_seed always assumes full entropy).
16448 * Do more iterations of Rabin-Miller probable prime test (specifically,
16449 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16450 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16451 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16452 false-positive rate of at most 2^-80 for random input.
16456 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16460 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16461 in the 0.9.5 release), this returns the chain
16462 from an X509_CTX structure with a dup of the stack and all
16463 the X509 reference counts upped: so the stack will exist
16464 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16467 Also make SSL_SESSION_print() print out the verify return
16472 * Add manpage for the pkcs12 command. Also change the default
16473 behaviour so MAC iteration counts are used unless the new
16474 -nomaciter option is used. This improves file security and
16475 only older versions of MSIE (4.0 for example) need it.
16479 * Honor the no-xxx Configure options when creating .DEF files.
16483 * Add PKCS#10 attributes to field table: challengePassword,
16484 unstructuredName and unstructuredAddress. These are taken from
16485 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16486 international characters are used.
16488 More changes to X509_ATTRIBUTE code: allow the setting of types
16489 based on strings. Remove the 'loc' parameter when adding
16490 attributes because these will be a SET OF encoding which is sorted
16495 * Initial changes to the 'req' utility to allow request generation
16496 automation. This will allow an application to just generate a template
16497 file containing all the field values and have req construct the
16500 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16501 used all over the place including certificate requests and PKCS#7
16502 structures. They are currently handled manually where necessary with
16503 some primitive wrappers for PKCS#7. The new functions behave in a
16504 manner analogous to the X509 extension functions: they allow
16505 attributes to be looked up by NID and added.
16507 Later something similar to the X509V3 code would be desirable to
16508 automatically handle the encoding, decoding and printing of the
16509 more complex types. The string types like challengePassword can
16510 be handled by the string table functions.
16512 Also modified the multi byte string table handling. Now there is
16513 a 'global mask' which masks out certain types. The table itself
16514 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16515 is useful when for example there is only one permissible type
16516 (as in countryName) and using the mask might result in no valid
16521 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16522 SSL_get_peer_finished to allow applications to obtain the latest
16523 Finished messages sent to the peer or expected from the peer,
16524 respectively. (SSL_get_peer_finished is usually the Finished message
16525 actually received from the peer, otherwise the protocol will be aborted.)
16527 As the Finished message are message digests of the complete handshake
16528 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16529 be used for external authentication procedures when the authentication
16530 provided by SSL/TLS is not desired or is not enough.
16534 * Enhanced support for Alpha Linux is added. Now ./config checks if
16535 the host supports BWX extension and if Compaq C is present on the
16536 $PATH. Just exploiting of the BWX extension results in 20-30%
16537 performance kick for some algorithms, e.g. DES and RC4 to mention
16538 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16543 * Add support for MS "fast SGC". This is arguably a violation of the
16544 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16545 weak crypto and after checking the certificate is SGC a second one
16546 with strong crypto. MS SGC stops the first handshake after receiving
16547 the server certificate message and sends a second client hello. Since
16548 a server will typically do all the time consuming operations before
16549 expecting any further messages from the client (server key exchange
16550 is the most expensive) there is little difference between the two.
16552 To get OpenSSL to support MS SGC we have to permit a second client
16553 hello message after we have sent server done. In addition we have to
16554 reset the MAC if we do get this second client hello.
16558 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16559 if a DER encoded private key is RSA or DSA traditional format. Changed
16560 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16561 format DER encoded private key. Newer code should use PKCS#8 format which
16562 has the key type encoded in the ASN1 structure. Added DER private key
16563 support to pkcs8 application.
16567 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16568 ciphersuites has been selected (as required by the SSL 3/TLS 1
16569 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16570 is set, we interpret this as a request to violate the specification
16571 (the worst that can happen is a handshake failure, and 'correct'
16572 behaviour would result in a handshake failure anyway).
16576 * In SSL_CTX_add_session, take into account that there might be multiple
16577 SSL_SESSION structures with the same session ID (e.g. when two threads
16578 concurrently obtain them from an external cache).
16579 The internal cache can handle only one SSL_SESSION with a given ID,
16580 so if there's a conflict, we now throw out the old one to achieve
16585 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16586 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16587 some routines that use cipher OIDs: some ciphers do not have OIDs
16588 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16593 * Simplify the trust setting structure and code. Now we just have
16594 two sequences of OIDs for trusted and rejected settings. These will
16595 typically have values the same as the extended key usage extension
16596 and any application specific purposes.
16598 The trust checking code now has a default behaviour: it will just
16599 check for an object with the same NID as the passed id. Functions can
16600 be provided to override either the default behaviour or the behaviour
16601 for a given id. SSL client, server and email already have functions
16602 in place for compatibility: they check the NID and also return "trusted"
16603 if the certificate is self signed.
16607 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16608 traditional format into an EVP_PKEY structure.
16612 * Add a password callback function PEM_cb() which either prompts for
16613 a password if usr_data is NULL or otherwise assumes it is a null
16614 terminated password. Allow passwords to be passed on command line
16615 environment or config files in a few more utilities.
16619 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16620 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16621 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16622 Update documentation.
16626 * Support for ASN1 "NULL" type. This could be handled before by using
16627 ASN1_TYPE but there wasn't any function that would try to read a NULL
16628 and produce an error if it couldn't. For compatibility we also have
16629 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16630 don't allocate anything because they don't need to.
16634 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16637 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16639 * Rebuild of the memory allocation routines used by OpenSSL code and
16640 possibly others as well. The purpose is to make an interface that
16641 provide hooks so anyone can build a separate set of allocation and
16642 deallocation routines to be used by OpenSSL, for example memory
16643 pool implementations, or something else, which was previously hard
16644 since Malloc(), Realloc() and Free() were defined as macros having
16645 the values malloc, realloc and free, respectively (except for Win32
16646 compilations). The same is provided for memory debugging code.
16647 OpenSSL already comes with functionality to find memory leaks, but
16648 this gives people a chance to debug other memory problems.
16650 With these changes, a new set of functions and macros have appeared:
16652 CRYPTO_set_mem_debug_functions() [F]
16653 CRYPTO_get_mem_debug_functions() [F]
16654 CRYPTO_dbg_set_options() [F]
16655 CRYPTO_dbg_get_options() [F]
16656 CRYPTO_malloc_debug_init() [M]
16658 The memory debug functions are NULL by default, unless the library
16659 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16660 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16661 gives the standard debugging functions that come with OpenSSL) or
16662 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16663 provided by the library user) must be used. When the standard
16664 debugging functions are used, CRYPTO_dbg_set_options can be used to
16665 request additional information:
16666 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16667 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16669 Also, things like CRYPTO_set_mem_functions will always give the
16670 expected result (the new set of functions is used for allocation
16671 and deallocation) at all times, regardless of platform and compiler
16674 To finish it up, some functions that were never use in any other
16675 way than through macros have a new API and new semantic:
16677 CRYPTO_dbg_malloc()
16678 CRYPTO_dbg_realloc()
16681 All macros of value have retained their old syntax.
16683 *Richard Levitte and Bodo Moeller*
16685 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16686 ordering of SMIMECapabilities wasn't in "strength order" and there
16687 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16692 * Some ASN1 types with illegal zero length encoding (INTEGER,
16693 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16695 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16697 * Merge in my S/MIME library for OpenSSL. This provides a simple
16698 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16699 functionality to handle multipart/signed properly) and a utility
16700 called 'smime' to call all this stuff. This is based on code I
16701 originally wrote for Celo who have kindly allowed it to be
16702 included in OpenSSL.
16706 * Add variants des_set_key_checked and des_set_key_unchecked of
16707 des_set_key (aka des_key_sched). Global variable des_check_key
16708 decides which of these is called by des_set_key; this way
16709 des_check_key behaves as it always did, but applications and
16710 the library itself, which was buggy for des_check_key == 1,
16711 have a cleaner way to pick the version they need.
16715 * New function PKCS12_newpass() which changes the password of a
16720 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16721 dynamic mix. In both cases the ids can be used as an index into the
16722 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16723 functions so they accept a list of the field values and the
16724 application doesn't need to directly manipulate the X509_TRUST
16729 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16734 * Modify the way the V3 extension code looks up extensions. This now
16735 works in a similar way to the object code: we have some "standard"
16736 extensions in a static table which is searched with OBJ_bsearch()
16737 and the application can add dynamic ones if needed. The file
16738 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16739 updated whenever a new extension is added to the core code and kept
16740 in ext_nid order. There is a simple program 'tabtest.c' which checks
16741 this. New extensions are not added too often so this file can readily
16742 be maintained manually.
16744 There are two big advantages in doing things this way. The extensions
16745 can be looked up immediately and no longer need to be "added" using
16746 X509V3_add_standard_extensions(): this function now does nothing.
16747 Side note: I get *lots* of email saying the extension code doesn't
16748 work because people forget to call this function.
16749 Also no dynamic allocation is done unless new extensions are added:
16750 so if we don't add custom extensions there is no need to call
16751 X509V3_EXT_cleanup().
16755 * Modify enc utility's salting as follows: make salting the default. Add a
16756 magic header, so unsalted files fail gracefully instead of just decrypting
16757 to garbage. This is because not salting is a big security hole, so people
16758 should be discouraged from doing it.
16762 * Fixes and enhancements to the 'x509' utility. It allowed a message
16763 digest to be passed on the command line but it only used this
16764 parameter when signing a certificate. Modified so all relevant
16765 operations are affected by the digest parameter including the
16766 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16767 DSA key was used because it didn't fix the digest.
16771 * Initial certificate chain verify code. Currently tests the untrusted
16772 certificates for consistency with the verify purpose (which is set
16773 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16775 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16776 this is because it will reject chains with invalid extensions whereas
16777 every previous version of OpenSSL and SSLeay made no checks at all.
16779 Trust code: checks the root CA for the relevant trust settings. Trust
16780 settings have an initial value consistent with the verify purpose: e.g.
16781 if the verify purpose is for SSL client use it expects the CA to be
16782 trusted for SSL client use. However the default value can be changed to
16783 permit custom trust settings: one example of this would be to only trust
16784 certificates from a specific "secure" set of CAs.
16786 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16787 which should be used for version portability: especially since the
16788 verify structure is likely to change more often now.
16790 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16791 to set them. If not set then assume SSL clients will verify SSL servers
16794 Two new options to the verify program: -untrusted allows a set of
16795 untrusted certificates to be passed in and -purpose which sets the
16796 intended purpose of the certificate. If a purpose is set then the
16797 new chain verify code is used to check extension consistency.
16801 * Support for the authority information access extension.
16805 * Modify RSA and DSA PEM read routines to transparently handle
16806 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16807 public keys in a format compatible with certificate
16808 SubjectPublicKeyInfo structures. Unfortunately there were already
16809 functions called *_PublicKey_* which used various odd formats so
16810 these are retained for compatibility: however the DSA variants were
16811 never in a public release so they have been deleted. Changed dsa/rsa
16812 utilities to handle the new format: note no releases ever handled public
16813 keys so we should be OK.
16815 The primary motivation for this change is to avoid the same fiasco
16816 that dogs private keys: there are several incompatible private key
16817 formats some of which are standard and some OpenSSL specific and
16818 require various evil hacks to allow partial transparent handling and
16819 even then it doesn't work with DER formats. Given the option anything
16820 other than PKCS#8 should be dumped: but the other formats have to
16821 stay in the name of compatibility.
16823 With public keys and the benefit of hindsight one standard format
16824 is used which works with EVP_PKEY, RSA or DSA structures: though
16825 it clearly returns an error if you try to read the wrong kind of key.
16827 Added a -pubkey option to the 'x509' utility to output the public key.
16828 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16829 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16830 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16831 that do the same as the `EVP_PKEY_assign_*()` except they up the
16832 reference count of the added key (they don't "swallow" the
16837 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16838 CRLs would fail if the file contained no certificates or no CRLs:
16839 added a new function to read in both types and return the number
16840 read: this means that if none are read it will be an error. The
16841 DER versions of the certificate and CRL reader would always fail
16842 because it isn't possible to mix certificates and CRLs in DER format
16843 without choking one or the other routine. Changed this to just read
16844 a certificate: this is the best we can do. Also modified the code
16845 in `apps/verify.c` to take notice of return codes: it was previously
16846 attempting to read in certificates from NULL pointers and ignoring
16847 any errors: this is one reason why the cert and CRL reader seemed
16848 to work. It doesn't check return codes from the default certificate
16849 routines: these may well fail if the certificates aren't installed.
16853 * Code to support otherName option in GeneralName.
16857 * First update to verify code. Change the verify utility
16858 so it warns if it is passed a self signed certificate:
16859 for consistency with the normal behaviour. X509_verify
16860 has been modified to it will now verify a self signed
16861 certificate if *exactly* the same certificate appears
16862 in the store: it was previously impossible to trust a
16863 single self signed certificate. This means that:
16864 openssl verify ss.pem
16865 now gives a warning about a self signed certificate but
16866 openssl verify -CAfile ss.pem ss.pem
16871 * For servers, store verify_result in SSL_SESSION data structure
16872 (and add it to external session representation).
16873 This is needed when client certificate verifications fails,
16874 but an application-provided verification callback (set by
16875 SSL_CTX_set_cert_verify_callback) allows accepting the session
16876 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16877 but returns 1): When the session is reused, we have to set
16878 ssl->verify_result to the appropriate error code to avoid
16881 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16883 * Fix a bug in the new PKCS#7 code: it didn't consider the
16884 case in PKCS7_dataInit() where the signed PKCS7 structure
16885 didn't contain any existing data because it was being created.
16887 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16889 * Add a salt to the key derivation routines in enc.c. This
16890 forms the first 8 bytes of the encrypted file. Also add a
16891 -S option to allow a salt to be input on the command line.
16895 * New function X509_cmp(). Oddly enough there wasn't a function
16896 to compare two certificates. We do this by working out the SHA1
16897 hash and comparing that. X509_cmp() will be needed by the trust
16902 * SSL_get1_session() is like SSL_get_session(), but increments
16903 the reference count in the SSL_SESSION returned.
16905 *Geoff Thorpe <geoff@eu.c2.net>*
16907 * Fix for 'req': it was adding a null to request attributes.
16908 Also change the X509_LOOKUP and X509_INFO code to handle
16909 certificate auxiliary information.
16913 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16918 * Add the possibility to add extra information to the memory leak
16919 detecting output, to form tracebacks, showing from where each
16920 allocation was originated: CRYPTO_push_info("constant string") adds
16921 the string plus current file name and line number to a per-thread
16922 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16923 is like calling CYRPTO_pop_info() until the stack is empty.
16924 Also updated memory leak detection code to be multi-thread-safe.
16928 * Add options -text and -noout to pkcs7 utility and delete the
16929 encryption options which never did anything. Update docs.
16933 * Add options to some of the utilities to allow the pass phrase
16934 to be included on either the command line (not recommended on
16935 OSes like Unix) or read from the environment. Update the
16936 manpages and fix a few bugs.
16940 * Add a few manpages for some of the openssl commands.
16944 * Fix the -revoke option in ca. It was freeing up memory twice,
16945 leaking and not finding already revoked certificates.
16949 * Extensive changes to support certificate auxiliary information.
16950 This involves the use of X509_CERT_AUX structure and X509_AUX
16951 functions. An X509_AUX function such as PEM_read_X509_AUX()
16952 can still read in a certificate file in the usual way but it
16953 will also read in any additional "auxiliary information". By
16954 doing things this way a fair degree of compatibility can be
16955 retained: existing certificates can have this information added
16956 using the new 'x509' options.
16958 Current auxiliary information includes an "alias" and some trust
16959 settings. The trust settings will ultimately be used in enhanced
16960 certificate chain verification routines: currently a certificate
16961 can only be trusted if it is self signed and then it is trusted
16966 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16967 The problem was that one of the replacement routines had not been working
16968 since SSLeay releases. For now the offending routine has been replaced
16969 with non-optimised assembler. Even so, this now gives around 95%
16970 performance improvement for 1024 bit RSA signs.
16974 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16975 handling. Most clients have the effective key size in bits equal to
16976 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16977 A few however don't do this and instead use the size of the decrypted key
16978 to determine the RC2 key length and the AlgorithmIdentifier to determine
16979 the effective key length. In this case the effective key length can still
16980 be 40 bits but the key length can be 168 bits for example. This is fixed
16981 by manually forcing an RC2 key into the EVP_PKEY structure because the
16982 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16983 the key length and effective key length are equal.
16987 * Add a bunch of functions that should simplify the creation of
16988 X509_NAME structures. Now you should be able to do:
16989 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16990 and have it automatically work out the correct field type and fill in
16991 the structures. The more adventurous can try:
16992 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16993 and it will (hopefully) work out the correct multibyte encoding.
16997 * Change the 'req' utility to use the new field handling and multibyte
16998 copy routines. Before the DN field creation was handled in an ad hoc
16999 way in req, ca, and x509 which was rather broken and didn't support
17000 BMPStrings or UTF8Strings. Since some software doesn't implement
17001 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17002 using the dirstring_type option. See the new comment in the default
17003 openssl.cnf for more info.
17007 * Make crypto/rand/md_rand.c more robust:
17008 - Assure unique random numbers after fork().
17009 - Make sure that concurrent threads access the global counter and
17010 md serializably so that we never lose entropy in them
17011 or use exactly the same state in multiple threads.
17012 Access to the large state is not always serializable because
17013 the additional locking could be a performance killer, and
17014 md should be large enough anyway.
17018 * New file `apps/app_rand.c` with commonly needed functionality
17019 for handling the random seed file.
17021 Use the random seed file in some applications that previously did not:
17023 dsaparam -genkey (which also ignored its '-rand' option),
17026 x509 (when signing).
17027 Except on systems with /dev/urandom, it is crucial to have a random
17028 seed file at least for key creation, DSA signing, and for DH exchanges;
17029 for RSA signatures we could do without one.
17031 gendh and gendsa (unlike genrsa) used to read only the first byte
17032 of each file listed in the '-rand' option. The function as previously
17033 found in genrsa is now in app_rand.c and is used by all programs
17034 that support '-rand'.
17038 * In RAND_write_file, use mode 0600 for creating files;
17039 don't just chmod when it may be too late.
17043 * Report an error from X509_STORE_load_locations
17044 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17048 * New function ASN1_mbstring_copy() this copies a string in either
17049 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17050 into an ASN1_STRING type. A mask of permissible types is passed
17051 and it chooses the "minimal" type to use or an error if not type
17056 * Add function equivalents to the various macros in asn1.h. The old
17057 macros are retained with an `M_` prefix. Code inside the library can
17058 use the `M_` macros. External code (including the openssl utility)
17059 should *NOT* in order to be "shared library friendly".
17063 * Add various functions that can check a certificate's extensions
17064 to see if it usable for various purposes such as SSL client,
17065 server or S/MIME and CAs of these types. This is currently
17066 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17067 verification. Also added a -purpose flag to x509 utility to
17068 print out all the purposes.
17072 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17077 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17078 for, obtain and decode and extension and obtain its critical flag.
17079 This allows all the necessary extension code to be handled in a
17080 single function call.
17084 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17085 platforms. See crypto/rc4/rc4_enc.c for further details.
17089 * New -noout option to asn1parse. This causes no output to be produced
17090 its main use is when combined with -strparse and -out to extract data
17091 from a file (which may not be in ASN.1 format).
17095 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17096 when producing the local key id.
17098 *Richard Levitte <levitte@stacken.kth.se>*
17100 * New option -dhparam in s_server. This allows a DH parameter file to be
17101 stated explicitly. If it is not stated then it tries the first server
17102 certificate file. The previous behaviour hard coded the filename
17107 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17108 a public key to be input or output. For example:
17109 openssl rsa -in key.pem -pubout -out pubkey.pem
17110 Also added necessary DSA public key functions to handle this.
17114 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17115 in the message. This was handled by allowing
17116 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17118 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17120 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17121 to the end of the strings whereas this didn't. This would cause problems
17122 if strings read with d2i_ASN1_bytes() were later modified.
17124 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17126 * Fix for base64 decode bug. When a base64 bio reads only one line of
17127 data and it contains EOF it will end up returning an error. This is
17128 caused by input 46 bytes long. The cause is due to the way base64
17129 BIOs find the start of base64 encoded data. They do this by trying a
17130 trial decode on each line until they find one that works. When they
17131 do a flag is set and it starts again knowing it can pass all the
17132 data directly through the decoder. Unfortunately it doesn't reset
17133 the context it uses. This means that if EOF is reached an attempt
17134 is made to pass two EOFs through the context and this causes the
17135 resulting error. This can also cause other problems as well. As is
17136 usual with these problems it takes *ages* to find and the fix is
17137 trivial: move one line.
17139 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17141 * Ugly workaround to get s_client and s_server working under Windows. The
17142 old code wouldn't work because it needed to select() on sockets and the
17143 tty (for keypresses and to see if data could be written). Win32 only
17144 supports select() on sockets so we select() with a 1s timeout on the
17145 sockets and then see if any characters are waiting to be read, if none
17146 are present then we retry, we also assume we can always write data to
17147 the tty. This isn't nice because the code then blocks until we've
17148 received a complete line of data and it is effectively polling the
17149 keyboard at 1s intervals: however it's quite a bit better than not
17150 working at all :-) A dedicated Windows application might handle this
17151 with an event loop for example.
17155 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17156 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17157 will be called when RSA_sign() and RSA_verify() are used. This is useful
17158 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17159 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17160 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17161 This necessitated the support of an extra signature type NID_md5_sha1
17162 for SSL signatures and modifications to the SSL library to use it instead
17163 of calling RSA_public_decrypt() and RSA_private_encrypt().
17167 * Add new -verify -CAfile and -CApath options to the crl program, these
17168 will lookup a CRL issuers certificate and verify the signature in a
17169 similar way to the verify program. Tidy up the crl program so it
17170 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17171 less strict. It will now permit CRL extensions even if it is not
17172 a V2 CRL: this will allow it to tolerate some broken CRLs.
17176 * Initialize all non-automatic variables each time one of the openssl
17177 sub-programs is started (this is necessary as they may be started
17178 multiple times from the "OpenSSL>" prompt).
17180 *Lennart Bang, Bodo Moeller*
17182 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17183 removing all other RSA functionality (this is what NO_RSA does). This
17184 is so (for example) those in the US can disable those operations covered
17185 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17190 * Non-copying interface to BIO pairs.
17191 (still largely untested)
17195 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17196 ASCII string. This was handled independently in various places before.
17200 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17201 UTF8 strings a character at a time.
17205 * Use client_version from client hello to select the protocol
17206 (s23_srvr.c) and for RSA client key exchange verification
17207 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17211 * Add various utility functions to handle SPKACs, these were previously
17212 handled by poking round in the structure internals. Added new function
17213 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17214 print, verify and generate SPKACs. Based on an original idea from
17215 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17219 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17223 * Allow the config file extension section to be overwritten on the
17224 command line. Based on an original idea from Massimiliano Pala
17225 <madwolf@comune.modena.it>. The new option is called -extensions
17226 and can be applied to ca, req and x509. Also -reqexts to override
17227 the request extensions in req and -crlexts to override the crl extensions
17232 * Add new feature to the SPKAC handling in ca. Now you can include
17233 the same field multiple times by preceding it by "XXXX." for example:
17236 this is the same syntax as used in the req config file.
17240 * Allow certificate extensions to be added to certificate requests. These
17241 are specified in a 'req_extensions' option of the req section of the
17242 config file. They can be printed out with the -text option to req but
17243 are otherwise ignored at present.
17247 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17248 data read consists of only the final block it would not decrypted because
17249 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17250 A misplaced 'break' also meant the decrypted final block might not be
17251 copied until the next read.
17255 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17256 a few extra parameters to the DH structure: these will be useful if
17257 for example we want the value of 'q' or implement X9.42 DH.
17261 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17262 provides hooks that allow the default DSA functions or functions on a
17263 "per key" basis to be replaced. This allows hardware acceleration and
17264 hardware key storage to be handled without major modification to the
17265 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17266 associated functions.
17270 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17271 as "read only": it can't be written to and the buffer it points to will
17272 not be freed. Reading from a read only BIO is much more efficient than
17273 a normal memory BIO. This was added because there are several times when
17274 an area of memory needs to be read from a BIO. The previous method was
17275 to create a memory BIO and write the data to it, this results in two
17276 copies of the data and an O(n^2) reading algorithm. There is a new
17277 function BIO_new_mem_buf() which creates a read only memory BIO from
17278 an area of memory. Also modified the PKCS#7 routines to use read only
17283 * Bugfix: ssl23_get_client_hello did not work properly when called in
17284 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17285 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17286 but a retry condition occurred while trying to read the rest.
17290 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17291 NID_pkcs7_encrypted by default: this was wrong since this should almost
17292 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17293 the encrypted data type: this is a more sensible place to put it and it
17294 allows the PKCS#12 code to be tidied up that duplicated this
17299 * Changed obj_dat.pl script so it takes its input and output files on
17300 the command line. This should avoid shell escape redirection problems
17305 * Initial support for certificate extension requests, these are included
17306 in things like Xenroll certificate requests. Included functions to allow
17307 extensions to be obtained and added.
17311 * -crlf option to s_client and s_server for sending newlines as
17312 CRLF (as required by many protocols).
17316 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17318 * Install libRSAglue.a when OpenSSL is built with RSAref.
17320 *Ralf S. Engelschall*
17322 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17324 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17326 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17331 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17332 DH parameters/keys (q is lost during that conversion, but the resulting
17333 DH parameters contain its length).
17335 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17336 much faster than DH_generate_parameters (which creates parameters
17337 where `p = 2*q + 1`), and also the smaller q makes DH computations
17338 much more efficient (160-bit exponentiation instead of 1024-bit
17339 exponentiation); so this provides a convenient way to support DHE
17340 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17341 utter importance to use
17342 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17344 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17345 when such DH parameters are used, because otherwise small subgroup
17346 attacks may become possible!
17350 * Avoid memory leak in i2d_DHparams.
17354 * Allow the -k option to be used more than once in the enc program:
17355 this allows the same encrypted message to be read by multiple recipients.
17359 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17360 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17361 it will always use the numerical form of the OID, even if it has a short
17366 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17367 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17368 otherwise bn_mod_exp was called. In the case of hardware keys for example
17369 no private key components need be present and it might store extra data
17370 in the RSA structure, which cannot be accessed from bn_mod_exp.
17371 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17372 private key operations.
17376 * Added support for SPARC Linux.
17380 * pem_password_cb function type incompatibly changed from
17381 typedef int pem_password_cb(char *buf, int size, int rwflag);
17383 ....(char *buf, int size, int rwflag, void *userdata);
17384 so that applications can pass data to their callbacks:
17385 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17386 additional void * argument, which is just handed through whenever
17387 the password callback is called.
17389 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17391 New function SSL_CTX_set_default_passwd_cb_userdata.
17393 Compatibility note: As many C implementations push function arguments
17394 onto the stack in reverse order, the new library version is likely to
17395 interoperate with programs that have been compiled with the old
17396 pem_password_cb definition (PEM_whatever takes some data that
17397 happens to be on the stack as its last argument, and the callback
17398 just ignores this garbage); but there is no guarantee whatsoever that
17401 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17402 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17403 problems not only on Windows, but also on some Unix platforms.
17404 To avoid problematic command lines, these definitions are now in an
17405 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17406 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17410 * MIPS III/IV assembler module is reimplemented.
17414 * More DES library cleanups: remove references to srand/rand and
17415 delete an unused file.
17419 * Add support for the free Netwide assembler (NASM) under Win32,
17420 since not many people have MASM (ml) and it can be hard to obtain.
17421 This is currently experimental but it seems to work OK and pass all
17422 the tests. Check out INSTALL.W32 for info.
17426 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17427 without temporary keys kept an extra copy of the server key,
17428 and connections with temporary keys did not free everything in case
17433 * New function RSA_check_key and new openssl rsa option -check
17434 for verifying the consistency of RSA keys.
17436 *Ulf Moeller, Bodo Moeller*
17438 * Various changes to make Win32 compile work:
17439 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17440 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17441 comparison" warnings.
17442 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17446 * Add a debugging option to PKCS#5 v2 key generation function: when
17447 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17448 derived keys are printed to stderr.
17452 * Copy the flags in ASN1_STRING_dup().
17454 *Roman E. Pavlov <pre@mo.msk.ru>*
17456 * The x509 application mishandled signing requests containing DSA
17457 keys when the signing key was also DSA and the parameters didn't match.
17459 It was supposed to omit the parameters when they matched the signing key:
17460 the verifying software was then supposed to automatically use the CA's
17461 parameters if they were absent from the end user certificate.
17463 Omitting parameters is no longer recommended. The test was also
17464 the wrong way round! This was probably due to unusual behaviour in
17465 EVP_cmp_parameters() which returns 1 if the parameters match.
17466 This meant that parameters were omitted when they *didn't* match and
17467 the certificate was useless. Certificates signed with 'ca' didn't have
17470 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17472 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17473 The interface is as follows:
17474 Applications can use
17475 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17476 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17477 "off" is now the default.
17478 The library internally uses
17479 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17480 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17481 to disable memory-checking temporarily.
17483 Some inconsistent states that previously were possible (and were
17484 even the default) are now avoided.
17486 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17487 with each memory chunk allocated; this is occasionally more helpful
17488 than just having a counter.
17490 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17492 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17497 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17498 which largely parallels "options", but is for changing API behaviour,
17499 whereas "options" are about protocol behaviour.
17500 Initial "mode" flags are:
17502 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17503 a single record has been written.
17504 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17505 retries use the same buffer location.
17506 (But all of the contents must be
17511 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17514 * Fix problems with no-hmac etc.
17516 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17518 * New functions RSA_get_default_method(), RSA_set_method() and
17519 RSA_get_method(). These allows replacement of RSA_METHODs without having
17520 to mess around with the internals of an RSA structure.
17524 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17525 Also really enable memory leak checks in openssl.c and in some
17528 *Chad C. Mulligan, Bodo Moeller*
17530 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17531 up the length of negative integers. This has now been simplified to just
17532 store the length when it is first determined and use it later, rather
17533 than trying to keep track of where data is copied and updating it to
17535 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17537 * Add a new function PKCS7_signatureVerify. This allows the verification
17538 of a PKCS#7 signature but with the signing certificate passed to the
17539 function itself. This contrasts with PKCS7_dataVerify which assumes the
17540 certificate is present in the PKCS#7 structure. This isn't always the
17541 case: certificates can be omitted from a PKCS#7 structure and be
17542 distributed by "out of band" means (such as a certificate database).
17546 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17547 function prototypes in pem.h, also change util/mkdef.pl to add the
17548 necessary function names.
17552 * mk1mf.pl (used by Windows builds) did not properly read the
17553 options set by Configure in the top level Makefile, and Configure
17554 was not even able to write more than one option correctly.
17555 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17559 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17560 file to be loaded from a BIO or FILE pointer. The BIO version will
17561 for example allow memory BIOs to contain config info.
17565 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17566 Whoever hopes to achieve shared-library compatibility across versions
17567 must use this, not the compile-time macro.
17568 (Exercise 0.9.4: Which is the minimum library version required by
17570 Note: All this applies only to multi-threaded programs, others don't
17575 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17576 through a BIO pair triggered the default case, i.e.
17577 SSLerr(...,SSL_R_UNKNOWN_STATE).
17581 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17582 can use the SSL library even if none of the specific BIOs is
17587 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17588 for the encoded length.
17590 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17592 * Add initial documentation of the X509V3 functions.
17596 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17597 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17598 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17599 secure PKCS#8 private key format with a high iteration count.
17603 * Fix determination of Perl interpreter: A perl or perl5
17604 *directory* in $PATH was also accepted as the interpreter.
17606 *Ralf S. Engelschall*
17608 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17609 wrong with it but it was very old and did things like calling
17610 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17611 unusual formatting.
17615 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17616 to use the new extension code.
17620 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17621 with macros. This should make it easier to change their form, add extra
17622 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17627 * Add to configuration table a new entry that can specify an alternative
17628 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17629 according to Mark Crispin <MRC@Panda.COM>.
17633 * DES CBC did not update the IV. Weird.
17637 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17638 Changing the behaviour of the former might break existing programs --
17639 where IV updating is needed, des_ncbc_encrypt can be used.
17642 * When bntest is run from "make test" it drives bc to check its
17643 calculations, as well as internally checking them. If an internal check
17644 fails, it needs to cause bc to give a non-zero result or make test carries
17645 on without noticing the failure. Fixed.
17649 * DES library cleanups.
17653 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17654 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17655 ciphers. NOTE: although the key derivation function has been verified
17656 against some published test vectors it has not been extensively tested
17657 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17662 * Instead of "mkdir -p", which is not fully portable, use new
17663 Perl script "util/mkdir-p.pl".
17667 * Rewrite the way password based encryption (PBE) is handled. It used to
17668 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17669 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17670 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17671 the 'parameter' field of the AlgorithmIdentifier is passed to the
17672 underlying key generation function so it must do its own ASN1 parsing.
17673 This has also changed the EVP_PBE_CipherInit() function which now has a
17674 'parameter' argument instead of literal salt and iteration count values
17675 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17679 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17680 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17681 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17682 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17683 value was just used as a "magic string" and not used directly its
17684 value doesn't matter.
17688 * Introduce some semblance of const correctness to BN. Shame C doesn't
17693 * "linux-sparc64" configuration (ultrapenguin).
17695 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17696 "linux-sparc" configuration.
17698 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17700 * config now generates no-xxx options for missing ciphers.
17704 * Support the EBCDIC character set (work in progress).
17705 File ebcdic.c not yet included because it has a different license.
17707 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17709 * Support BS2000/OSD-POSIX.
17711 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17713 * Make callbacks for key generation use `void *` instead of `char *`.
17717 * Make S/MIME samples compile (not yet tested).
17721 * Additional typesafe stacks.
17725 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17729 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17731 * New configuration variant "sco5-gcc".
17733 * Updated some demos.
17735 *Sean O Riordain, Wade Scholine*
17737 * Add missing BIO_free at exit of pkcs12 application.
17741 * Fix memory leak in conf.c.
17745 * Updates for Win32 to assembler version of MD5.
17749 * Set #! path to perl in `apps/der_chop` to where we found it
17750 instead of using a fixed path.
17754 * SHA library changes for irix64-mips4-cc.
17758 * Improvements for VMS support.
17762 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17764 * Bignum library bug fix. IRIX 6 passes "make test" now!
17765 This also avoids the problems with SC4.2 and unpatched SC5.
17767 *Andy Polyakov <appro@fy.chalmers.se>*
17769 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17770 These are required because of the typesafe stack would otherwise break
17771 existing code. If old code used a structure member which used to be STACK
17772 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17773 sk_num or sk_value it would produce an error because the num, data members
17774 are not present in STACK_OF. Now it just produces a warning. sk_set
17775 replaces the old method of assigning a value to sk_value
17776 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17777 that does this will no longer work (and should use sk_set instead) but
17778 this could be regarded as a "questionable" behaviour anyway.
17782 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17783 correctly handle encrypted S/MIME data.
17787 * Change type of various DES function arguments from des_cblock
17788 (which means, in function argument declarations, pointer to char)
17789 to des_cblock * (meaning pointer to array with 8 char elements),
17790 which allows the compiler to do more typechecking; it was like
17791 that back in SSLeay, but with lots of ugly casts.
17793 Introduce new type const_des_cblock.
17797 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17798 problems: find RecipientInfo structure that matches recipient certificate
17799 and initialise the ASN1 structures properly based on passed cipher.
17803 * Belatedly make the BN tests actually check the results.
17807 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17808 to and from BNs: it was completely broken. New compilation option
17809 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17810 key elements as negative integers.
17814 * Reorganize and speed up MD5.
17816 *Andy Polyakov <appro@fy.chalmers.se>*
17820 *Richard Levitte <richard@levitte.org>*
17822 * New option -out to asn1parse to allow the parsed structure to be
17823 output to a file. This is most useful when combined with the -strparse
17824 option to examine the output of things like OCTET STRINGS.
17828 * Make SSL library a little more fool-proof by not requiring any longer
17829 that `SSL_set_{accept,connect}_state` be called before
17830 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17831 in many applications because usually everything *appeared* to work as
17832 intended anyway -- now it really works as intended).
17836 * Move openssl.cnf out of lib/.
17840 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17841 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17842 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17844 *Ralf S. Engelschall*
17846 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17847 handle PKCS#7 enveloped data properly.
17849 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17851 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17852 copying pointers. The cert_st handling is changed by this in
17853 various ways (and thus what used to be known as ctx->default_cert
17854 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17855 any longer when s->cert does not give us what we need).
17856 ssl_cert_instantiate becomes obsolete by this change.
17857 As soon as we've got the new code right (possibly it already is?),
17858 we have solved a couple of bugs of the earlier code where s->cert
17859 was used as if it could not have been shared with other SSL structures.
17861 Note that using the SSL API in certain dirty ways now will result
17862 in different behaviour than observed with earlier library versions:
17863 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17864 does not influence s as it used to.
17866 In order to clean up things more thoroughly, inside SSL_SESSION
17867 we don't use CERT any longer, but a new structure SESS_CERT
17868 that holds per-session data (if available); currently, this is
17869 the peer's certificate chain and, for clients, the server's certificate
17870 and temporary key. CERT holds only those values that can have
17871 meaningful defaults in an SSL_CTX.
17875 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17876 from the internal representation. Various PKCS#7 fixes: remove some
17877 evil casts and set the enc_dig_alg field properly based on the signing
17882 * Allow PKCS#12 password to be set from the command line or the
17883 environment. Let 'ca' get its config file name from the environment
17884 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17889 * Allow certificate policies extension to use an IA5STRING for the
17890 organization field. This is contrary to the PKIX definition but
17891 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17896 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17897 without disallowing inline assembler and the like for non-pedantic builds.
17901 * Support Borland C++ builder.
17903 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17909 * SHA-1 cleanups and performance enhancements.
17911 *Andy Polyakov <appro@fy.chalmers.se>*
17913 * Sparc v8plus assembler for the bignum library.
17915 *Andy Polyakov <appro@fy.chalmers.se>*
17917 * Accept any -xxx and +xxx compiler options in Configure.
17921 * Update HPUX configuration.
17925 * Add missing `sk_<type>_unshift()` function to safestack.h
17927 *Ralf S. Engelschall*
17929 * New function SSL_CTX_use_certificate_chain_file that sets the
17930 "extra_cert"s in addition to the certificate. (This makes sense
17931 only for "PEM" format files, as chains as a whole are not
17936 * Support verify_depth from the SSL API.
17937 x509_vfy.c had what can be considered an off-by-one-error:
17938 Its depth (which was not part of the external interface)
17939 was actually counting the number of certificates in a chain;
17940 now it really counts the depth.
17944 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17945 instead of X509err, which often resulted in confusing error
17946 messages since the error codes are not globally unique
17947 (e.g. an alleged error in ssl3_accept when a certificate
17948 didn't match the private key).
17950 * New function SSL_CTX_set_session_id_context that allows to set a default
17951 value (so that you don't need SSL_set_session_id_context for each
17952 connection using the SSL_CTX).
17956 * OAEP decoding bug fix.
17960 * Support INSTALL_PREFIX for package builders, as proposed by
17965 * New Configure options "threads" and "no-threads". For systems
17966 where the proper compiler options are known (currently Solaris
17967 and Linux), "threads" is the default.
17971 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17975 * Install various scripts to $(OPENSSLDIR)/misc, not to
17976 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17977 such as /usr/local/bin.
17981 * "make linux-shared" to build shared libraries.
17983 *Niels Poppe <niels@netbox.org>*
17985 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17989 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17990 extension adding in x509 utility.
17994 * Remove NOPROTO sections and error code comments.
17998 * Partial rewrite of the DEF file generator to now parse the ANSI
18003 * New Configure options --prefix=DIR and --openssldir=DIR.
18007 * Complete rewrite of the error code script(s). It is all now handled
18008 by one script at the top level which handles error code gathering,
18009 header rewriting and C source file generation. It should be much better
18010 than the old method: it now uses a modified version of Ulf's parser to
18011 read the ANSI prototypes in all header files (thus the old K&R definitions
18012 aren't needed for error creation any more) and do a better job of
18013 translating function codes into names. The old 'ASN1 error code embedded
18014 in a comment' is no longer necessary and it doesn't use .err files which
18015 have now been deleted. Also the error code call doesn't have to appear all
18016 on one line (which resulted in some large lines...).
18020 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18024 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18025 0 (which usually indicates a closed connection), but continue reading.
18029 * Fix some race conditions.
18033 * Add support for CRL distribution points extension. Add Certificate
18034 Policies and CRL distribution points documentation.
18038 * Move the autogenerated header file parts to crypto/opensslconf.h.
18042 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18043 8 of keying material. Merlin has also confirmed interop with this fix
18044 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18046 *Merlin Hughes <merlin@baltimore.ie>*
18048 * Fix lots of warnings.
18050 *Richard Levitte <levitte@stacken.kth.se>*
18052 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18053 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18055 *Richard Levitte <levitte@stacken.kth.se>*
18057 * Fix problems with sizeof(long) == 8.
18059 *Andy Polyakov <appro@fy.chalmers.se>*
18061 * Change functions to ANSI C.
18065 * Fix typos in error codes.
18067 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18069 * Remove defunct assembler files from Configure.
18073 * SPARC v8 assembler BIGNUM implementation.
18075 *Andy Polyakov <appro@fy.chalmers.se>*
18077 * Support for Certificate Policies extension: both print and set.
18078 Various additions to support the r2i method this uses.
18082 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18083 return a const string when you are expecting an allocated buffer.
18087 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18088 types DirectoryString and DisplayText.
18092 * Add code to allow r2i extensions to access the configuration database,
18093 add an LHASH database driver and add several ctx helper functions.
18097 * Fix an evil bug in bn_expand2() which caused various BN functions to
18098 fail when they extended the size of a BIGNUM.
18102 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18103 support typesafe stack.
18107 * Fix typo in SSL_[gs]et_options().
18109 *Nils Frostberg <nils@medcom.se>*
18111 * Delete various functions and files that belonged to the (now obsolete)
18112 old X509V3 handling code.
18116 * New Configure option "rsaref".
18120 * Don't auto-generate pem.h.
18124 * Introduce type-safe ASN.1 SETs.
18128 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18130 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18132 * Introduce type-safe STACKs. This will almost certainly break lots of code
18133 that links with OpenSSL (well at least cause lots of warnings), but fear
18134 not: the conversion is trivial, and it eliminates loads of evil casts. A
18135 few STACKed things have been converted already. Feel free to convert more.
18136 In the fullness of time, I'll do away with the STACK type altogether.
18140 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18141 specified in `<certfile>` by updating the entry in the index.txt file.
18142 This way one no longer has to edit the index.txt file manually for
18143 revoking a certificate. The -revoke option does the gory details now.
18145 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18147 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18148 `-text` option at all and this way the `-noout -text` combination was
18149 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18151 *Ralf S. Engelschall*
18153 * Make sure a corresponding plain text error message exists for the
18154 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18155 verify callback function determined that a certificate was revoked.
18157 *Ralf S. Engelschall*
18159 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18160 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18161 all available ciphers including rc5, which was forgotten until now.
18162 In order to let the testing shell script know which algorithms
18163 are available, a new (up to now undocumented) command
18164 `openssl list-cipher-commands` is used.
18168 * Bugfix: s_client occasionally would sleep in select() when
18169 it should have checked SSL_pending() first.
18173 * New functions DSA_do_sign and DSA_do_verify to provide access to
18174 the raw DSA values prior to ASN.1 encoding.
18178 * Tweaks to Configure
18180 *Niels Poppe <niels@netbox.org>*
18182 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18187 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18191 * New config option to avoid instructions that are illegal on the 80386.
18192 The default code is faster, but requires at least a 486.
18196 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18197 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18198 same as SSL2_VERSION anyway.
18202 * New "-showcerts" option for s_client.
18206 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18207 application. Various cleanups and fixes.
18211 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18212 modify error routines to work internally. Add error codes and PBE init
18213 to library startup routines.
18217 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18218 packing functions to asn1 and evp. Changed function names and error
18219 codes along the way.
18223 * PKCS12 integration: and so it begins... First of several patches to
18224 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18225 objects to objects.h
18229 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18230 and display support for Thawte strong extranet extension.
18234 * Add LinuxPPC support.
18236 *Jeff Dubrule <igor@pobox.org>*
18238 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18239 bn_div_words in alpha.s.
18241 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18243 * Make sure the RSA OAEP test is skipped under -DRSAref because
18244 OAEP isn't supported when OpenSSL is built with RSAref.
18246 *Ulf Moeller <ulf@fitug.de>*
18248 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18249 so they no longer are missing under -DNOPROTO.
18251 *Soren S. Jorvang <soren@t.dk>*
18253 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18255 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18256 doesn't work when the session is reused. Coming soon!
18260 * Fix a security hole, that allows sessions to be reused in the wrong
18261 context thus bypassing client cert protection! All software that uses
18262 client certs and session caches in multiple contexts NEEDS PATCHING to
18263 allow session reuse! A fuller solution is in the works.
18265 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18267 * Some more source tree cleanups (removed obsolete files
18268 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18269 permission on "config" script to be executable) and a fix for the INSTALL
18272 *Ulf Moeller <ulf@fitug.de>*
18274 * Remove some legacy and erroneous uses of malloc, free instead of
18277 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18279 * Make rsa_oaep_test return non-zero on error.
18281 *Ulf Moeller <ulf@fitug.de>*
18283 * Add support for native Solaris shared libraries. Configure
18284 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18285 if someone would make that last step automatic.
18287 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18289 * ctx_size was not built with the right compiler during "make links". Fixed.
18293 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18294 except NULL ciphers". This means the default cipher list will no longer
18295 enable NULL ciphers. They need to be specifically enabled e.g. with
18296 the string "DEFAULT:eNULL".
18300 * Fix to RSA private encryption routines: if p < q then it would
18301 occasionally produce an invalid result. This will only happen with
18302 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18306 * Be less restrictive and allow also `perl util/perlpath.pl
18307 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18308 because this way one can also use an interpreter named `perl5` (which is
18309 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18310 installed as `perl`).
18312 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18314 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18316 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18318 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18319 advapi32.lib to Win32 build and change the pem test comparison
18320 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18321 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18322 and crypto/des/ede_cbcm_enc.c.
18326 * DES quad checksum was broken on big-endian architectures. Fixed.
18330 * Comment out two functions in bio.h that aren't implemented. Fix up the
18331 Win32 test batch file so it (might) work again. The Win32 test batch file
18332 is horrible: I feel ill....
18336 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18337 in e_os.h. Audit of header files to check ANSI and non ANSI
18338 sections: 10 functions were absent from non ANSI section and not exported
18339 from Windows DLLs. Fixed up libeay.num for new functions.
18343 * Make `openssl version` output lines consistent.
18345 *Ralf S. Engelschall*
18347 * Fix Win32 symbol export lists for BIO functions: Added
18348 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18349 to ms/libeay{16,32}.def.
18351 *Ralf S. Engelschall*
18353 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18354 fine under Unix and passes some trivial tests I've now added. But the
18355 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18356 added to make sure no one expects that this stuff really works in the
18357 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18358 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18361 *Ralf S. Engelschall*
18363 * Fix the generation of two part addresses in perl.
18365 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18367 * Add config entry for Linux on MIPS.
18369 *John Tobey <jtobey@channel1.com>*
18371 * Make links whenever Configure is run, unless we are on Windoze.
18375 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18376 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18381 * Add a useful kludge to allow package maintainers to specify compiler and
18382 other platforms details on the command line without having to patch the
18383 Configure script every time: One now can use
18384 `perl Configure <id>:<details>`,
18385 i.e. platform ids are allowed to have details appended
18386 to them (separated by colons). This is treated as there would be a static
18387 pre-configured entry in Configure's %table under key `<id>` with value
18388 `<details>` and `perl Configure <id>` is called. So, when you want to
18389 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18390 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18391 now, which overrides the FreeBSD-elf entry on-the-fly.
18393 *Ralf S. Engelschall*
18395 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18399 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18400 on the `perl Configure ...` command line. This way one can compile
18401 OpenSSL libraries with Position Independent Code (PIC) which is needed
18402 for linking it into DSOs.
18404 *Ralf S. Engelschall*
18406 * Remarkably, export ciphers were totally broken and no-one had noticed!
18411 * Cleaned up the LICENSE document: The official contact for any license
18412 questions now is the OpenSSL core team under openssl-core@openssl.org.
18413 And add a paragraph about the dual-license situation to make sure people
18414 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18415 to the OpenSSL toolkit.
18417 *Ralf S. Engelschall*
18419 * General source tree makefile cleanups: Made `making xxx in yyy...`
18420 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18421 Additionally cleaned up the `make links` target: Remove unnecessary
18422 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18423 to speed processing and no longer clutter the display with confusing
18424 stuff. Instead only the actually done links are displayed.
18426 *Ralf S. Engelschall*
18428 * Permit null encryption ciphersuites, used for authentication only. It used
18429 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18430 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18435 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18436 signed attributes when verifying signatures (this would break them),
18437 the detached data encoding was wrong and public keys obtained using
18438 X509_get_pubkey() weren't freed.
18442 * Add text documentation for the BUFFER functions. Also added a work around
18443 to a Win95 console bug. This was triggered by the password read stuff: the
18444 last character typed gets carried over to the next fread(). If you were
18445 generating a new cert request using 'req' for example then the last
18446 character of the passphrase would be CR which would then enter the first
18451 * Added the new 'Includes OpenSSL Cryptography Software' button as
18452 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18453 button and can be used by applications based on OpenSSL to show the
18454 relationship to the OpenSSL project.
18456 *Ralf S. Engelschall*
18458 * Remove confusing variables in function signatures in files
18459 ssl/ssl_lib.c and ssl/ssl.h.
18461 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18463 * Don't install bss_file.c under PREFIX/include/
18465 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18467 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18468 functions that return function pointers and has support for NT specific
18469 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18470 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18471 unsigned to signed types: this was killing the Win32 compile.
18475 * Add new certificate file to stack functions,
18476 SSL_add_dir_cert_subjects_to_stack() and
18477 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18478 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18479 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18480 This means that Apache-SSL and similar packages don't have to mess around
18481 to add as many CAs as they want to the preferred list.
18485 * Experiment with doxygen documentation. Currently only partially applied to
18487 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18488 openssl.doxy as the configuration file.
18492 * Get rid of remaining C++-style comments which strict C compilers hate.
18494 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18496 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18497 compiled in by default: it has problems with large keys.
18501 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18502 DH private keys and/or callback functions which directly correspond to
18503 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18504 is needed for applications which have to configure certificates on a
18505 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18507 For the RSA certificate situation is makes no difference, but
18508 for the DSA certificate situation this fixes the "no shared cipher"
18509 problem where the OpenSSL cipher selection procedure failed because the
18510 temporary keys were not overtaken from the context and the API provided
18511 no way to reconfigure them.
18512 The new functions now let applications reconfigure the stuff and they
18513 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18514 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18515 non-public-API function ssl_cert_instantiate() is used as a helper
18516 function and also to reduce code redundancy inside ssl_rsa.c.
18518 *Ralf S. Engelschall*
18520 * Move s_server -dcert and -dkey options out of the undocumented feature
18521 area because they are useful for the DSA situation and should be
18522 recognized by the users.
18524 *Ralf S. Engelschall*
18526 * Fix the cipher decision scheme for export ciphers: the export bits are
18527 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18528 SSL_EXP_MASK. So, the original variable has to be used instead of the
18529 already masked variable.
18531 *Richard Levitte <levitte@stacken.kth.se>*
18533 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18535 *Richard Levitte <levitte@stacken.kth.se>*
18537 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18538 from `int` to `unsigned int` because it is a length and initialized by
18539 EVP_DigestFinal() which expects an `unsigned int *`.
18541 *Richard Levitte <levitte@stacken.kth.se>*
18543 * Don't hard-code path to Perl interpreter on shebang line of Configure
18544 script. Instead use the usual Shell->Perl transition trick.
18546 *Ralf S. Engelschall*
18548 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18549 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18550 -noout -modulus` as it's already the case for `openssl rsa -noout
18551 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18552 currently the public key is printed (a decision which was already done by
18553 `openssl dsa -modulus` in the past) which serves a similar purpose.
18554 Additionally the NO_RSA no longer completely removes the whole -modulus
18555 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18558 *Ralf S. Engelschall*
18560 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18561 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18563 *Arne Ansper <arne@ats.cyber.ee>*
18565 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18566 to be added. Now both 'req' and 'ca' can use new objects defined in the
18571 * Add cool BIO that does syslog (or event log on NT).
18573 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18575 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18576 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18577 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18578 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18582 * Add preliminary config info for new extension code.
18586 * Make RSA_NO_PADDING really use no padding.
18588 *Ulf Moeller <ulf@fitug.de>*
18590 * Generate errors when private/public key check is done.
18594 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18595 for some CRL extensions and new objects added.
18599 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18600 key usage extension and fuller support for authority key id.
18604 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18605 padding method for RSA, which is recommended for new applications in PKCS
18606 #1 v2.0 (RFC 2437, October 1998).
18607 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18608 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18609 against Bleichbacher's attack on RSA.
18610 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18613 * Updates to the new SSL compression code
18615 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18617 * Fix so that the version number in the master secret, when passed
18618 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18619 (because the server will not accept higher), that the version number
18620 is 0x03,0x01, not 0x03,0x00
18622 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18624 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18625 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18626 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18630 * Support for RAW extensions where an arbitrary extension can be
18631 created by including its DER encoding. See `apps/openssl.cnf` for
18636 * Make sure latest Perl versions don't interpret some generated C array
18637 code as Perl array code in the crypto/err/err_genc.pl script.
18639 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18641 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18642 not many people have the assembler. Various Win32 compilation fixes and
18643 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18644 build instructions.
18648 * Modify configure script 'Configure' to automatically create crypto/date.h
18649 file under Win32 and also build pem.h from pem.org. New script
18650 util/mkfiles.pl to create the MINFO file on environments that can't do a
18651 'make files': perl util/mkfiles.pl >MINFO should work.
18655 * Major rework of DES function declarations, in the pursuit of correctness
18656 and purity. As a result, many evil casts evaporated, and some weirdness,
18657 too. You may find this causes warnings in your code. Zapping your evil
18658 casts will probably fix them. Mostly.
18662 * Fix for a typo in asn1.h. Bug fix to object creation script
18663 obj_dat.pl. It considered a zero in an object definition to mean
18664 "end of object": none of the objects in objects.h have any zeros
18665 so it wasn't spotted.
18667 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18669 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18670 Masking (CBCM). In the absence of test vectors, the best I have been able
18671 to do is check that the decrypt undoes the encrypt, so far. Send me test
18672 vectors if you have them.
18676 * Correct calculation of key length for export ciphers (too much space was
18677 allocated for null ciphers). This has not been tested!
18681 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18682 message is now correct (it understands "crypto" and "ssl" on its
18683 command line). There is also now an "update" option. This will update
18684 the util/ssleay.num and util/libeay.num files with any new functions.
18686 perl util/mkdef.pl crypto ssl update
18687 it will update them.
18691 * Overhauled the Perl interface:
18692 - ported BN stuff to OpenSSL's different BN library
18693 - made the perl/ source tree CVS-aware
18694 - renamed the package from SSLeay to OpenSSL (the files still contain
18695 their history because I've copied them in the repository)
18696 - removed obsolete files (the test scripts will be replaced
18697 by better Test::Harness variants in the future)
18699 *Ralf S. Engelschall*
18701 * First cut for a very conservative source tree cleanup:
18702 1. merge various obsolete readme texts into doc/ssleay.txt
18703 where we collect the old documents and readme texts.
18704 2. remove the first part of files where I'm already sure that we no
18705 longer need them because of three reasons: either they are just temporary
18706 files which were left by Eric or they are preserved original files where
18707 I've verified that the diff is also available in the CVS via "cvs diff
18708 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18709 the crypto/md/ stuff).
18711 *Ralf S. Engelschall*
18713 * More extension code. Incomplete support for subject and issuer alt
18714 name, issuer and authority key id. Change the i2v function parameters
18715 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18716 what that's for :-) Fix to ASN1 macro which messed up
18717 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18721 * Preliminary support for ENUMERATED type. This is largely copied from the
18726 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18728 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18730 * Make sure `make rehash` target really finds the `openssl` program.
18732 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18734 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18735 like to hear about it if this slows down other processors.
18739 * Add CygWin32 platform information to Configure script.
18741 *Alan Batie <batie@aahz.jf.intel.com>*
18743 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18745 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18747 * New program nseq to manipulate netscape certificate sequences
18751 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18756 * Fixes to BN code. Previously the default was to define BN_RECURSION
18757 but the BN code had some problems that would cause failures when
18758 doing certificate verification and some other functions.
18760 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18762 * Add ASN1 and PEM code to support netscape certificate sequences.
18766 * Add ASN1 and PEM code to support netscape certificate sequences.
18770 * Add several PKIX and private extended key usage OIDs.
18774 * Modify the 'ca' program to handle the new extension code. Modify
18775 openssl.cnf for new extension format, add comments.
18779 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18780 and add a sample to openssl.cnf so req -x509 now adds appropriate
18785 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18786 error code, add initial support to X509_print() and x509 application.
18790 * Takes a deep breath and start adding X509 V3 extension support code. Add
18791 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18792 stuff is currently isolated and isn't even compiled yet.
18796 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18797 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18798 Removed the versions check from X509 routines when loading extensions:
18799 this allows certain broken certificates that don't set the version
18800 properly to be processed.
18804 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18805 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18806 can still be regenerated with "make depend".
18810 * Spelling mistake in C version of CAST-128.
18812 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18814 * Changes to the error generation code. The perl script err-code.pl
18815 now reads in the old error codes and retains the old numbers, only
18816 adding new ones if necessary. It also only changes the .err files if new
18817 codes are added. The makefiles have been modified to only insert errors
18818 when needed (to avoid needlessly modifying header files). This is done
18819 by only inserting errors if the .err file is newer than the auto generated
18820 C file. To rebuild all the error codes from scratch (the old behaviour)
18821 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18822 or delete all the .err files.
18826 * CAST-128 was incorrectly implemented for short keys. The C version has
18827 been fixed, but is untested. The assembler versions are also fixed, but
18828 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18829 to regenerate it if needed.
18830 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18831 Hagino <itojun@kame.net>*
18833 * File was opened incorrectly in randfile.c.
18835 *Ulf Möller <ulf@fitug.de>*
18837 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18838 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18839 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18840 al: it's just almost always a UTCTime. Note this patch adds new error
18841 codes so do a "make errors" if there are problems.
18845 * Correct Linux 1 recognition in config.
18847 *Ulf Möller <ulf@fitug.de>*
18849 * Remove pointless MD5 hash when using DSA keys in ca.
18851 *Anonymous <nobody@replay.com>*
18853 * Generate an error if given an empty string as a cert directory. Also
18854 generate an error if handed NULL (previously returned 0 to indicate an
18855 error, but didn't set one).
18857 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18859 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18863 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18864 parameters. This was causing a warning which killed off the Win32 compile.
18868 * Remove C++ style comments from crypto/bn/bn_local.h.
18870 *Neil Costigan <neil.costigan@celocom.com>*
18872 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18873 based on a text string, looking up short and long names and finally
18874 "dot" format. The "dot" format stuff didn't work. Added new function
18875 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18876 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18877 OID is not part of the table.
18881 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18882 X509_LOOKUP_by_alias().
18886 * Sort openssl functions by name.
18890 * Get the `gendsa` command working and add it to the `list` command. Remove
18891 encryption from sample DSA keys (in case anyone is interested the password
18896 * Make *all* `*_free` functions accept a NULL pointer.
18898 *Frans Heymans <fheymans@isaserver.be>*
18900 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18903 *Anonymous <nobody@replay.com>*
18905 * s_server should send the CAfile as acceptable CAs, not its own cert.
18907 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18909 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18911 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18913 * Temp key "for export" tests were wrong in s3_srvr.c.
18915 *Anonymous <nobody@replay.com>*
18917 * Add prototype for temp key callback functions
18918 SSL_CTX_set_tmp_{rsa,dh}_callback().
18922 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18923 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18927 * X509_name_add_entry() freed the wrong thing after an error.
18929 *Arne Ansper <arne@ats.cyber.ee>*
18931 * rsa_eay.c would attempt to free a NULL context.
18933 *Arne Ansper <arne@ats.cyber.ee>*
18935 * BIO_s_socket() had a broken should_retry() on Windoze.
18937 *Arne Ansper <arne@ats.cyber.ee>*
18939 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18941 *Arne Ansper <arne@ats.cyber.ee>*
18943 * Make sure the already existing X509_STORE->depth variable is initialized
18944 in X509_STORE_new(), but document the fact that this variable is still
18945 unused in the certificate verification process.
18947 *Ralf S. Engelschall*
18949 * Fix the various library and `apps/` files to free up pkeys obtained from
18950 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18954 * Fix reference counting in X509_PUBKEY_get(). This makes
18955 demos/maurice/example2.c work, amongst others, probably.
18957 *Steve Henson and Ben Laurie*
18959 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18960 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18961 are no longer created. This way we have a single and consistent command
18962 line interface `openssl <command>`, similar to `cvs <command>`.
18964 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18966 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18967 BIT STRING wrapper always have zero unused bits.
18971 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18975 * Make the top-level INSTALL documentation easier to understand.
18979 * Makefiles updated to exit if an error occurs in a sub-directory
18980 make (including if user presses ^C) [Paul Sutton]
18982 * Make Montgomery context stuff explicit in RSA data structure.
18986 * Fix build order of pem and err to allow for generated pem.h.
18990 * Fix renumbering bug in X509_NAME_delete_entry().
18994 * Enhanced the err-ins.pl script so it makes the error library number
18995 global and can add a library name. This is needed for external ASN1 and
18996 other error libraries.
19000 * Fixed sk_insert which never worked properly.
19004 * Fix ASN1 macros so they can handle indefinite length constructed
19005 EXPLICIT tags. Some non standard certificates use these: they can now
19010 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19011 into a single doc/ssleay.txt bundle. This way the information is still
19012 preserved but no longer messes up this directory. Now it's new room for
19013 the new set of documentation files.
19015 *Ralf S. Engelschall*
19017 * SETs were incorrectly DER encoded. This was a major pain, because they
19018 shared code with SEQUENCEs, which aren't coded the same. This means that
19019 almost everything to do with SETs or SEQUENCEs has either changed name or
19020 number of arguments.
19022 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19024 * Fix test data to work with the above.
19028 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19029 was already fixed by Eric for 0.9.1 it seems.
19031 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19033 * Autodetect FreeBSD3.
19037 * Fix various bugs in Configure. This affects the following platforms:
19041 unixware-2.0-pentium
19046 * Eliminate generated files from CVS. Reorder tests to regenerate files
19047 before they are needed.
19051 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19055 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19057 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19058 changed SSLeay to OpenSSL in version strings.
19060 *Ralf S. Engelschall*
19062 * Some fixups to the top-level documents.
19066 * Fixed the nasty bug where rsaref.h was not found under compile-time
19067 because the symlink to include/ was missing.
19069 *Ralf S. Engelschall*
19071 * Incorporated the popular no-RSA/DSA-only patches
19072 which allow to compile a RSA-free SSLeay.
19074 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19076 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19077 when "ssleay" is still not found.
19079 *Ralf S. Engelschall*
19081 * Added more platforms to Configure: Cray T3E, HPUX 11,
19083 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19085 * Updated the README file.
19087 *Ralf S. Engelschall*
19089 * Added various .cvsignore files in the CVS repository subdirs
19090 to make a "cvs update" really silent.
19092 *Ralf S. Engelschall*
19094 * Recompiled the error-definition header files and added
19095 missing symbols to the Win32 linker tables.
19097 *Ralf S. Engelschall*
19099 * Cleaned up the top-level documents;
19100 o new files: CHANGES and LICENSE
19101 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19102 o merged COPYRIGHT into LICENSE
19103 o removed obsolete TODO file
19104 o renamed MICROSOFT to INSTALL.W32
19106 *Ralf S. Engelschall*
19108 * Removed dummy files from the 0.9.1b source tree:
19109 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19110 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19111 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19112 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19113 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19115 *Ralf S. Engelschall*
19117 * Added various platform portability fixes.
19121 * The Genesis of the OpenSSL rpject:
19122 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19123 Young and Tim J. Hudson created while they were working for C2Net until
19126 *The OpenSSL Project*
19128 ### Changes between 0.9.0b and 0.9.1b [not released]
19130 * Updated a few CA certificates under certs/
19134 * Changed some BIGNUM api stuff.
19138 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19139 DGUX x86, Linux Alpha, etc.
19143 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19144 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19149 * Add -strparse option to asn1pars program which parses nested
19152 *Dr Stephen Henson <shenson@bigfoot.com>*
19154 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19158 * DSA fix for "ca" program.
19162 * Added "-genkey" option to "dsaparam" program.
19166 * Added RIPE MD160 (rmd160) message digest.
19170 * Added -a (all) option to "ssleay version" command.
19174 * Added PLATFORM define which is the id given to Configure.
19178 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19182 * Extended the ASN.1 parser routines.
19186 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19190 * Added a BN_CTX to the BN library.
19194 * Fixed the weak key values in DES library
19198 * Changed API in EVP library for cipher aliases.
19202 * Added support for RC2/64bit cipher.
19206 * Converted the lhash library to the crypto/mem.c functions.
19210 * Added more recognized ASN.1 object ids.
19214 * Added more RSA padding checks for SSL/TLS.
19218 * Added BIO proxy/filter functionality.
19222 * Added extra_certs to SSL_CTX which can be used
19223 send extra CA certificates to the client in the CA cert chain sending
19224 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19228 * Now Fortezza is denied in the authentication phase because
19229 this is key exchange mechanism is not supported by SSLeay at all.
19233 * Additional PKCS1 checks.
19237 * Support the string "TLSv1" for all TLS v1 ciphers.
19241 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19242 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19246 * Fixed a few memory leaks.
19250 * Fixed various code and comment typos.
19254 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19255 bytes sent in the client random.
19257 *Edward Bishop <ebishop@spyglass.com>*
19261 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19262 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19263 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19264 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19265 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19266 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19267 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19268 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19269 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19270 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19271 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19272 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19273 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19274 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19275 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19276 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19277 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19278 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19279 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19280 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19281 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19282 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19283 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19284 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19285 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19286 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19287 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19288 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19289 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19290 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19291 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19292 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19293 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19294 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19295 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19296 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19297 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19298 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19299 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19300 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19301 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19302 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19303 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19304 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19305 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19306 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19307 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19308 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19309 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19310 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19311 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19312 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19313 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19314 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19315 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19316 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19317 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19318 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19319 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19320 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19321 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19322 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19323 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19324 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19325 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19326 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19327 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19328 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19329 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19330 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19331 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19332 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19333 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19334 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19335 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19336 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19337 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19338 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19339 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19340 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19341 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19342 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19343 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19344 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19345 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19346 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19347 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19348 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19349 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19350 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19351 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19352 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19353 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19354 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19355 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19356 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19357 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19358 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19359 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19360 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19361 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19362 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19363 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19364 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19365 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19366 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19367 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19368 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19369 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19370 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19371 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19372 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19373 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19374 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19375 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19376 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19377 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19378 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19379 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19380 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19381 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19382 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19383 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19384 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19385 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19386 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19387 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19388 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19389 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19390 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19391 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19392 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19393 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19394 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19395 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19396 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19397 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19398 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19399 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19400 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19401 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19402 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19403 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19404 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19405 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19406 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19407 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19408 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19409 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19410 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19411 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19412 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19413 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19414 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19415 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19416 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19417 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19418 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19419 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19420 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19421 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19422 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19423 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19424 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655