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 3.2](#openssl-32)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.2 [xx XXX xxxx]
27 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT parameters
29 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
30 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
31 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
32 Using these invalid names may cause algorithms to use slower methods
33 that ignore the CRT parameters.
37 * Add support for certificate compression (RFC8879), including
38 library support for Brotli and Zstandard compression.
42 * Add the ability to add custom attributes to PKCS12 files. Add a new API
43 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
44 for a user specified callback and optional argument.
45 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
46 added to the existing STACK_OF attrs.
50 * Major refactor of the libssl record layer
54 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
58 * Add a mac salt length option for the pkcs12 command.
62 * Add more SRTP protection profiles from RFC8723 and RFC8269.
66 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
68 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
70 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
71 supported and enabled.
75 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
76 to the list of ciphersuites providing Perfect Forward Secrecy as
77 required by SECLEVEL >= 3.
79 *Dmitry Belyavskiy, Nicola Tuveri*
81 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
82 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
83 SSL_get0_iana_groups() function-like macro, retrieves the list of
84 supported groups sent by the peer.
85 The function SSL_client_hello_get_extension_order() populates
86 a caller-supplied array with the list of extension types present in the
87 ClientHello, in order of appearance.
91 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
92 to make it possible to use empty passphrase strings.
96 * RNDR and RNDRRS support in provider functions to provide
97 random number generation for Arm CPUs (aarch64).
101 * s_client and s_server apps now explicitly say when the TLS version
102 does not include the renegotiation mechanism. This avoids confusion
103 between that scenario versus when the TLS version includes secure
104 renegotiation but the peer lacks support for it.
108 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
110 *Tomasz Kantecki, Andrey Matyukov*
112 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
113 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
114 of 160 bits and above and less than 224 bits were previously accepted by
115 default but are now no longer allowed. By default TLS compression was
116 already disabled in previous OpenSSL versions. At security level 2 it cannot
121 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
126 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
127 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
128 will need to load the legacy crypto provider.
132 * The various OBJ_* functions have been made thread safe.
136 * CCM8 cipher suites in TLS have been downgraded to security level zero
137 because they use a short authentication tag which lowers their strength.
141 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
146 * Parallel dual-prime 1536/2048-bit modular exponentiation for
147 AVX512_IFMA capable processors.
149 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
151 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
152 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
153 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
154 marked deprecated from OpenSSL 3.2 onwards and can be disabled by defining
155 `OPENSSL_NO_DEPRECATED_3_2`.
157 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
158 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
159 definitions for these functions regardless of whether
160 `OPENSSL_NO_DEPRECATED_3_2` is defined.
162 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
163 functions regardless of whether they are using them. It is recommended that
164 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
168 * When generating safe-prime DH parameters set the recommended private key
169 length equivalent to minimum key lengths as in RFC 7919.
173 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
174 such as adding a trace facility for debugging certificate chain building.
178 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
179 in particular supporting requests for central key generation, generalized
180 polling, and various types of genm/genp exchanges defined in CMP Updates.
184 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
185 like correcting the TLS and proxy support and adding tracing for debugging.
189 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
193 * Fixed and extended `util/check-format.pl` for checking adherence to the
194 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
195 The checks are meanwhile more complete and yield fewer false positives.
199 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
200 sending and receiving multiple messages in a single call. An implementation
201 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
205 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
206 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
207 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
208 paths which are searched for root certificates.
210 The existing `SSL_CERT_DIR` environment variable is deprecated.
211 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
212 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
213 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
214 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
215 for the purposes of determining root certificate stores.
219 * Support for loading root certificates from the Windows certificate store
220 has been added. The support is in the form of a store which recognises the
221 URI string of `org.openssl.winstore://`. This store is enabled by default and
222 can be disabled using the new compile-time option `no-winstore`.
229 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
230 listed here are only a brief description.
231 The migration guide contains more detailed information related to new features,
232 breaking changes, and mappings for the large list of deprecated functions.
234 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
236 ### Changes between 3.0.6 and 3.0.7 [xx XXX xxxx]
238 * Added RIPEMD160 to the default provider.
242 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
244 * OpenSSL supports creating a custom cipher via the legacy
245 EVP_CIPHER_meth_new() function and associated function calls. This function
246 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
247 to use the new provider mechanism in order to implement custom ciphers.
249 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
250 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
251 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
252 and decryption initialisation functions). Instead of using the custom cipher
253 directly it incorrectly tries to fetch an equivalent cipher from the
254 available providers. An equivalent cipher is found based on the NID passed to
255 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
256 given cipher. However it is possible for an application to incorrectly pass
257 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
258 is used in this way the OpenSSL encryption/decryption initialisation function
259 will match the NULL cipher as being equivalent and will fetch this from the
260 available providers. This will succeed if the default provider has been
261 loaded (or if a third party provider has been loaded that offers this
262 cipher). Using the NULL cipher means that the plaintext is emitted as the
265 Applications are only affected by this issue if they call
266 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
267 encryption/decryption initialisation function. Applications that only use
268 SSL/TLS are not impacted by this issue.
273 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
278 * Fixed the linux-mips64 Configure target which was missing the
279 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
284 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
289 * Correctly handle a retransmitted ClientHello in DTLS
293 * Fixed detection of ktls support in cross-compile environment on Linux
297 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
302 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
303 report correct results in some cases
307 * Fix UWP builds by defining VirtualLock
311 * For known safe primes use the minimum key length according to RFC 7919.
312 Longer private key sizes unnecessarily raise the cycles needed to compute the
313 shared secret without any increase of the real security. This fixes a
314 regression from 1.1.1 where these shorter keys were generated for the known
319 * Added the loongarch64 target
323 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
324 only passed to the FIPS provider and not to the default or legacy provider.
328 * Fixed reported performance degradation on aarch64. Restored the
329 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
330 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
331 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
332 The new algorithm is still used for 32 bit targets.
336 * Added a missing header for memcmp that caused compilation failure on some
341 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
343 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
344 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
345 This issue makes the RSA implementation with 2048 bit private keys
346 incorrect on such machines and memory corruption will happen during
347 the computation. As a consequence of the memory corruption an attacker
348 may be able to trigger a remote code execution on the machine performing
351 SSL/TLS servers or other servers using 2048 bit RSA private keys running
352 on machines supporting AVX512IFMA instructions of the X86_64 architecture
353 are affected by this issue.
358 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
359 implementation would not encrypt the entirety of the data under some
360 circumstances. This could reveal sixteen bytes of data that was
361 preexisting in the memory that wasn't written. In the special case of
362 "in place" encryption, sixteen bytes of the plaintext would be revealed.
364 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
365 they are both unaffected.
368 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
370 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
372 * In addition to the c_rehash shell command injection identified in
373 CVE-2022-1292, further bugs where the c_rehash script does not
374 properly sanitise shell metacharacters to prevent command injection have been
377 When the CVE-2022-1292 was fixed it was not discovered that there
378 are other places in the script where the file names of certificates
379 being hashed were possibly passed to a command executed through the shell.
381 This script is distributed by some operating systems in a manner where
382 it is automatically executed. On such operating systems, an attacker
383 could execute arbitrary commands with the privileges of the script.
385 Use of the c_rehash script is considered obsolete and should be replaced
386 by the OpenSSL rehash command line tool.
389 *Daniel Fiala, Tomáš Mráz*
391 * Case insensitive string comparison no longer uses locales. It has instead
392 been directly implemented.
396 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
398 * Case insensitive string comparison is reimplemented via new locale-agnostic
399 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
400 comparison. The previous implementation had problems when the Turkish locale
405 * Fixed a bug in the c_rehash script which was not properly sanitising shell
406 metacharacters to prevent command injection. This script is distributed by
407 some operating systems in a manner where it is automatically executed. On
408 such operating systems, an attacker could execute arbitrary commands with the
409 privileges of the script.
411 Use of the c_rehash script is considered obsolete and should be replaced
412 by the OpenSSL rehash command line tool.
417 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
418 certificate on an OCSP response. The bug caused the function in the case
419 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
420 response (meaning a successful verification) even in the case where the
421 response signing certificate fails to verify.
423 It is anticipated that most users of `OCSP_basic_verify` will not use the
424 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
425 a negative value (indicating a fatal error) in the case of a certificate
426 verification failure. The normal expected return value in this case would be
429 This issue also impacts the command line OpenSSL "ocsp" application. When
430 verifying an ocsp response with the "-no_cert_checks" option the command line
431 application will report that the verification is successful even though it
432 has in fact failed. In this case the incorrect successful response will also
433 be accompanied by error messages showing the failure and contradicting the
434 apparently successful result.
439 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
440 AAD data as the MAC key. This made the MAC key trivially predictable.
442 An attacker could exploit this issue by performing a man-in-the-middle attack
443 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
444 that the modified data would still pass the MAC integrity check.
446 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
447 endpoint will always be rejected by the recipient and the connection will
448 fail at that point. Many application protocols require data to be sent from
449 the client to the server first. Therefore, in such a case, only an OpenSSL
450 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
452 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
453 sent in both directions. In this case both clients and servers could be
454 affected, regardless of the application protocol.
456 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
457 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
458 the handshake when using this ciphersuite.
460 The confidentiality of data is not impacted by this issue, i.e. an attacker
461 cannot decrypt data that has been encrypted using this ciphersuite - they can
464 In order for this attack to work both endpoints must legitimately negotiate
465 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
466 OpenSSL 3.0, and is not available within the default provider or the default
467 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
468 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
469 following must have occurred:
471 1) OpenSSL must have been compiled with the (non-default) compile time option
472 enable-weak-ssl-ciphers
474 2) OpenSSL must have had the legacy provider explicitly loaded (either
475 through application code or via configuration)
477 3) The ciphersuite must have been explicitly added to the ciphersuite list
479 4) The libssl security level must have been set to 0 (default is 1)
481 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
483 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
484 others that both endpoints have in common
489 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
490 occupied by the removed hash table entries.
492 This function is used when decoding certificates or keys. If a long lived
493 process periodically decodes certificates or keys its memory usage will
494 expand without bounds and the process might be terminated by the operating
495 system causing a denial of service. Also traversing the empty hash table
496 entries will take increasingly more time.
498 Typically such long lived processes might be TLS clients or TLS servers
499 configured to accept client certificate authentication.
502 *Hugo Landau, Aliaksei Levin*
504 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
505 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
506 statistics are no longer supported. For compatibility, these statistics are
507 still listed in the output but are now always reported as zero.
511 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
513 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
514 for non-prime moduli.
516 Internally this function is used when parsing certificates that contain
517 elliptic curve public keys in compressed form or explicit elliptic curve
518 parameters with a base point encoded in compressed form.
520 It is possible to trigger the infinite loop by crafting a certificate that
521 has invalid explicit curve parameters.
523 Since certificate parsing happens prior to verification of the certificate
524 signature, any process that parses an externally supplied certificate may thus
525 be subject to a denial of service attack. The infinite loop can also be
526 reached when parsing crafted private keys as they can contain explicit
527 elliptic curve parameters.
529 Thus vulnerable situations include:
531 - TLS clients consuming server certificates
532 - TLS servers consuming client certificates
533 - Hosting providers taking certificates or private keys from customers
534 - Certificate authorities parsing certification requests from subscribers
535 - Anything else which parses ASN.1 elliptic curve parameters
537 Also any other applications that use the BN_mod_sqrt() where the attacker
538 can control the parameter values are vulnerable to this DoS issue.
543 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
544 to the list of ciphersuites providing Perfect Forward Secrecy as
545 required by SECLEVEL >= 3.
547 *Dmitry Belyavskiy, Nicola Tuveri*
549 * Made the AES constant time code for no-asm configurations
550 optional due to the resulting 95% performance degradation.
551 The AES constant time code can be enabled, for no assembly
552 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
556 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
561 * The negative return value handling of the certificate verification callback
562 was reverted. The replacement is to set the verification retry state with
563 the SSL_set_retry_verify() function.
567 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
569 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
570 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
571 verify a certificate supplied by a server. That function may return a
572 negative return value to indicate an internal error (for example out of
573 memory). Such a negative return value is mishandled by OpenSSL and will cause
574 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
575 success and a subsequent call to SSL_get_error() to return the value
576 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
577 returned by OpenSSL if the application has previously called
578 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
579 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
580 totally unexpected and applications may not behave correctly as a result. The
581 exact behaviour will depend on the application but it could result in
582 crashes, infinite loops or other similar incorrect responses.
584 This issue is made more serious in combination with a separate bug in OpenSSL
585 3.0 that will cause X509_verify_cert() to indicate an internal error when
586 processing a certificate chain. This will occur where a certificate does not
587 include the Subject Alternative Name extension but where a Certificate
588 Authority has enforced name constraints. This issue can occur even with valid
594 * Corrected a few file name and file reference bugs in the build,
595 installation and setup scripts, which lead to installation verification
596 failures. Slightly enhanced the installation verification script.
600 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
605 * Fixed PVK encoder to properly query for the passphrase.
609 * Multiple fixes in the OSSL_HTTP API functions.
613 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
614 OSSL_PARAM_INTEGER data type and return error on negative numbers
615 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
616 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
620 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
624 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
628 * Multiple threading fixes.
632 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
636 * Allow fetching an operation from the provider that owns an unexportable key
637 as a fallback if that is still allowed by the property query.
641 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
643 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
648 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
649 S390X capability vector to zero. This simplifies testing of different code
650 paths on S390X architecture.
654 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
655 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
656 SP 800-38D". The communication will fail at this point.
660 * The EC_GROUP_clear_free() function is deprecated as there is nothing
661 confidential in EC_GROUP data.
665 * The byte order mark (BOM) character is ignored if encountered at the
666 beginning of a PEM-formatted file.
670 * Added CMS support for the Russian GOST algorithms.
674 * Due to move of the implementation of cryptographic operations
675 to the providers, validation of various operation parameters can
676 be postponed until the actual operation is executed where previously
677 it happened immediately when an operation parameter was set.
679 For example when setting an unsupported curve with
680 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
681 fail but later keygen operations with the EVP_PKEY_CTX will fail.
683 *OpenSSL team members and many third party contributors*
685 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
686 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
687 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
688 instead to retrieve these algorithms from a provider.
692 * On build targets where the multilib postfix is set in the build
693 configuration the libdir directory was changing based on whether
694 the lib directory with the multilib postfix exists on the system
695 or not. This unpredictable behavior was removed and eventual
696 multilib postfix is now always added to the default libdir. Use
697 `--libdir=lib` to override the libdir if adding the postfix is
702 * The triple DES key wrap functionality now conforms to RFC 3217 but is
703 no longer interoperable with OpenSSL 1.1.1.
707 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
708 function codes, this function can only cause problems for calling
713 * Add a configurable flag to output date formats as ISO 8601. Does not
714 change the default date format.
718 * Version of MSVC earlier than 1300 could get link warnings, which could
719 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
720 Support for this flag has been removed.
724 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
725 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
726 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
727 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
728 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
732 * The signatures of the functions to get and set options on SSL and
733 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
734 Some source code changes may be required.
738 * The public definitions of conf_method_st and conf_st have been
739 deprecated. They will be made opaque in a future release.
741 *Rich Salz and Tomáš Mráz*
743 * Client-initiated renegotiation is disabled by default. To allow it, use
744 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
745 flag, or the "ClientRenegotiation" config parameter as appropriate.
749 * Add "abspath" and "includedir" pragma's to config files, to prevent,
750 or modify relative pathname inclusion.
754 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
755 validated. Please consult the README-FIPS and
756 README-PROVIDERS files, as well as the migration guide.
758 *OpenSSL team members and many third party contributors*
760 * For the key types DH and DHX the allowed settable parameters are now different.
764 * The openssl commands that read keys, certificates, and CRLs now
765 automatically detect the PEM or DER format of the input files.
767 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
769 * Added enhanced PKCS#12 APIs which accept a library context.
773 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
777 * Added support for Kernel TLS (KTLS).
779 *Boris Pismenny, John Baldwin and Andrew Gallatin*
781 * Support for RFC 5746 secure renegotiation is now required by default for
782 SSL or TLS connections to succeed.
786 * The signature of the `copy` functional parameter of the
787 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
788 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
789 the signature of the `pub_decode` functional parameter of the
790 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
791 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
795 * The error return values from some control calls (ctrl) have changed.
799 * A public key check is now performed during EVP_PKEY_derive_set_peer().
803 * Many functions in the EVP_ namespace that are getters of values from
804 implementations or contexts were renamed to include get or get0 in their
805 names. Old names are provided as macro aliases for compatibility and
810 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
811 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
812 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
817 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
820 * The output from the command line applications may have minor
825 * The output from numerous "printing" may have minor changes.
829 * Windows thread synchronization uses read/write primitives (SRWLock) when
830 supported by the OS, otherwise CriticalSection continues to be used.
834 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
835 work on read only BIO source/sinks that do not support these functions.
836 This allows piping or redirection of a file BIO using stdin to be buffered
837 into memory. This is used internally in OSSL_DECODER_from_bio().
841 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
842 this function would return one of the values OSSL_STORE_INFO_NAME,
843 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
844 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
845 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
846 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
847 using this function should be amended to handle the changed return value.
851 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
852 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
853 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
854 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
855 contains more than one certificate identifier: This means that all
856 certificates referenced there MUST be part of the validation chain.
860 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
861 RC5, DESX and DES have been moved to the legacy provider.
865 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
866 RIPEMD-160 have been moved to the legacy provider.
870 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
875 * The deprecated functions EVP_PKEY_get0_RSA(),
876 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
877 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
878 well as the similarly named "get1" functions behave differently in
883 * A number of functions handling low-level keys or engines were deprecated
884 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
885 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
886 EVP_PKEY_get0_siphash().
890 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
891 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
892 will need to load the legacy crypto provider. This includes these PBE
893 algorithms which use this KDF:
894 - NID_pbeWithMD2AndDES_CBC
895 - NID_pbeWithMD5AndDES_CBC
896 - NID_pbeWithSHA1AndRC2_CBC
897 - NID_pbeWithMD2AndRC2_CBC
898 - NID_pbeWithMD5AndRC2_CBC
899 - NID_pbeWithSHA1AndDES_CBC
903 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
904 BIO_debug_callback() functions.
908 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
909 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
913 * The RAND_METHOD APIs have been deprecated.
917 * The SRP APIs have been deprecated.
921 * Add a compile time option to prevent the caching of provider fetched
922 algorithms. This is enabled by including the no-cached-fetch option
923 at configuration time.
927 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
928 count of PKCS12_DEFAULT_ITER.
930 *Tomáš Mráz and Sahana Prasad*
932 * The openssl speed command does not use low-level API calls anymore.
936 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
939 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
941 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
945 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
946 providers may supply their own group implementations (using either the "key
947 exchange" or the "key encapsulation" methods) which will automatically be
948 detected and used by libssl.
950 *Matt Caswell, Nicola Tuveri*
952 * The undocumented function X509_certificate_type() has been deprecated;
956 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
960 * Removed RSA padding mode for SSLv23 (which was only used for
961 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
962 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
967 * Deprecated the obsolete X9.31 RSA key generation related functions.
969 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
970 is not allowed to return a value > 1, this is no more taken as failure.
972 *Viktor Dukhovni and David von Oheimb*
974 * Deprecated the obsolete X9.31 RSA key generation related functions
975 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
976 BN_X931_generate_prime_ex().
980 * The default key generation method for the regular 2-prime RSA keys was
981 changed to the FIPS 186-4 B.3.6 method.
985 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
989 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
993 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
994 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
996 *Rich Salz, Richard Levitte, and David von Oheimb*
998 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1002 * Deprecated `OCSP_parse_url()`.
1006 * Validation of SM2 keys has been separated from the validation of regular EC
1011 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1012 switches: a validation failure triggers an early exit, returning a failure
1013 exit status to the parent process.
1017 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1018 to ignore unknown ciphers.
1022 * The `-cipher-commands` and `-digest-commands` options
1023 of the command line utility `list` have been deprecated.
1024 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1028 * Added convenience functions for generating asymmetric key pairs:
1029 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1030 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1034 * All of the low-level EC_KEY functions have been deprecated.
1036 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1038 * Deprecated all the libcrypto and libssl error string loading
1043 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1044 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1049 * The `-crypt` option to the `passwd` command line tool has been removed.
1053 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1058 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1062 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1063 EVP_PKEY_get1_tls_encodedpoint().
1067 * The security callback, which can be customised by application code, supports
1068 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1069 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1073 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1074 interface. Their functionality remains unchanged.
1078 * Added new option for 'openssl list', '-providers', which will display the
1079 list of loaded providers, their names, version and status. It optionally
1080 displays their gettable parameters.
1084 * Removed EVP_PKEY_set_alias_type().
1088 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1089 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1093 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1094 parameters are still checked for type safety at compile time via helper
1099 * Remove the RAND_DRBG API
1101 *Paul Dale and Matthias St. Pierre*
1103 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1104 as well as actual hostnames.
1108 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1109 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1110 conversely, silently ignore DTLS protocol version bounds when configuring
1111 TLS-based contexts. The commands can be repeated to set bounds of both
1112 types. The same applies with the corresponding "min_protocol" and
1113 "max_protocol" command-line switches, in case some application uses both TLS
1116 SSL_CTX instances that are created for a fixed protocol version (e.g.
1117 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1118 attempts to apply bounds to these protocol versions would result in an
1119 error. Now only the "version-flexible" SSL_CTX instances are subject to
1120 limits in configuration files in command-line options.
1124 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1129 * Reworked the recorded ERR codes to make better space for system errors.
1130 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1131 given code is a system error (true) or an OpenSSL error (false).
1135 * Reworked the test perl framework to better allow parallel testing.
1137 *Nicola Tuveri and David von Oheimb*
1139 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1140 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1144 * 'Configure' has been changed to figure out the configuration target if
1145 none is given on the command line. Consequently, the 'config' script is
1146 now only a mere wrapper. All documentation is changed to only mention
1149 *Rich Salz and Richard Levitte*
1151 * Added a library context `OSSL_LIB_CTX` that applications as well as
1152 other libraries can use to form a separate context within which
1153 libcrypto operations are performed.
1157 * Added various `_ex` functions to the OpenSSL API that support using
1158 a non-default `OSSL_LIB_CTX`.
1162 * Handshake now fails if Extended Master Secret extension is dropped
1167 * Dropped interactive mode from the `openssl` program.
1171 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1173 *David von Oheimb and Shane Lontis*
1175 * Deprecated `EC_METHOD_get_field_type()`.
1179 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1180 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1181 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1185 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1189 * Add CAdES-BES signature verification support, mostly derived
1190 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1192 *Filipe Raimundo da Silva*
1194 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1198 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1199 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1203 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1207 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1208 EC_KEY_precompute_mult().
1212 * Deprecated EC_POINTs_mul().
1216 * Removed FIPS_mode() and FIPS_mode_set().
1220 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1224 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1225 EC_POINT_get_Jprojective_coordinates_GFp().
1229 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1230 arrays to be more easily constructed via a series of utility functions.
1231 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1232 the various push functions and finally convert to a passable OSSL_PARAM
1233 array using OSSL_PARAM_BLD_to_param().
1237 * The security strength of SHA1 and MD5 based signatures in TLS has been
1242 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1243 contain a provider side internal key.
1247 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1251 * Project text documents not yet having a proper file name extension
1252 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1253 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1257 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1258 have been converted to Markdown with the goal to produce documents
1259 which not only look pretty when viewed online in the browser, but
1260 remain well readable inside a plain text editor.
1262 To achieve this goal, a 'minimalistic' Markdown style has been applied
1263 which avoids formatting elements that interfere too much with the
1264 reading flow in the text file. For example, it
1266 * avoids [ATX headings][] and uses [setext headings][] instead
1267 (which works for `<h1>` and `<h2>` headings only).
1268 * avoids [inline links][] and uses [reference links][] instead.
1269 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1271 [ATX headings]: https://github.github.com/gfm/#atx-headings
1272 [setext headings]: https://github.github.com/gfm/#setext-headings
1273 [inline links]: https://github.github.com/gfm/#inline-link
1274 [reference links]: https://github.github.com/gfm/#reference-link
1275 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1276 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1278 *Matthias St. Pierre*
1280 * The test suite is changed to preserve results of each test recipe.
1281 A new directory test-runs/ with subdirectories named like the
1282 test recipes are created in the build tree for this purpose.
1286 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1287 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1288 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1290 *David von Oheimb, Martin Peylo*
1292 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1293 It supports arbitrary request and response content types, GET redirection,
1294 TLS, connections via HTTP(S) proxies, connections and exchange via
1295 user-defined BIOs (allowing implicit connections), persistent connections,
1296 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1297 The legacy OCSP-focused (and only partly documented) API
1298 is retained for backward compatibility, while most of it is deprecated.
1302 * Added `util/check-format.pl`, a tool for checking adherence to the
1303 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1304 The checks performed are incomplete and yield some false positives.
1305 Still the tool should be useful for detecting most typical glitches.
1309 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1310 If domain name resolution yields multiple IP addresses all of them are tried
1311 after `connect()` failures.
1315 * All of the low-level RSA functions have been deprecated.
1319 * X509 certificates signed using SHA1 are no longer allowed at security
1324 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1325 modified to use PKEY APIs. These commands are now in maintenance mode
1326 and no new features will be added to them.
1330 * The command line utility rsautl has been deprecated.
1334 * The command line utilities genrsa and rsa have been modified to use PKEY
1335 APIs. They now write PKCS#8 keys by default. These commands are now in
1336 maintenance mode and no new features will be added to them.
1340 * All of the low-level DH functions have been deprecated.
1342 *Paul Dale and Matt Caswell*
1344 * All of the low-level DSA functions have been deprecated.
1348 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1349 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1353 * Deprecated low-level ECDH and ECDSA functions.
1357 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1361 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1362 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1363 a new formulation to include all the things it can be used for,
1364 as well as words of caution.
1368 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1372 * All of the low-level HMAC functions have been deprecated.
1374 *Paul Dale and David von Oheimb*
1376 * Over two thousand fixes were made to the documentation, including:
1377 - Common options (such as -rand/-writerand, TLS version control, etc)
1378 were refactored and point to newly-enhanced descriptions in openssl.pod.
1379 - Added style conformance for all options (with help from Richard Levitte),
1380 documented all reported missing options, added a CI build to check
1381 that all options are documented and that no unimplemented options
1383 - Documented some internals, such as all use of environment variables.
1384 - Addressed all internal broken L<> references.
1388 * All of the low-level CMAC functions have been deprecated.
1392 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1393 functions have been deprecated.
1395 *Paul Dale and David von Oheimb*
1397 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1398 set of functions. The documentation mentioned negative values for some
1399 errors, but this was never the case, so the mention of negative values
1402 Code that followed the documentation and thereby check with something
1403 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1407 * All of the low-level cipher functions have been deprecated.
1409 *Matt Caswell and Paul Dale*
1411 * Removed include/openssl/opensslconf.h.in and replaced it with
1412 include/openssl/configuration.h.in, which differs in not including
1413 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1414 was added to include both.
1416 This allows internal hacks where one might need to modify the set
1417 of configured macros, for example this if deprecated symbols are
1418 still supposed to be available internally:
1420 #include <openssl/configuration.h>
1422 #undef OPENSSL_NO_DEPRECATED
1423 #define OPENSSL_SUPPRESS_DEPRECATED
1425 #include <openssl/macros.h>
1427 This should not be used by applications that use the exported
1428 symbols, as that will lead to linking errors.
1432 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1433 used in exponentiation with 512-bit moduli. No EC algorithms are
1434 affected. Analysis suggests that attacks against 2-prime RSA1024,
1435 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1436 difficult to perform and are not believed likely. Attacks against DH512
1437 are considered just feasible. However, for an attack the target would
1438 have to re-use the DH512 private key, which is not recommended anyway.
1439 Also applications directly using the low-level API BN_mod_exp may be
1440 affected if they use BN_FLG_CONSTTIME.
1445 * Most memory-debug features have been deprecated, and the functionality
1446 replaced with no-ops.
1450 * Added documentation for the STACK API.
1454 * Introduced a new method type and API, OSSL_ENCODER, to represent
1455 generic encoders. These do the same sort of job that PEM writers
1456 and d2i functions do, but with support for methods supplied by
1457 providers, and the possibility for providers to support other
1462 * Introduced a new method type and API, OSSL_DECODER, to represent
1463 generic decoders. These do the same sort of job that PEM readers
1464 and i2d functions do, but with support for methods supplied by
1465 providers, and the possibility for providers to support other
1470 * Added a .pragma directive to the syntax of configuration files, to
1471 allow varying behavior in a supported and predictable manner.
1472 Currently added pragma:
1476 This allows dollar signs to be a keyword character unless it's
1477 followed by a opening brace or parenthesis. This is useful for
1478 platforms where dollar signs are commonly used in names, such as
1479 volume names and system directory names on VMS.
1483 * Added functionality to create an EVP_PKEY from user data.
1487 * Change the interpretation of the '--api' configuration option to
1488 mean that this is a desired API compatibility level with no
1489 further meaning. The previous interpretation, that this would
1490 also mean to remove all deprecated symbols up to and including
1491 the given version, no requires that 'no-deprecated' is also used
1492 in the configuration.
1494 When building applications, the desired API compatibility level
1495 can be set with the OPENSSL_API_COMPAT macro like before. For
1496 API compatibility version below 3.0, the old style numerical
1497 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1498 For version 3.0 and on, the value is expected to be the decimal
1499 value calculated from the major and minor version like this:
1501 MAJOR * 10000 + MINOR * 100
1505 -DOPENSSL_API_COMPAT=30000 For 3.0
1506 -DOPENSSL_API_COMPAT=30200 For 3.2
1508 To hide declarations that are deprecated up to and including the
1509 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1510 given when building the application as well.
1514 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1515 access to certificate and CRL stores via URIs and OSSL_STORE
1518 This adds the following functions:
1520 - X509_LOOKUP_store()
1521 - X509_STORE_load_file()
1522 - X509_STORE_load_path()
1523 - X509_STORE_load_store()
1524 - SSL_add_store_cert_subjects_to_stack()
1525 - SSL_CTX_set_default_verify_store()
1526 - SSL_CTX_load_verify_file()
1527 - SSL_CTX_load_verify_dir()
1528 - SSL_CTX_load_verify_store()
1532 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1533 The presence of this system service is determined at run-time.
1537 * Added functionality to create an EVP_PKEY context based on data
1538 for methods from providers. This takes an algorithm name and a
1539 property query string and simply stores them, with the intent
1540 that any operation that uses this context will use those strings
1541 to fetch the needed methods implicitly, thereby making the port
1542 of application written for pre-3.0 OpenSSL easier.
1546 * The undocumented function NCONF_WIN32() has been deprecated; for
1547 conversion details see the HISTORY section of doc/man5/config.pod
1551 * Introduced the new functions EVP_DigestSignInit_ex() and
1552 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1553 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1554 pages for further details.
1558 * Over two thousand fixes were made to the documentation, including:
1559 adding missing command flags, better style conformance, documentation
1562 *Rich Salz, Richard Levitte*
1564 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1565 X25519, X448, Ed25519 and Ed448.
1569 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1574 * Deprecated the public definition of `ERR_STATE` as well as the function
1575 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1580 * Added ERR functionality to give callers access to the stored function
1581 names that have replaced the older function code based functions.
1583 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1584 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1585 ERR_peek_error_all() and ERR_peek_last_error_all().
1587 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1588 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1589 ERR_func_error_string().
1593 * Extended testing to be verbose for failing tests only. The make variables
1594 VERBOSE_FAILURE or VF can be used to enable this:
1596 $ make VF=1 test # Unix
1597 $ mms /macro=(VF=1) test ! OpenVMS
1598 $ nmake VF=1 test # Windows
1602 * Added the `-copy_extensions` option to the `x509` command for use with
1603 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1604 all extensions in the request are copied to the certificate or vice versa.
1606 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1608 * Added the `-copy_extensions` option to the `req` command for use with
1609 `-x509`. When given with the `copy` or `copyall` argument,
1610 all extensions in the certification request are copied to the certificate.
1614 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1615 they generate are by default RFC 5280 compliant in the following sense:
1616 There is a subjectKeyIdentifier extension with a hash value of the public key
1617 and for not self-signed certs there is an authorityKeyIdentifier extension
1618 with a keyIdentifier field or issuer information identifying the signing key.
1619 This is done unless some configuration overrides the new default behavior,
1620 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1624 * Added several checks to `X509_verify_cert()` according to requirements in
1625 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1626 (which may be done by using the CLI option `-x509_strict`):
1627 * The basicConstraints of CA certificates must be marked critical.
1628 * CA certificates must explicitly include the keyUsage extension.
1629 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1630 * The issuer name of any certificate must not be empty.
1631 * The subject name of CA certs, certs with keyUsage crlSign,
1632 and certs without subjectAlternativeName must not be empty.
1633 * If a subjectAlternativeName extension is given it must not be empty.
1634 * The signatureAlgorithm field and the cert signature must be consistent.
1635 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1636 must not be marked critical.
1637 * The authorityKeyIdentifier must be given for X.509v3 certs
1638 unless they are self-signed.
1639 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1643 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1644 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1648 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1649 used even when parsing explicit parameters, when loading a encoded key
1650 or calling `EC_GROUP_new_from_ecpkparameters()`/
1651 `EC_GROUP_new_from_ecparameters()`.
1652 This prevents bypass of security hardening and performance gains,
1653 especially for curves with specialized EC_METHODs.
1654 By default, if a key encoded with explicit parameters is loaded and later
1655 encoded, the output is still encoded with explicit parameters, even if
1656 internally a "named" EC_GROUP is used for computation.
1660 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1661 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1662 NULL. After this change, only the cofactor parameter can be NULL. It also
1663 does some minimal sanity checks on the passed order.
1668 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1669 An attack is simple, if the first CMS_recipientInfo is valid but the
1670 second CMS_recipientInfo is chosen ciphertext. If the second
1671 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1672 encryption key will be replaced by garbage, and the message cannot be
1673 decoded, but if the RSA decryption fails, the correct encryption key is
1674 used and the recipient will not notice the attack.
1675 As a work around for this potential attack the length of the decrypted
1676 key must be equal to the cipher default key length, in case the
1677 certificate is not given and all recipientInfo are tried out.
1678 The old behaviour can be re-enabled in the CMS code by setting the
1679 CMS_DEBUG_DECRYPT flag.
1683 * Early start up entropy quality from the DEVRANDOM seed source has been
1684 improved for older Linux systems. The RAND subsystem will wait for
1685 /dev/random to be producing output before seeding from /dev/urandom.
1686 The seeded state is stored for future library initialisations using
1687 a system global shared memory segment. The shared memory identifier
1688 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1689 the desired value. The default identifier is 114.
1693 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1694 when primes for RSA keys are computed.
1695 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1696 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1697 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1698 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1699 This avoids possible fingerprinting of newly generated RSA modules.
1703 * Correct the extended master secret constant on EBCDIC systems. Without this
1704 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1705 negotiate EMS will fail. Unfortunately this also means that TLS connections
1706 between EBCDIC systems with this fix, and EBCDIC systems without this
1707 fix will fail if they negotiate EMS.
1711 * Changed the library initialisation so that the config file is now loaded
1712 by default. This was already the case for libssl. It now occurs for both
1713 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1714 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1718 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1719 where the former acts as a replacement for `ERR_put_error()`, and the
1720 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1721 `ERR_raise_data()` adds more flexibility by taking a format string and
1722 an arbitrary number of arguments following it, to be processed with
1727 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1728 to check if a named provider is loaded and available. When called, it
1729 will also activate all fallback providers if such are still present.
1733 * Enforce a minimum DH modulus size of 512 bits.
1737 * Changed DH parameters to generate the order q subgroup instead of 2q.
1738 Previously generated DH parameters are still accepted by DH_check
1739 but DH_generate_key works around that by clearing bit 0 of the
1740 private key for those. This avoids leaking bit 0 of the private key.
1744 * Significantly reduce secure memory usage by the randomness pools.
1748 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1753 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1754 algorithms. An implementation of a key exchange algorithm can be obtained
1755 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1756 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1757 the older EVP_PKEY_derive_init() function. See the man pages for the new
1758 functions for further details.
1762 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1766 * Removed the function names from error messages and deprecated the
1771 * Removed NextStep support and the macro OPENSSL_UNISTD
1775 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1776 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1777 Also removed "export var as function" capability; we do not export
1778 variables, only functions.
1782 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1783 an error and 1 indicating success. In previous versions of OpenSSL this
1784 was a void type. If a key was set longer than the maximum possible this
1789 * Support SM2 signing and verification schemes with X509 certificate.
1793 * Use SHA256 as the default digest for TS query in the `ts` app.
1797 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1801 * Default cipher lists/suites are now available via a function, the
1802 #defines are deprecated.
1806 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1807 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1808 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1812 * Join the directories crypto/x509 and crypto/x509v3
1816 * Added command 'openssl kdf' that uses the EVP_KDF API.
1820 * Added command 'openssl mac' that uses the EVP_MAC API.
1824 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1825 as default directories. Also added the command 'openssl info'
1826 for scripting purposes.
1830 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1835 * Add prediction resistance to the DRBG reseeding process.
1839 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1840 mandated by IEEE Std 1619-2018.
1844 * Added newline escaping functionality to a filename when using openssl dgst.
1845 This output format is to replicate the output format found in the `*sum`
1846 checksum programs. This aims to preserve backward compatibility.
1848 *Matt Eaton, Richard Levitte, and Paul Dale*
1850 * Removed the heartbeat message in DTLS feature, as it has very
1851 little usage and doesn't seem to fulfill a valuable purpose.
1852 The configuration option is now deprecated.
1856 * Changed the output of 'openssl {digestname} < file' to display the
1857 digest name in its output.
1861 * Added a new generic trace API which provides support for enabling
1862 instrumentation through trace output.
1864 *Richard Levitte & Matthias St. Pierre*
1866 * Added build tests for C++. These are generated files that only do one
1867 thing, to include one public OpenSSL head file each. This tests that
1868 the public header files can be usefully included in a C++ application.
1870 This test isn't enabled by default. It can be enabled with the option
1871 'enable-buildtest-c++'.
1875 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1879 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1883 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1887 * Added KMAC to EVP_MAC.
1891 * Added property based algorithm implementation selection framework to
1896 * Added SCA hardening for modular field inversion in EC_GROUP through
1897 a new dedicated field_inv() pointer in EC_METHOD.
1898 This also addresses a leakage affecting conversions from projective
1899 to affine coordinates.
1901 *Billy Bob Brumley, Nicola Tuveri*
1903 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1904 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1905 those algorithms that were already supported through the EVP_PKEY API
1906 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1907 and scrypt are now wrappers that call EVP_KDF.
1911 * Build devcrypto engine as a dynamic engine.
1913 *Eneas U de Queiroz*
1915 * Add keyed BLAKE2 to EVP_MAC.
1919 * Fix a bug in the computation of the endpoint-pair shared secret used
1920 by DTLS over SCTP. This breaks interoperability with older versions
1921 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1922 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1923 interoperability with such broken implementations. However, enabling
1924 this switch breaks interoperability with correct implementations.
1926 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1927 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1931 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1935 * Changed the license to the Apache License v2.0.
1939 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1941 - Major releases (indicated by incrementing the MAJOR release number)
1942 may introduce incompatible API/ABI changes.
1943 - Minor releases (indicated by incrementing the MINOR release number)
1944 may introduce new features but retain API/ABI compatibility.
1945 - Patch releases (indicated by incrementing the PATCH number)
1946 are intended for bug fixes and other improvements of existing
1947 features only (like improving performance or adding documentation)
1948 and retain API/ABI compatibility.
1952 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1956 * Remove the 'dist' target and add a tarball building script. The
1957 'dist' target has fallen out of use, and it shouldn't be
1958 necessary to configure just to create a source distribution.
1962 * Recreate the OS390-Unix config target. It no longer relies on a
1963 special script like it did for OpenSSL pre-1.1.0.
1967 * Instead of having the source directories listed in Configure, add
1968 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1973 * Add GMAC to EVP_MAC.
1977 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1981 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1982 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1983 to facilitate the continued use of MACs through raw private keys in
1984 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1988 * Deprecate ECDH_KDF_X9_62().
1992 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1993 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1994 are retained for backwards compatibility.
1998 * AES-XTS mode now enforces that its two keys are different to mitigate
1999 the attacked described in "Efficient Instantiations of Tweakable
2000 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2001 Details of this attack can be obtained from:
2002 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2006 * Rename the object files, i.e. give them other names than in previous
2007 versions. Their names now include the name of the final product, as
2008 well as its type mnemonic (bin, lib, shlib).
2012 * Added new option for 'openssl list', '-objects', which will display the
2013 list of built in objects, i.e. OIDs with names.
2017 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2018 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2023 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2024 improves application performance by removing data copies and providing
2025 applications with zero-copy system calls such as sendfile and splice.
2029 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2033 * `PKCS12_parse` now maintains the order of the parsed certificates
2034 when outputting them via `*ca` (rather than reversing it).
2038 * Deprecated pthread fork support methods.
2042 * Added support for FFDHE key exchange in TLS 1.3.
2046 * Added a new concept for OpenSSL plugability: providers. This
2047 functionality is designed to replace the ENGINE API and ENGINE
2048 implementations, and to be much more dynamic, allowing provider
2049 authors to introduce new algorithms among other things, as long as
2050 there's an API that supports the algorithm type.
2052 With this concept comes a new core API for interaction between
2053 libcrypto and provider implementations. Public libcrypto functions
2054 that want to use providers do so through this core API.
2056 The main documentation for this core API is found in
2057 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2058 refer to other manuals describing the API specific for supported
2059 algorithm types (also called operations).
2066 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2068 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2070 * Avoid loading of a dynamic engine twice.
2074 * Prioritise DANE TLSA issuer certs over peer certs
2078 * Fixed random API for MacOS prior to 10.12
2080 These MacOS versions don't support the CommonCrypto APIs
2084 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2086 * Fixed an SM2 Decryption Buffer Overflow.
2088 In order to decrypt SM2 encrypted data an application is expected to
2089 call the API function EVP_PKEY_decrypt(). Typically an application will
2090 call this function twice. The first time, on entry, the "out" parameter
2091 can be NULL and, on exit, the "outlen" parameter is populated with the
2092 buffer size required to hold the decrypted plaintext. The application
2093 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2094 again, but this time passing a non-NULL value for the "out" parameter.
2096 A bug in the implementation of the SM2 decryption code means that the
2097 calculation of the buffer size required to hold the plaintext returned
2098 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2099 size required by the second call. This can lead to a buffer overflow
2100 when EVP_PKEY_decrypt() is called by the application a second time with
2101 a buffer that is too small.
2103 A malicious attacker who is able present SM2 content for decryption to
2104 an application could cause attacker chosen data to overflow the buffer
2105 by up to a maximum of 62 bytes altering the contents of other data held
2106 after the buffer, possibly changing application behaviour or causing
2107 the application to crash. The location of the buffer is application
2108 dependent but is typically heap allocated.
2113 * Fixed various read buffer overruns processing ASN.1 strings
2115 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2116 structure which contains a buffer holding the string data and a field
2117 holding the buffer length. This contrasts with normal C strings which
2118 are represented as a buffer for the string data which is terminated
2119 with a NUL (0) byte.
2121 Although not a strict requirement, ASN.1 strings that are parsed using
2122 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2123 well as any string whose value has been set with the ASN1_STRING_set()
2124 function will additionally NUL terminate the byte array in the
2125 ASN1_STRING structure.
2127 However, it is possible for applications to directly construct valid
2128 ASN1_STRING structures which do not NUL terminate the byte array by
2129 directly setting the "data" and "length" fields in the ASN1_STRING
2130 array. This can also happen by using the ASN1_STRING_set0() function.
2132 Numerous OpenSSL functions that print ASN.1 data have been found to
2133 assume that the ASN1_STRING byte array will be NUL terminated, even
2134 though this is not guaranteed for strings that have been directly
2135 constructed. Where an application requests an ASN.1 structure to be
2136 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2137 been directly constructed by the application without NUL terminating
2138 the "data" field, then a read buffer overrun can occur.
2140 The same thing can also occur during name constraints processing
2141 of certificates (for example if a certificate has been directly
2142 constructed by the application instead of loading it via the OpenSSL
2143 parsing functions, and the certificate contains non NUL terminated
2144 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2145 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2147 If a malicious actor can cause an application to directly construct an
2148 ASN1_STRING and then process it through one of the affected OpenSSL
2149 functions then this issue could be hit. This might result in a crash
2150 (causing a Denial of Service attack). It could also result in the
2151 disclosure of private memory contents (such as private keys, or
2152 sensitive plaintext).
2157 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2159 * Fixed a problem with verifying a certificate chain when using the
2160 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2161 the certificates present in a certificate chain. It is not set by default.
2163 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2164 the chain that have explicitly encoded elliptic curve parameters was added
2165 as an additional strict check.
2167 An error in the implementation of this check meant that the result of a
2168 previous check to confirm that certificates in the chain are valid CA
2169 certificates was overwritten. This effectively bypasses the check
2170 that non-CA certificates must not be able to issue other certificates.
2172 If a "purpose" has been configured then there is a subsequent opportunity
2173 for checks that the certificate is a valid CA. All of the named "purpose"
2174 values implemented in libcrypto perform this check. Therefore, where
2175 a purpose is set the certificate chain will still be rejected even when the
2176 strict flag has been used. A purpose is set by default in libssl client and
2177 server certificate verification routines, but it can be overridden or
2178 removed by an application.
2180 In order to be affected, an application must explicitly set the
2181 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2182 for the certificate verification or, in the case of TLS client or server
2183 applications, override the default purpose.
2188 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2189 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2190 renegotiation ClientHello omits the signature_algorithms extension (where it
2191 was present in the initial ClientHello), but includes a
2192 signature_algorithms_cert extension then a NULL pointer dereference will
2193 result, leading to a crash and a denial of service attack.
2195 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2196 (which is the default configuration). OpenSSL TLS clients are not impacted by
2200 *Peter Kästle and Samuel Sapalski*
2202 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2204 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2205 create a unique hash value based on the issuer and serial number data
2206 contained within an X509 certificate. However, it was failing to correctly
2207 handle any errors that may occur while parsing the issuer field (which might
2208 occur if the issuer field is maliciously constructed). This may subsequently
2209 result in a NULL pointer deref and a crash leading to a potential denial of
2215 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2216 padding mode to correctly check for rollback attacks. This is considered a
2217 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2222 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2223 functions. Previously they could overflow the output length argument in some
2224 cases where the input length is close to the maximum permissible length for
2225 an integer on the platform. In such cases the return value from the function
2226 call would be 1 (indicating success), but the output length value would be
2227 negative. This could cause applications to behave incorrectly or crash.
2232 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2233 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2234 could be exploited in a side channel attack to recover the password. Since
2235 the attack is local host only this is outside of the current OpenSSL
2236 threat model and therefore no CVE is assigned.
2238 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2243 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2245 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2246 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2247 If an attacker can control both items being compared then this could lead
2248 to a possible denial of service attack. OpenSSL itself uses the
2249 GENERAL_NAME_cmp function for two purposes:
2250 1) Comparing CRL distribution point names between an available CRL and a
2251 CRL distribution point embedded in an X509 certificate
2252 2) When verifying that a timestamp response token signer matches the
2253 timestamp authority name (exposed via the API functions
2254 TS_RESP_verify_response and TS_RESP_verify_token)
2259 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2261 * Certificates with explicit curve parameters are now disallowed in
2262 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2266 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2267 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2268 conversely, silently ignore DTLS protocol version bounds when configuring
2269 TLS-based contexts. The commands can be repeated to set bounds of both
2270 types. The same applies with the corresponding "min_protocol" and
2271 "max_protocol" command-line switches, in case some application uses both TLS
2274 SSL_CTX instances that are created for a fixed protocol version (e.g.
2275 TLSv1_server_method()) also silently ignore version bounds. Previously
2276 attempts to apply bounds to these protocol versions would result in an
2277 error. Now only the "version-flexible" SSL_CTX instances are subject to
2278 limits in configuration files in command-line options.
2282 * Handshake now fails if Extended Master Secret extension is dropped
2287 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2289 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2291 * Fixed segmentation fault in SSL_check_chain()
2292 Server or client applications that call the SSL_check_chain() function
2293 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2294 dereference as a result of incorrect handling of the
2295 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2296 or unrecognised signature algorithm is received from the peer. This could
2297 be exploited by a malicious peer in a Denial of Service attack.
2302 * Added AES consttime code for no-asm configurations
2303 an optional constant time support for AES was added
2304 when building openssl for no-asm.
2305 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2306 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2307 At this time this feature is by default disabled.
2308 It will be enabled by default in 3.0.
2312 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2314 * Revert the change of EOF detection while reading in libssl to avoid
2315 regressions in applications depending on the current way of reporting
2316 the EOF. As the existing method is not fully accurate the change to
2317 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2318 branch and will be present in the 3.0 release.
2322 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2323 when primes for RSA keys are computed.
2324 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2325 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2326 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2327 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2328 This avoids possible fingerprinting of newly generated RSA modules.
2332 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2334 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2335 while reading in libssl then we would report an error back to the
2336 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2337 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2338 therefore give a hint as to what went wrong.
2342 * Check that ed25519 and ed448 are allowed by the security level. Previously
2343 signature algorithms not using an MD were not being checked that they were
2344 allowed by the security level.
2348 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2349 was not quite right. The behaviour was not consistent between resumption
2350 and normal handshakes, and also not quite consistent with historical
2351 behaviour. The behaviour in various scenarios has been clarified and
2352 it has been updated to make it match historical behaviour as closely as
2357 * *[VMS only]* The header files that the VMS compilers include automatically,
2358 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2359 that the C++ compiler doesn't understand. This is a shortcoming in the
2360 compiler, but can be worked around with `__cplusplus` guards.
2362 C++ applications that use OpenSSL libraries must be compiled using the
2363 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2364 functions. Otherwise, only functions with symbols of less than 31
2365 characters can be used, as the linker will not be able to successfully
2366 resolve symbols with longer names.
2370 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2371 The presence of this system service is determined at run-time.
2375 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2380 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2382 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2383 number generator (RNG). This was intended to include protection in the
2384 event of a fork() system call in order to ensure that the parent and child
2385 processes did not share the same RNG state. However, this protection was not
2386 being used in the default case.
2388 A partial mitigation for this issue is that the output from a high
2389 precision timer is mixed into the RNG state so the likelihood of a parent
2390 and child process sharing state is significantly reduced.
2392 If an application already calls OPENSSL_init_crypto() explicitly using
2393 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2396 *Matthias St. Pierre*
2398 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2399 used even when parsing explicit parameters, when loading a encoded key
2400 or calling `EC_GROUP_new_from_ecpkparameters()`/
2401 `EC_GROUP_new_from_ecparameters()`.
2402 This prevents bypass of security hardening and performance gains,
2403 especially for curves with specialized EC_METHODs.
2404 By default, if a key encoded with explicit parameters is loaded and later
2405 encoded, the output is still encoded with explicit parameters, even if
2406 internally a "named" EC_GROUP is used for computation.
2410 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2411 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2412 NULL. After this change, only the cofactor parameter can be NULL. It also
2413 does some minimal sanity checks on the passed order.
2418 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2419 An attack is simple, if the first CMS_recipientInfo is valid but the
2420 second CMS_recipientInfo is chosen ciphertext. If the second
2421 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2422 encryption key will be replaced by garbage, and the message cannot be
2423 decoded, but if the RSA decryption fails, the correct encryption key is
2424 used and the recipient will not notice the attack.
2425 As a work around for this potential attack the length of the decrypted
2426 key must be equal to the cipher default key length, in case the
2427 certificate is not given and all recipientInfo are tried out.
2428 The old behaviour can be re-enabled in the CMS code by setting the
2429 CMS_DEBUG_DECRYPT flag.
2434 * Early start up entropy quality from the DEVRANDOM seed source has been
2435 improved for older Linux systems. The RAND subsystem will wait for
2436 /dev/random to be producing output before seeding from /dev/urandom.
2437 The seeded state is stored for future library initialisations using
2438 a system global shared memory segment. The shared memory identifier
2439 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2440 the desired value. The default identifier is 114.
2444 * Correct the extended master secret constant on EBCDIC systems. Without this
2445 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2446 negotiate EMS will fail. Unfortunately this also means that TLS connections
2447 between EBCDIC systems with this fix, and EBCDIC systems without this
2448 fix will fail if they negotiate EMS.
2452 * Use Windows installation paths in the mingw builds
2454 Mingw isn't a POSIX environment per se, which means that Windows
2455 paths should be used for installation.
2460 * Changed DH_check to accept parameters with order q and 2q subgroups.
2461 With order 2q subgroups the bit 0 of the private key is not secret
2462 but DH_generate_key works around that by clearing bit 0 of the
2463 private key for those. This avoids leaking bit 0 of the private key.
2467 * Significantly reduce secure memory usage by the randomness pools.
2471 * Revert the DEVRANDOM_WAIT feature for Linux systems
2473 The DEVRANDOM_WAIT feature added a select() call to wait for the
2474 /dev/random device to become readable before reading from the
2475 /dev/urandom device.
2477 It turned out that this change had negative side effects on
2478 performance which were not acceptable. After some discussion it
2479 was decided to revert this feature and leave it up to the OS
2480 resp. the platform maintainer to ensure a proper initialization
2481 during early boot time.
2483 *Matthias St. Pierre*
2485 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2487 * Add build tests for C++. These are generated files that only do one
2488 thing, to include one public OpenSSL head file each. This tests that
2489 the public header files can be usefully included in a C++ application.
2491 This test isn't enabled by default. It can be enabled with the option
2492 'enable-buildtest-c++'.
2496 * Enable SHA3 pre-hashing for ECDSA and DSA.
2500 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2501 This changes the size when using the `genpkey` command when no size is given.
2502 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2503 generation commands to use 2048 bits by default.
2507 * Reorganize the manual pages to consistently have RETURN VALUES,
2508 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2509 util/fix-doc-nits accordingly.
2511 *Paul Yang, Joshua Lock*
2513 * Add the missing accessor EVP_PKEY_get0_engine()
2517 * Have commands like `s_client` and `s_server` output the signature scheme
2518 along with other cipher suite parameters when debugging.
2522 * Make OPENSSL_config() error agnostic again.
2526 * Do the error handling in RSA decryption constant time.
2530 * Prevent over long nonces in ChaCha20-Poly1305.
2532 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2533 for every encryption operation. RFC 7539 specifies that the nonce value
2534 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2535 and front pads the nonce with 0 bytes if it is less than 12
2536 bytes. However it also incorrectly allows a nonce to be set of up to 16
2537 bytes. In this case only the last 12 bytes are significant and any
2538 additional leading bytes are ignored.
2540 It is a requirement of using this cipher that nonce values are
2541 unique. Messages encrypted using a reused nonce value are susceptible to
2542 serious confidentiality and integrity attacks. If an application changes
2543 the default nonce length to be longer than 12 bytes and then makes a
2544 change to the leading bytes of the nonce expecting the new value to be a
2545 new unique nonce then such an application could inadvertently encrypt
2546 messages with a reused nonce.
2548 Additionally the ignored bytes in a long nonce are not covered by the
2549 integrity guarantee of this cipher. Any application that relies on the
2550 integrity of these ignored leading bytes of a long nonce may be further
2551 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2552 is safe because no such use sets such a long nonce value. However user
2553 applications that use this cipher directly and set a non-default nonce
2554 length to be longer than 12 bytes may be vulnerable.
2556 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2562 * Add DEVRANDOM_WAIT feature for Linux systems
2564 On older Linux systems where the getrandom() system call is not available,
2565 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2566 Contrary to getrandom(), the /dev/urandom device will not block during
2567 early boot when the kernel CSPRNG has not been seeded yet.
2569 To mitigate this known weakness, use select() to wait for /dev/random to
2570 become readable before reading from /dev/urandom.
2572 * Ensure that SM2 only uses SM3 as digest algorithm
2576 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2578 * Change the info callback signals for the start and end of a post-handshake
2579 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2580 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2581 confused by this and assume that a TLSv1.2 renegotiation has started. This
2582 can break KeyUpdate handling. Instead we no longer signal the start and end
2583 of a post handshake message exchange (although the messages themselves are
2584 still signalled). This could break some applications that were expecting
2585 the old signals. However without this KeyUpdate is not usable for many
2590 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2592 * Timing vulnerability in DSA signature generation
2594 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2595 timing side channel attack. An attacker could use variations in the signing
2596 algorithm to recover the private key.
2598 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2603 * Timing vulnerability in ECDSA signature generation
2605 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2606 timing side channel attack. An attacker could use variations in the signing
2607 algorithm to recover the private key.
2609 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2614 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2615 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2616 of two gigabytes and the error handling improved.
2618 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2619 categorized as a normal bug, not a security issue, because the DRBG reseeds
2620 automatically and is fully functional even without additional randomness
2621 provided by the application.
2623 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2625 * Add a new ClientHello callback. Provides a callback interface that gives
2626 the application the ability to adjust the nascent SSL object at the
2627 earliest stage of ClientHello processing, immediately after extensions have
2628 been collected but before they have been processed. In particular, this
2629 callback can adjust the supported TLS versions in response to the contents
2634 * Add SM2 base algorithm support.
2638 * s390x assembly pack: add (improved) hardware-support for the following
2639 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2640 aes-cfb/cfb8, aes-ecb.
2644 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2645 parameter is no longer accepted, as it leads to a corrupt table. NULL
2646 pem_str is reserved for alias entries only.
2650 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2651 step for prime curves. The new implementation is based on formulae from
2652 differential addition-and-doubling in homogeneous projective coordinates
2653 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2654 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2655 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2656 to work in projective coordinates.
2658 *Billy Bob Brumley, Nicola Tuveri*
2660 * Change generating and checking of primes so that the error rate of not
2661 being prime depends on the intended use based on the size of the input.
2662 For larger primes this will result in more rounds of Miller-Rabin.
2663 The maximal error rate for primes with more than 1080 bits is lowered
2666 *Kurt Roeckx, Annie Yousar*
2668 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2672 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2673 moving between systems, and to avoid confusion when a Windows build is
2674 done with mingw vs with MSVC. For POSIX installs, there's still a
2675 symlink or copy named 'tsget' to avoid that confusion as well.
2679 * Revert blinding in ECDSA sign and instead make problematic addition
2680 length-invariant. Switch even to fixed-length Montgomery multiplication.
2684 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2685 step for binary curves. The new implementation is based on formulae from
2686 differential addition-and-doubling in mixed Lopez-Dahab projective
2687 coordinates, modified to independently blind the operands.
2689 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2691 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2692 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2693 EC_METHODs to implement their own specialized "ladder step", to take
2694 advantage of more favorable coordinate systems or more efficient
2695 differential addition-and-doubling algorithms.
2697 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2699 * Modified the random device based seed sources to keep the relevant
2700 file descriptors open rather than reopening them on each access.
2701 This allows such sources to operate in a chroot() jail without
2702 the associated device nodes being available. This behaviour can be
2703 controlled using RAND_keep_random_devices_open().
2707 * Numerous side-channel attack mitigations have been applied. This may have
2708 performance impacts for some algorithms for the benefit of improved
2709 security. Specific changes are noted in this change log by their respective
2714 * AIX shared library support overhaul. Switch to AIX "natural" way of
2715 handling shared libraries, which means collecting shared objects of
2716 different versions and bitnesses in one common archive. This allows to
2717 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2718 doesn't affect the way 3rd party applications are linked, only how
2719 multi-version installation is managed.
2723 * Make ec_group_do_inverse_ord() more robust and available to other
2724 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2725 mitigations are applied to the fallback BN_mod_inverse().
2726 When using this function rather than BN_mod_inverse() directly, new
2727 EC cryptosystem implementations are then safer-by-default.
2731 * Add coordinate blinding for EC_POINT and implement projective
2732 coordinate blinding for generic prime curves as a countermeasure to
2733 chosen point SCA attacks.
2735 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2737 * Add blinding to ECDSA and DSA signatures to protect against side channel
2738 attacks discovered by Keegan Ryan (NCC Group).
2742 * Enforce checking in the `pkeyutl` command to ensure that the input
2743 length does not exceed the maximum supported digest length when performing
2744 a sign, verify or verifyrecover operation.
2748 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2749 I/O in combination with something like select() or poll() will hang. This
2750 can be turned off again using SSL_CTX_clear_mode().
2751 Many applications do not properly handle non-application data records, and
2752 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2753 around the problems in those applications, but can also break some.
2754 It's recommended to read the manpages about SSL_read(), SSL_write(),
2755 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2756 SSL_CTX_set_read_ahead() again.
2760 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2761 now allow empty (zero character) pass phrases.
2765 * Apply blinding to binary field modular inversion and remove patent
2766 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2770 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2771 binary and prime elliptic curves.
2775 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2776 constant time fixed point multiplication.
2780 * Revise elliptic curve scalar multiplication with timing attack
2781 defenses: ec_wNAF_mul redirects to a constant time implementation
2782 when computing fixed point and variable point multiplication (which
2783 in OpenSSL are mostly used with secret scalars in keygen, sign,
2784 ECDH derive operations).
2785 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2788 * Updated CONTRIBUTING
2792 * Updated DRBG / RAND to request nonce and additional low entropy
2793 randomness from the system.
2795 *Matthias St. Pierre*
2797 * Updated 'openssl rehash' to use OpenSSL consistent default.
2801 * Moved the load of the ssl_conf module to libcrypto, which helps
2802 loading engines that libssl uses before libssl is initialised.
2806 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2810 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2812 *Ingo Schwarze, Rich Salz*
2814 * Added output of accepting IP address and port for 'openssl s_server'
2818 * Added a new API for TLSv1.3 ciphersuites:
2819 SSL_CTX_set_ciphersuites()
2820 SSL_set_ciphersuites()
2824 * Memory allocation failures consistently add an error to the error
2829 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2830 in libcrypto when run as setuid/setgid.
2834 * Load any config file by default when libssl is used.
2838 * Added new public header file <openssl/rand_drbg.h> and documentation
2839 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2841 *Matthias St. Pierre*
2843 * QNX support removed (cannot find contributors to get their approval
2844 for the license change).
2848 * TLSv1.3 replay protection for early data has been implemented. See the
2849 SSL_read_early_data() man page for further details.
2853 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2854 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2855 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2856 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2857 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2858 configuration has been separated out. See the ciphers man page or the
2859 SSL_CTX_set_ciphersuites() man page for more information.
2863 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2864 in responder mode now supports the new "-multi" option, which
2865 spawns the specified number of child processes to handle OCSP
2866 requests. The "-timeout" option now also limits the OCSP
2867 responder's patience to wait to receive the full client request
2868 on a newly accepted connection. Child processes are respawned
2869 as needed, and the CA index file is automatically reloaded
2870 when changed. This makes it possible to run the "ocsp" responder
2871 as a long-running service, making the OpenSSL CA somewhat more
2872 feature-complete. In this mode, most diagnostic messages logged
2873 after entering the event loop are logged via syslog(3) rather than
2878 * Added support for X448 and Ed448. Heavily based on original work by
2883 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2884 objects loaded. This adds the functions OSSL_STORE_expect() and
2885 OSSL_STORE_find() as well as needed tools to construct searches and
2886 get the search data out of them.
2890 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2891 version of OpenSSL should review their configuration settings to ensure
2892 that they are still appropriate for TLSv1.3. For further information see:
2893 <https://wiki.openssl.org/index.php/TLS1.3>
2897 * Grand redesign of the OpenSSL random generator
2899 The default RAND method now utilizes an AES-CTR DRBG according to
2900 NIST standard SP 800-90Ar1. The new random generator is essentially
2901 a port of the default random generator from the OpenSSL FIPS 2.0
2902 object module. It is a hybrid deterministic random bit generator
2903 using an AES-CTR bit stream and which seeds and reseeds itself
2904 automatically using trusted system entropy sources.
2906 Some of its new features are:
2907 - Support for multiple DRBG instances with seed chaining.
2908 - The default RAND method makes use of a DRBG.
2909 - There is a public and private DRBG instance.
2910 - The DRBG instances are fork-safe.
2911 - Keep all global DRBG instances on the secure heap if it is enabled.
2912 - The public and private DRBG instance are per thread for lock free
2915 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2917 * Changed Configure so it only says what it does and doesn't dump
2918 so much data. Instead, ./configdata.pm should be used as a script
2919 to display all sorts of configuration data.
2923 * Added processing of "make variables" to Configure.
2927 * Added SHA512/224 and SHA512/256 algorithm support.
2931 * The last traces of Netware support, first removed in 1.1.0, have
2936 * Get rid of Makefile.shared, and in the process, make the processing
2937 of certain files (rc.obj, or the .def/.map/.opt files produced from
2938 the ordinal files) more visible and hopefully easier to trace and
2939 debug (or make silent).
2943 * Make it possible to have environment variable assignments as
2944 arguments to config / Configure.
2948 * Add multi-prime RSA (RFC 8017) support.
2952 * Add SM3 implemented according to GB/T 32905-2016
2953 *Jack Lloyd <jack.lloyd@ribose.com>,*
2954 *Ronald Tse <ronald.tse@ribose.com>,*
2955 *Erick Borsboom <erick.borsboom@ribose.com>*
2957 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2958 as documented in RFC6066.
2959 Based on a patch from Tomasz Moń
2961 *Filipe Raimundo da Silva*
2963 * Add SM4 implemented according to GB/T 32907-2016.
2964 *Jack Lloyd <jack.lloyd@ribose.com>,*
2965 *Ronald Tse <ronald.tse@ribose.com>,*
2966 *Erick Borsboom <erick.borsboom@ribose.com>*
2968 * Reimplement -newreq-nodes and ERR_error_string_n; the
2969 original author does not agree with the license change.
2973 * Add ARIA AEAD TLS support.
2977 * Some macro definitions to support VS6 have been removed. Visual
2978 Studio 6 has not worked since 1.1.0
2982 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2983 without clearing the errors.
2987 * Add "atfork" functions. If building on a system that without
2988 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2989 requirements. The RAND facility now uses/requires this.
2997 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2998 not possible to disable entirely. However, it's still possible to
2999 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3002 To disable, configure with 'no-ui-console'. 'no-ui' is still
3003 possible to use as an alias. Check at compile time with the
3004 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3005 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3009 * Add a STORE module, which implements a uniform and URI based reader of
3010 stores that can contain keys, certificates, CRLs and numerous other
3011 objects. The main API is loosely based on a few stdio functions,
3012 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3013 OSSL_STORE_error and OSSL_STORE_close.
3014 The implementation uses backends called "loaders" to implement arbitrary
3015 URI schemes. There is one built in "loader" for the 'file' scheme.
3019 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3020 then adjusted to work on FreeBSD 8.4 as well.
3021 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3022 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3026 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3027 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3028 error code calls like this:
3030 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3032 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3033 that can be encoded in C. For the foreseeable future, this will only
3036 *Richard Levitte and Tim Hudson*
3038 * Removed BSD cryptodev engine.
3042 * Add a build target 'build_all_generated', to build all generated files
3043 and only that. This can be used to prepare everything that requires
3044 things like perl for a system that lacks perl and then move everything
3045 to that system and do the rest of the build there.
3049 * In the UI interface, make it possible to duplicate the user data. This
3050 can be used by engines that need to retain the data for a longer time
3051 than just the call where this user data is passed.
3055 * Ignore the '-named_curve auto' value for compatibility of applications
3058 *Tomáš Mráz <tmraz@fedoraproject.org>*
3060 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3061 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3062 alerts across multiple records (some of which could be empty). In practice
3063 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3064 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3065 support this at all. Supporting it adds significant complexity to the
3066 record layer, and its removal is unlikely to cause interoperability
3071 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3072 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3073 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3078 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3079 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3081 *Richard Levitte, Andy Polyakov*
3083 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3088 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3089 platform rather than 'mingw'.
3093 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3094 success if they are asked to add an object which already exists
3095 in the store. This change cascades to other functions which load
3096 certificates and CRLs.
3100 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3101 facilitate stack unwinding even from assembly subroutines.
3105 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3106 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3110 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3111 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3112 which is the minimum version we support.
3116 * Certificate time validation (X509_cmp_time) enforces stricter
3117 compliance with RFC 5280. Fractional seconds and timezone offsets
3118 are no longer allowed.
3122 * Add support for ARIA
3126 * s_client will now send the Server Name Indication (SNI) extension by
3127 default unless the new "-noservername" option is used. The server name is
3128 based on the host provided to the "-connect" option unless overridden by
3129 using "-servername".
3133 * Add support for SipHash
3137 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3138 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3139 prevent issues where no progress is being made and the peer continually
3140 sends unrecognised record types, using up resources processing them.
3144 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3145 using the algorithm defined in
3146 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3150 * Heartbeat support has been removed; the ABI is changed for now.
3152 *Richard Levitte, Rich Salz*
3154 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3158 * The RSA "null" method, which was partially supported to avoid patent
3159 issues, has been replaced to always returns NULL.
3166 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3168 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3169 used even when parsing explicit parameters, when loading a encoded key
3170 or calling `EC_GROUP_new_from_ecpkparameters()`/
3171 `EC_GROUP_new_from_ecparameters()`.
3172 This prevents bypass of security hardening and performance gains,
3173 especially for curves with specialized EC_METHODs.
3174 By default, if a key encoded with explicit parameters is loaded and later
3175 encoded, the output is still encoded with explicit parameters, even if
3176 internally a "named" EC_GROUP is used for computation.
3180 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3181 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3182 NULL. After this change, only the cofactor parameter can be NULL. It also
3183 does some minimal sanity checks on the passed order.
3188 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3189 An attack is simple, if the first CMS_recipientInfo is valid but the
3190 second CMS_recipientInfo is chosen ciphertext. If the second
3191 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3192 encryption key will be replaced by garbage, and the message cannot be
3193 decoded, but if the RSA decryption fails, the correct encryption key is
3194 used and the recipient will not notice the attack.
3195 As a work around for this potential attack the length of the decrypted
3196 key must be equal to the cipher default key length, in case the
3197 certificate is not given and all recipientInfo are tried out.
3198 The old behaviour can be re-enabled in the CMS code by setting the
3199 CMS_DEBUG_DECRYPT flag.
3204 * Use Windows installation paths in the mingw builds
3206 Mingw isn't a POSIX environment per se, which means that Windows
3207 paths should be used for installation.
3212 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3214 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3215 This changes the size when using the `genpkey` command when no size is given.
3216 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3217 generation commands to use 2048 bits by default.
3221 * Prevent over long nonces in ChaCha20-Poly1305.
3223 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3224 for every encryption operation. RFC 7539 specifies that the nonce value
3225 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3226 and front pads the nonce with 0 bytes if it is less than 12
3227 bytes. However it also incorrectly allows a nonce to be set of up to 16
3228 bytes. In this case only the last 12 bytes are significant and any
3229 additional leading bytes are ignored.
3231 It is a requirement of using this cipher that nonce values are
3232 unique. Messages encrypted using a reused nonce value are susceptible to
3233 serious confidentiality and integrity attacks. If an application changes
3234 the default nonce length to be longer than 12 bytes and then makes a
3235 change to the leading bytes of the nonce expecting the new value to be a
3236 new unique nonce then such an application could inadvertently encrypt
3237 messages with a reused nonce.
3239 Additionally the ignored bytes in a long nonce are not covered by the
3240 integrity guarantee of this cipher. Any application that relies on the
3241 integrity of these ignored leading bytes of a long nonce may be further
3242 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3243 is safe because no such use sets such a long nonce value. However user
3244 applications that use this cipher directly and set a non-default nonce
3245 length to be longer than 12 bytes may be vulnerable.
3247 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3253 * Added SCA hardening for modular field inversion in EC_GROUP through
3254 a new dedicated field_inv() pointer in EC_METHOD.
3255 This also addresses a leakage affecting conversions from projective
3256 to affine coordinates.
3258 *Billy Bob Brumley, Nicola Tuveri*
3260 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3261 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3265 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3269 * Remove the 'dist' target and add a tarball building script. The
3270 'dist' target has fallen out of use, and it shouldn't be
3271 necessary to configure just to create a source distribution.
3275 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3277 * Timing vulnerability in DSA signature generation
3279 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3280 timing side channel attack. An attacker could use variations in the signing
3281 algorithm to recover the private key.
3283 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3288 * Timing vulnerability in ECDSA signature generation
3290 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3291 timing side channel attack. An attacker could use variations in the signing
3292 algorithm to recover the private key.
3294 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3299 * Add coordinate blinding for EC_POINT and implement projective
3300 coordinate blinding for generic prime curves as a countermeasure to
3301 chosen point SCA attacks.
3303 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3305 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3307 * Client DoS due to large DH parameter
3309 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3310 malicious server can send a very large prime value to the client. This will
3311 cause the client to spend an unreasonably long period of time generating a
3312 key for this prime resulting in a hang until the client has finished. This
3313 could be exploited in a Denial Of Service attack.
3315 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3320 * Cache timing vulnerability in RSA Key Generation
3322 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3323 a cache timing side channel attack. An attacker with sufficient access to
3324 mount cache timing attacks during the RSA key generation process could
3325 recover the private key.
3327 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3328 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3333 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3334 parameter is no longer accepted, as it leads to a corrupt table. NULL
3335 pem_str is reserved for alias entries only.
3339 * Revert blinding in ECDSA sign and instead make problematic addition
3340 length-invariant. Switch even to fixed-length Montgomery multiplication.
3344 * Change generating and checking of primes so that the error rate of not
3345 being prime depends on the intended use based on the size of the input.
3346 For larger primes this will result in more rounds of Miller-Rabin.
3347 The maximal error rate for primes with more than 1080 bits is lowered
3350 *Kurt Roeckx, Annie Yousar*
3352 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3356 * Add blinding to ECDSA and DSA signatures to protect against side channel
3357 attacks discovered by Keegan Ryan (NCC Group).
3361 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3362 now allow empty (zero character) pass phrases.
3366 * Certificate time validation (X509_cmp_time) enforces stricter
3367 compliance with RFC 5280. Fractional seconds and timezone offsets
3368 are no longer allowed.
3372 * Fixed a text canonicalisation bug in CMS
3374 Where a CMS detached signature is used with text content the text goes
3375 through a canonicalisation process first prior to signing or verifying a
3376 signature. This process strips trailing space at the end of lines, converts
3377 line terminators to CRLF and removes additional trailing line terminators
3378 at the end of a file. A bug in the canonicalisation process meant that
3379 some characters, such as form-feed, were incorrectly treated as whitespace
3380 and removed. This is contrary to the specification (RFC5485). This fix
3381 could mean that detached text data signed with an earlier version of
3382 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3383 signed with a fixed OpenSSL may fail to verify with an earlier version of
3384 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3385 and use the "-binary" flag (for the "cms" command line application) or set
3386 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3390 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3392 * Constructed ASN.1 types with a recursive definition could exceed the stack
3394 Constructed ASN.1 types with a recursive definition (such as can be found
3395 in PKCS7) could eventually exceed the stack given malicious input with
3396 excessive recursion. This could result in a Denial Of Service attack. There
3397 are no such structures used within SSL/TLS that come from untrusted sources
3398 so this is considered safe.
3400 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3406 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3408 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3409 effectively reduced to only comparing the least significant bit of each
3410 byte. This allows an attacker to forge messages that would be considered as
3411 authenticated in an amount of tries lower than that guaranteed by the
3412 security claims of the scheme. The module can only be compiled by the
3413 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3415 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3421 * Add a build target 'build_all_generated', to build all generated files
3422 and only that. This can be used to prepare everything that requires
3423 things like perl for a system that lacks perl and then move everything
3424 to that system and do the rest of the build there.
3428 * Backport SSL_OP_NO_RENGOTIATION
3430 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3431 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3432 changes this is no longer possible in 1.1.0. Therefore, the new
3433 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3434 1.1.0 to provide equivalent functionality.
3436 Note that if an application built against 1.1.0h headers (or above) is run
3437 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3438 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3442 * Removed the OS390-Unix config target. It relied on a script that doesn't
3447 * rsaz_1024_mul_avx2 overflow bug on x86_64
3449 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3450 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3451 Analysis suggests that attacks against RSA and DSA as a result of this
3452 defect would be very difficult to perform and are not believed likely.
3453 Attacks against DH1024 are considered just feasible, because most of the
3454 work necessary to deduce information about a private key may be performed
3455 offline. The amount of resources required for such an attack would be
3456 significant. However, for an attack on TLS to be meaningful, the server
3457 would have to share the DH1024 private key among multiple clients, which is
3458 no longer an option since CVE-2016-0701.
3460 This only affects processors that support the AVX2 but not ADX extensions
3461 like Intel Haswell (4th generation).
3463 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3464 was originally found via the OSS-Fuzz project.
3469 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3471 * bn_sqrx8x_internal carry bug on x86_64
3473 There is a carry propagating bug in the x86_64 Montgomery squaring
3474 procedure. No EC algorithms are affected. Analysis suggests that attacks
3475 against RSA and DSA as a result of this defect would be very difficult to
3476 perform and are not believed likely. Attacks against DH are considered just
3477 feasible (although very difficult) because most of the work necessary to
3478 deduce information about a private key may be performed offline. The amount
3479 of resources required for such an attack would be very significant and
3480 likely only accessible to a limited number of attackers. An attacker would
3481 additionally need online access to an unpatched system using the target
3482 private key in a scenario with persistent DH parameters and a private
3483 key that is shared between multiple clients.
3485 This only affects processors that support the BMI1, BMI2 and ADX extensions
3486 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3488 This issue was reported to OpenSSL by the OSS-Fuzz project.
3493 * Malformed X.509 IPAddressFamily could cause OOB read
3495 If an X.509 certificate has a malformed IPAddressFamily extension,
3496 OpenSSL could do a one-byte buffer overread. The most likely result
3497 would be an erroneous display of the certificate in text format.
3499 This issue was reported to OpenSSL by the OSS-Fuzz project.
3504 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3506 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3507 platform rather than 'mingw'.
3511 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3512 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3513 which is the minimum version we support.
3517 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3519 * Encrypt-Then-Mac renegotiation crash
3521 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3522 negotiated where it was not in the original handshake (or vice-versa) then
3523 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3524 and servers are affected.
3526 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3531 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3533 * Truncated packet could crash via OOB read
3535 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3536 cipher is being used, then a truncated packet can cause that host to
3537 perform an out-of-bounds read, usually resulting in a crash.
3539 This issue was reported to OpenSSL by Robert Święcki of Google.
3544 * Bad (EC)DHE parameters cause a client crash
3546 If a malicious server supplies bad parameters for a DHE or ECDHE key
3547 exchange then this can result in the client attempting to dereference a
3548 NULL pointer leading to a client crash. This could be exploited in a Denial
3551 This issue was reported to OpenSSL by Guido Vranken.
3556 * BN_mod_exp may produce incorrect results on x86_64
3558 There is a carry propagating bug in the x86_64 Montgomery squaring
3559 procedure. No EC algorithms are affected. Analysis suggests that attacks
3560 against RSA and DSA as a result of this defect would be very difficult to
3561 perform and are not believed likely. Attacks against DH are considered just
3562 feasible (although very difficult) because most of the work necessary to
3563 deduce information about a private key may be performed offline. The amount
3564 of resources required for such an attack would be very significant and
3565 likely only accessible to a limited number of attackers. An attacker would
3566 additionally need online access to an unpatched system using the target
3567 private key in a scenario with persistent DH parameters and a private
3568 key that is shared between multiple clients. For example this can occur by
3569 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3570 similar to CVE-2015-3193 but must be treated as a separate problem.
3572 This issue was reported to OpenSSL by the OSS-Fuzz project.
3577 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3579 * ChaCha20/Poly1305 heap-buffer-overflow
3581 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3582 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3583 crash. This issue is not considered to be exploitable beyond a DoS.
3585 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3590 * CMS Null dereference
3592 Applications parsing invalid CMS structures can crash with a NULL pointer
3593 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3594 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3595 structure callback if an attempt is made to free certain invalid encodings.
3596 Only CHOICE structures using a callback which do not handle NULL value are
3599 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3604 * Montgomery multiplication may produce incorrect results
3606 There is a carry propagating bug in the Broadwell-specific Montgomery
3607 multiplication procedure that handles input lengths divisible by, but
3608 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3609 and DH private keys are impossible. This is because the subroutine in
3610 question is not used in operations with the private key itself and an input
3611 of the attacker's direct choice. Otherwise the bug can manifest itself as
3612 transient authentication and key negotiation failures or reproducible
3613 erroneous outcome of public-key operations with specially crafted input.
3614 Among EC algorithms only Brainpool P-512 curves are affected and one
3615 presumably can attack ECDH key negotiation. Impact was not analyzed in
3616 detail, because pre-requisites for attack are considered unlikely. Namely
3617 multiple clients have to choose the curve in question and the server has to
3618 share the private key among them, neither of which is default behaviour.
3619 Even then only clients that chose the curve will be affected.
3621 This issue was publicly reported as transient failures and was not
3622 initially recognized as a security issue. Thanks to Richard Morgan for
3623 providing reproducible case.
3628 * Removed automatic addition of RPATH in shared libraries and executables,
3629 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3633 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3635 * Fix Use After Free for large message sizes
3637 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3638 message larger than approx 16k is received then the underlying buffer to
3639 store the incoming message is reallocated and moved. Unfortunately a
3640 dangling pointer to the old location is left which results in an attempt to
3641 write to the previously freed location. This is likely to result in a
3642 crash, however it could potentially lead to execution of arbitrary code.
3644 This issue only affects OpenSSL 1.1.0a.
3646 This issue was reported to OpenSSL by Robert Święcki.
3651 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3653 * OCSP Status Request extension unbounded memory growth
3655 A malicious client can send an excessively large OCSP Status Request
3656 extension. If that client continually requests renegotiation, sending a
3657 large OCSP Status Request extension each time, then there will be unbounded
3658 memory growth on the server. This will eventually lead to a Denial Of
3659 Service attack through memory exhaustion. Servers with a default
3660 configuration are vulnerable even if they do not support OCSP. Builds using
3661 the "no-ocsp" build time option are not affected.
3663 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3668 * SSL_peek() hang on empty record
3670 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3671 sends an empty record. This could be exploited by a malicious peer in a
3672 Denial Of Service attack.
3674 This issue was reported to OpenSSL by Alex Gaynor.
3679 * Excessive allocation of memory in tls_get_message_header() and
3680 dtls1_preprocess_fragment()
3682 A (D)TLS message includes 3 bytes for its length in the header for the
3683 message. This would allow for messages up to 16Mb in length. Messages of
3684 this length are excessive and OpenSSL includes a check to ensure that a
3685 peer is sending reasonably sized messages in order to avoid too much memory
3686 being consumed to service a connection. A flaw in the logic of version
3687 1.1.0 means that memory for the message is allocated too early, prior to
3688 the excessive message length check. Due to way memory is allocated in
3689 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3690 to service a connection. This could lead to a Denial of Service through
3691 memory exhaustion. However, the excessive message length check still takes
3692 place, and this would cause the connection to immediately fail. Assuming
3693 that the application calls SSL_free() on the failed connection in a timely
3694 manner then the 21Mb of allocated memory will then be immediately freed
3695 again. Therefore, the excessive memory allocation will be transitory in
3696 nature. This then means that there is only a security impact if:
3698 1) The application does not call SSL_free() in a timely manner in the event
3699 that the connection fails
3701 2) The application is working in a constrained environment where there is
3702 very little free memory
3704 3) The attacker initiates multiple connection attempts such that there are
3705 multiple connections in a state where memory has been allocated for the
3706 connection; SSL_free() has not yet been called; and there is insufficient
3707 memory to service the multiple requests.
3709 Except in the instance of (1) above any Denial Of Service is likely to be
3710 transitory because as soon as the connection fails the memory is
3711 subsequently freed again in the SSL_free() call. However there is an
3712 increased risk during this period of application crashes due to the lack of
3713 memory - which would then mean a more serious Denial of Service.
3715 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3716 (CVE-2016-6307 and CVE-2016-6308)
3720 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3721 had to be removed. Primary reason is that vendor assembler can't
3722 assemble our modules with -KPIC flag. As result it, assembly
3723 support, was not even available as option. But its lack means
3724 lack of side-channel resistant code, which is incompatible with
3725 security by todays standards. Fortunately gcc is readily available
3726 prepackaged option, which we firmly point at...
3730 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3732 * Windows command-line tool supports UTF-8 opt-in option for arguments
3733 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3734 (to any value) allows Windows user to access PKCS#12 file generated
3735 with Windows CryptoAPI and protected with non-ASCII password, as well
3736 as files generated under UTF-8 locale on Linux also protected with
3741 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3742 have been disabled by default and removed from DEFAULT, just like RC4.
3743 See the RC4 item below to re-enable both.
3747 * The method for finding the storage location for the Windows RAND seed file
3748 has changed. First we check %RANDFILE%. If that is not set then we check
3749 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3750 all else fails we fall back to C:\.
3754 * The EVP_EncryptUpdate() function has had its return type changed from void
3755 to int. A return of 0 indicates and error while a return of 1 indicates
3760 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3761 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3762 off the constant time implementation for RSA, DSA and DH have been made
3763 no-ops and deprecated.
3767 * Windows RAND implementation was simplified to only get entropy by
3768 calling CryptGenRandom(). Various other RAND-related tickets
3771 *Joseph Wylie Yandle, Rich Salz*
3773 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3774 and `OPENSSL_LH_`, respectively. The old names are available
3775 with API compatibility. They new names are now completely documented.
3779 * Unify TYPE_up_ref(obj) methods signature.
3780 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3781 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3782 int (instead of void) like all others TYPE_up_ref() methods.
3783 So now these methods also check the return value of CRYPTO_atomic_add(),
3784 and the validity of object reference counter.
3786 *fdasilvayy@gmail.com*
3788 * With Windows Visual Studio builds, the .pdb files are installed
3789 alongside the installed libraries and executables. For a static
3790 library installation, ossl_static.pdb is the associate compiler
3791 generated .pdb file to be used when linking programs.
3795 * Remove openssl.spec. Packaging files belong with the packagers.
3799 * Automatic Darwin/OSX configuration has had a refresh, it will now
3800 recognise x86_64 architectures automatically. You can still decide
3801 to build for a different bitness with the environment variable
3802 KERNEL_BITS (can be 32 or 64), for example:
3804 KERNEL_BITS=32 ./config
3808 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3809 256 bit AES and HMAC with SHA256.
3813 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3817 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3821 * To enable users to have their own config files and build file templates,
3822 Configure looks in the directory indicated by the environment variable
3823 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3824 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3825 name and is used as is.
3829 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3830 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3831 X509_CERT_FILE_CTX was removed.
3835 * "shared" builds are now the default. To create only static libraries use
3836 the "no-shared" Configure option.
3840 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3841 All of these option have not worked for some while and are fundamental
3846 * Make various cleanup routines no-ops and mark them as deprecated. Most
3847 global cleanup functions are no longer required because they are handled
3848 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3849 Explicitly de-initing can cause problems (e.g. where a library that uses
3850 OpenSSL de-inits, but an application is still using it). The affected
3851 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3852 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3853 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3854 COMP_zlib_cleanup().
3858 * --strict-warnings no longer enables runtime debugging options
3859 such as REF_DEBUG. Instead, debug options are automatically
3860 enabled with '--debug' builds.
3862 *Andy Polyakov, Emilia Käsper*
3864 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3865 have been moved out of the public header files. New functions for managing
3866 these have been added.
3870 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3871 objects have been moved out of the public header files. New
3872 functions for managing these have been added.
3876 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3877 have been moved out of the public header files. New functions for managing
3878 these have been added.
3882 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3883 moved out of the public header files. New functions for managing these
3888 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3892 * Removed the mk1mf build scripts.
3896 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3897 it is always safe to #include a header now.
3901 * Removed the aged BC-32 config and all its supporting scripts
3905 * Removed support for Ultrix, Netware, and OS/2.
3909 * Add support for HKDF.
3911 *Alessandro Ghedini*
3913 * Add support for blake2b and blake2s
3917 * Added support for "pipelining". Ciphers that have the
3918 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3919 encryptions/decryptions simultaneously. There are currently no built-in
3920 ciphers with this property but the expectation is that engines will be able
3921 to offer it to significantly improve throughput. Support has been extended
3922 into libssl so that multiple records for a single connection can be
3923 processed in one go (for >=TLS 1.1).
3927 * Added the AFALG engine. This is an async capable engine which is able to
3928 offload work to the Linux kernel. In this initial version it only supports
3929 AES128-CBC. The kernel must be version 4.1.0 or greater.
3933 * OpenSSL now uses a new threading API. It is no longer necessary to
3934 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3935 are two supported threading models: pthreads and windows threads. It is
3936 also possible to configure OpenSSL at compile time for "no-threads". The
3937 old threading API should no longer be used. The functions have been
3938 replaced with "no-op" compatibility macros.
3940 *Alessandro Ghedini, Matt Caswell*
3942 * Modify behavior of ALPN to invoke callback after SNI/servername
3943 callback, such that updates to the SSL_CTX affect ALPN.
3947 * Add SSL_CIPHER queries for authentication and key-exchange.
3951 * Changes to the DEFAULT cipherlist:
3952 - Prefer (EC)DHE handshakes over plain RSA.
3953 - Prefer AEAD ciphers over legacy ciphers.
3954 - Prefer ECDSA over RSA when both certificates are available.
3955 - Prefer TLSv1.2 ciphers/PRF.
3956 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3961 * Change the ECC default curve list to be this, in order: x25519,
3962 secp256r1, secp521r1, secp384r1.
3966 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3967 disabled by default. They can be re-enabled using the
3968 enable-weak-ssl-ciphers option to Configure.
3972 * If the server has ALPN configured, but supports no protocols that the
3973 client advertises, send a fatal "no_application_protocol" alert.
3974 This behaviour is SHALL in RFC 7301, though it isn't universally
3975 implemented by other servers.
3979 * Add X25519 support.
3980 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3981 for public and private key encoding using the format documented in
3982 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3983 key generation and key derivation.
3985 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3990 * Deprecate SRP_VBASE_get_by_user.
3991 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3992 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3993 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3994 seed, even if the seed is configured.
3996 Users should use SRP_VBASE_get1_by_user instead. Note that in
3997 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3998 also that even though configuring the SRP seed attempts to hide
3999 invalid usernames by continuing the handshake with fake
4000 credentials, this behaviour is not constant time and no strong
4001 guarantees are made that the handshake is indistinguishable from
4002 that of a valid user.
4006 * Configuration change; it's now possible to build dynamic engines
4007 without having to build shared libraries and vice versa. This
4008 only applies to the engines in `engines/`, those in `crypto/engine/`
4009 will always be built into libcrypto (i.e. "static").
4011 Building dynamic engines is enabled by default; to disable, use
4012 the configuration option "disable-dynamic-engine".
4014 The only requirements for building dynamic engines are the
4015 presence of the DSO module and building with position independent
4016 code, so they will also automatically be disabled if configuring
4017 with "disable-dso" or "disable-pic".
4019 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4020 are also taken away from openssl/opensslconf.h, as they are
4025 * Configuration change; if there is a known flag to compile
4026 position independent code, it will always be applied on the
4027 libcrypto and libssl object files, and never on the application
4028 object files. This means other libraries that use routines from
4029 libcrypto / libssl can be made into shared libraries regardless
4030 of how OpenSSL was configured.
4032 If this isn't desirable, the configuration options "disable-pic"
4033 or "no-pic" can be used to disable the use of PIC. This will
4034 also disable building shared libraries and dynamic engines.
4038 * Removed JPAKE code. It was experimental and has no wide use.
4042 * The INSTALL_PREFIX Makefile variable has been renamed to
4043 DESTDIR. That makes for less confusion on what this variable
4044 is for. Also, the configuration option --install_prefix is
4049 * Heartbeat for TLS has been removed and is disabled by default
4050 for DTLS; configure with enable-heartbeats. Code that uses the
4051 old #define's might need to be updated.
4053 *Emilia Käsper, Rich Salz*
4055 * Rename REF_CHECK to REF_DEBUG.
4059 * New "unified" build system
4061 The "unified" build system is aimed to be a common system for all
4062 platforms we support. With it comes new support for VMS.
4064 This system builds supports building in a different directory tree
4065 than the source tree. It produces one Makefile (for unix family
4066 or lookalikes), or one descrip.mms (for VMS).
4068 The source of information to make the Makefile / descrip.mms is
4069 small files called 'build.info', holding the necessary
4070 information for each directory with source to compile, and a
4071 template in Configurations, like unix-Makefile.tmpl or
4074 With this change, the library names were also renamed on Windows
4075 and on VMS. They now have names that are closer to the standard
4076 on Unix, and include the major version number, and in certain
4077 cases, the architecture they are built for. See "Notes on shared
4078 libraries" in INSTALL.
4080 We rely heavily on the perl module Text::Template.
4084 * Added support for auto-initialisation and de-initialisation of the library.
4085 OpenSSL no longer requires explicit init or deinit routines to be called,
4086 except in certain circumstances. See the OPENSSL_init_crypto() and
4087 OPENSSL_init_ssl() man pages for further information.
4091 * The arguments to the DTLSv1_listen function have changed. Specifically the
4092 "peer" argument is now expected to be a BIO_ADDR object.
4094 * Rewrite of BIO networking library. The BIO library lacked consistent
4095 support of IPv6, and adding it required some more extensive
4096 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4097 which hold all types of addresses and chains of address information.
4098 It also introduces a new API, with functions like BIO_socket,
4099 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4100 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4101 have been adapted accordingly.
4105 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4110 * CRIME protection: disable compression by default, even if OpenSSL is
4111 compiled with zlib enabled. Applications can still enable compression
4112 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4113 using the SSL_CONF library to configure compression.
4117 * The signature of the session callback configured with
4118 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4119 was explicitly marked as `const unsigned char*` instead of
4124 * Always DPURIFY. Remove the use of uninitialized memory in the
4125 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4129 * Removed many obsolete configuration items, including
4130 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4131 MD2_CHAR, MD2_INT, MD2_LONG
4133 IDEA_SHORT, IDEA_LONG
4134 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4136 *Rich Salz, with advice from Andy Polyakov*
4138 * Many BN internals have been moved to an internal header file.
4140 *Rich Salz with help from Andy Polyakov*
4142 * Configuration and writing out the results from it has changed.
4143 Files such as Makefile include/openssl/opensslconf.h and are now
4144 produced through general templates, such as Makefile.in and
4145 crypto/opensslconf.h.in and some help from the perl module
4148 Also, the center of configuration information is no longer
4149 Makefile. Instead, Configure produces a perl module in
4150 configdata.pm which holds most of the config data (in the hash
4151 table %config), the target data that comes from the target
4152 configuration in one of the `Configurations/*.conf` files (in
4157 * To clarify their intended purposes, the Configure options
4158 --prefix and --openssldir change their semantics, and become more
4159 straightforward and less interdependent.
4161 --prefix shall be used exclusively to give the location INSTALLTOP
4162 where programs, scripts, libraries, include files and manuals are
4163 going to be installed. The default is now /usr/local.
4165 --openssldir shall be used exclusively to give the default
4166 location OPENSSLDIR where certificates, private keys, CRLs are
4167 managed. This is also where the default openssl.cnf gets
4169 If the directory given with this option is a relative path, the
4170 values of both the --prefix value and the --openssldir value will
4171 be combined to become OPENSSLDIR.
4172 The default for --openssldir is INSTALLTOP/ssl.
4174 Anyone who uses --openssldir to specify where OpenSSL is to be
4175 installed MUST change to use --prefix instead.
4179 * The GOST engine was out of date and therefore it has been removed. An up
4180 to date GOST engine is now being maintained in an external repository.
4181 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4182 support for GOST ciphersuites (these are only activated if a GOST engine
4187 * EGD is no longer supported by default; use enable-egd when
4190 *Ben Kaduk and Rich Salz*
4192 * The distribution now has Makefile.in files, which are used to
4193 create Makefile's when Configure is run. *Configure must be run
4194 before trying to build now.*
4198 * The return value for SSL_CIPHER_description() for error conditions
4203 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4205 Obtaining and performing DNSSEC validation of TLSA records is
4206 the application's responsibility. The application provides
4207 the TLSA records of its choice to OpenSSL, and these are then
4208 used to authenticate the peer.
4210 The TLSA records need not even come from DNS. They can, for
4211 example, be used to implement local end-entity certificate or
4212 trust-anchor "pinning", where the "pin" data takes the form
4213 of TLSA records, which can augment or replace verification
4214 based on the usual WebPKI public certification authorities.
4218 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4219 continues to support deprecated interfaces in default builds.
4220 However, applications are strongly advised to compile their
4221 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4222 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4223 or the 1.1.0 releases.
4225 In environments in which all applications have been ported to
4226 not use any deprecated interfaces OpenSSL's Configure script
4227 should be used with the --api=1.1.0 option to entirely remove
4228 support for the deprecated features from the library and
4229 unconditionally disable them in the installed headers.
4230 Essentially the same effect can be achieved with the "no-deprecated"
4231 argument to Configure, except that this will always restrict
4232 the build to just the latest API, rather than a fixed API
4235 As applications are ported to future revisions of the API,
4236 they should update their compile-time OPENSSL_API_COMPAT define
4237 accordingly, but in most cases should be able to continue to
4238 compile with later releases.
4240 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4241 0x10000000L and 0x00908000L, respectively. However those
4242 versions did not support the OPENSSL_API_COMPAT feature, and
4243 so applications are not typically tested for explicit support
4244 of just the undeprecated features of either release.
4248 * Add support for setting the minimum and maximum supported protocol.
4249 It can bet set via the SSL_set_min_proto_version() and
4250 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4251 MaxProtocol. It's recommended to use the new APIs to disable
4252 protocols instead of disabling individual protocols using
4253 SSL_set_options() or SSL_CONF's Protocol. This change also
4254 removes support for disabling TLS 1.2 in the OpenSSL TLS
4255 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4259 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4263 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4264 and integrates ECDSA and ECDH functionality into EC. Implementations can
4265 now redirect key generation and no longer need to convert to or from
4268 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4269 include the ec.h header file instead.
4273 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4274 ciphers who are no longer supported and drops support the ephemeral RSA key
4275 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4279 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4280 opaque. For HMAC_CTX, the following constructors and destructors
4283 HMAC_CTX *HMAC_CTX_new(void);
4284 void HMAC_CTX_free(HMAC_CTX *ctx);
4286 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4287 destroy such methods has been added. See EVP_MD_meth_new(3) and
4288 EVP_CIPHER_meth_new(3) for documentation.
4291 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4292 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4293 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4294 an already created structure.
4295 2) For consistency with the majority of our object creators and
4296 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4297 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4298 for deprecated builds.
4302 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4303 cryptographic operations to be performed asynchronously as long as an
4304 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4305 further details. Libssl has also had this capability integrated with the
4306 introduction of the new mode SSL_MODE_ASYNC and associated error
4307 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4308 pages. This work was developed in partnership with Intel Corp.
4312 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4313 always enabled now. If you want to disable the support you should
4314 exclude it using the list of supported ciphers. This also means that the
4315 "-no_ecdhe" option has been removed from s_server.
4319 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4320 SSL_{CTX_}set1_curves() which can set a list.
4324 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4325 curve you want to support using SSL_{CTX_}set1_curves().
4329 * State machine rewrite. The state machine code has been significantly
4330 refactored in order to remove much duplication of code and solve issues
4331 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4332 further details). This change does have some associated API changes.
4333 Notably the SSL_state() function has been removed and replaced by
4334 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4335 SSL_set_state() has been removed altogether. The previous handshake states
4336 defined in ssl.h and ssl3.h have also been removed.
4340 * All instances of the string "ssleay" in the public API were replaced
4341 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4342 Some error codes related to internal RSA_eay API's were renamed.
4346 * The demo files in crypto/threads were moved to demo/threads.
4350 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4353 *Matt Caswell, Rich Salz*
4355 * New ASN.1 embed macro.
4357 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4358 structure is not allocated: it is part of the parent. That is instead of
4366 This reduces memory fragmentation and make it impossible to accidentally
4367 set a mandatory field to NULL.
4369 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4370 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4371 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4376 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4380 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4381 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4382 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4383 DES and RC4 ciphersuites.
4387 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4388 This changes the decoding behaviour for some invalid messages,
4389 though the change is mostly in the more lenient direction, and
4390 legacy behaviour is preserved as much as possible.
4394 * Fix no-stdio build.
4395 *David Woodhouse <David.Woodhouse@intel.com> and also*
4396 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4398 * New testing framework
4399 The testing framework has been largely rewritten and is now using
4400 perl and the perl modules Test::Harness and an extended variant of
4401 Test::More called OpenSSL::Test to do its work. All test scripts in
4402 test/ have been rewritten into test recipes, and all direct calls to
4403 executables in test/Makefile have become individual recipes using the
4404 simplified testing OpenSSL::Test::Simple.
4406 For documentation on our testing modules, do:
4408 perldoc test/testlib/OpenSSL/Test/Simple.pm
4409 perldoc test/testlib/OpenSSL/Test.pm
4413 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4414 are used; the latter aborts on memory leaks (usually checked on exit).
4415 Some undocumented "set malloc, etc., hooks" functions were removed
4416 and others were changed. All are now documented.
4420 * In DSA_generate_parameters_ex, if the provided seed is too short,
4423 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4425 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4426 from RFC4279, RFC4785, RFC5487, RFC5489.
4428 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4429 original RSA_PSK patch.
4433 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4434 era flag was never set throughout the codebase (only read). Also removed
4435 SSL3_FLAGS_POP_BUFFER which was only used if
4436 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4440 * Changed the default name options in the "ca", "crl", "req" and "x509"
4441 to be "oneline" instead of "compat".
4445 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4446 not aware of clients that still exhibit this bug, and the workaround
4447 hasn't been working properly for a while.
4451 * The return type of BIO_number_read() and BIO_number_written() as well as
4452 the corresponding num_read and num_write members in the BIO structure has
4453 changed from unsigned long to uint64_t. On platforms where an unsigned
4454 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4459 * Given the pervasive nature of TLS extensions it is inadvisable to run
4460 OpenSSL without support for them. It also means that maintaining
4461 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4462 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4466 * Removed support for the two export grade static DH ciphersuites
4467 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4468 were newly added (along with a number of other static DH ciphersuites) to
4469 1.0.2. However the two export ones have *never* worked since they were
4470 introduced. It seems strange in any case to be adding new export
4471 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4475 * Version negotiation has been rewritten. In particular SSLv23_method(),
4476 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4477 and turned into macros which simply call the new preferred function names
4478 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4479 should use the new names instead. Also as part of this change the ssl23.h
4480 header file has been removed.
4484 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4485 code and the associated standard is no longer considered fit-for-purpose.
4489 * RT2547 was closed. When generating a private key, try to make the
4490 output file readable only by the owner. This behavior change might
4491 be noticeable when interacting with other software.
4493 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4498 * Added HTTP GET support to the ocsp command.
4502 * Changed default digest for the dgst and enc commands from MD5 to
4507 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4511 * Added support for TLS extended master secret from
4512 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4513 initial patch which was a great help during development.
4517 * All libssl internal structures have been removed from the public header
4518 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4519 now redundant). Users should not attempt to access internal structures
4520 directly. Instead they should use the provided API functions.
4524 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4525 Access to deprecated functions can be re-enabled by running config with
4526 "enable-deprecated". In addition applications wishing to use deprecated
4527 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4528 will, by default, disable some transitive includes that previously existed
4529 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4533 * Added support for OCB mode. OpenSSL has been granted a patent license
4534 compatible with the OpenSSL license for use of OCB. Details are available
4535 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4536 for OCB can be removed by calling config with no-ocb.
4540 * SSLv2 support has been removed. It still supports receiving an SSLv2
4541 compatible client hello.
4545 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4546 done while fixing the error code for the key-too-small case.
4548 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4550 * CA.sh has been removed; use CA.pl instead.
4554 * Removed old DES API.
4558 * Remove various unsupported platforms:
4564 Sinix/ReliantUNIX RM400
4569 16-bit platforms such as WIN16
4573 * Clean up OPENSSL_NO_xxx #define's
4574 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4575 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4576 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4577 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4578 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4579 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4580 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4581 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4582 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4583 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4587 * Cleaned up dead code
4588 Remove all but one '#ifdef undef' which is to be looked at.
4592 * Clean up calling of xxx_free routines.
4593 Just like free(), fix most of the xxx_free routines to accept
4594 NULL. Remove the non-null checks from callers. Save much code.
4598 * Add secure heap for storage of private keys (when possible).
4599 Add BIO_s_secmem(), CBIGNUM, etc.
4600 Contributed by Akamai Technologies under our Corporate CLA.
4604 * Experimental support for a new, fast, unbiased prime candidate generator,
4605 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4607 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4609 * New output format NSS in the sess_id command line tool. This allows
4610 exporting the session id and the master key in NSS keylog format.
4612 *Martin Kaiser <martin@kaiser.cx>*
4614 * Harmonize version and its documentation. -f flag is used to display
4617 *mancha <mancha1@zoho.com>*
4619 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4620 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4622 *mancha <mancha1@zoho.com>*
4624 * Fix some double frees. These are not thought to be exploitable.
4626 *mancha <mancha1@zoho.com>*
4628 * A missing bounds check in the handling of the TLS heartbeat extension
4629 can be used to reveal up to 64k of memory to a connected client or
4632 Thanks for Neel Mehta of Google Security for discovering this bug and to
4633 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4634 preparing the fix ([CVE-2014-0160])
4636 *Adam Langley, Bodo Moeller*
4638 * Fix for the attack described in the paper "Recovering OpenSSL
4639 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4640 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4641 <http://eprint.iacr.org/2014/140>
4643 Thanks to Yuval Yarom and Naomi Benger for discovering this
4644 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4646 *Yuval Yarom and Naomi Benger*
4648 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4649 this fixes a limitation in previous versions of OpenSSL.
4653 * Experimental encrypt-then-mac support.
4655 Experimental support for encrypt then mac from
4656 draft-gutmann-tls-encrypt-then-mac-02.txt
4658 To enable it set the appropriate extension number (0x42 for the test
4659 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4661 For non-compliant peers (i.e. just about everything) this should have no
4664 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4668 * Add EVP support for key wrapping algorithms, to avoid problems with
4669 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4670 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4671 algorithms and include tests cases.
4675 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4680 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4681 MGF1 digest and OAEP label.
4685 * Make openssl verify return errors.
4687 *Chris Palmer <palmer@google.com> and Ben Laurie*
4689 * New function ASN1_TIME_diff to calculate the difference between two
4690 ASN1_TIME structures or one structure and the current time.
4694 * Update fips_test_suite to support multiple command line options. New
4695 test to induce all self test errors in sequence and check expected
4700 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4701 sign or verify all in one operation.
4705 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4706 test programs and fips_test_suite. Includes functionality to parse
4707 the minimal script output of fipsalgest.pl directly.
4711 * Add authorisation parameter to FIPS_module_mode_set().
4715 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4719 * Use separate DRBG fields for internal and external flags. New function
4720 FIPS_drbg_health_check() to perform on demand health checking. Add
4721 generation tests to fips_test_suite with reduced health check interval to
4722 demonstrate periodic health checking. Add "nodh" option to
4723 fips_test_suite to skip very slow DH test.
4727 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4732 * More extensive health check for DRBG checking many more failure modes.
4733 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4734 combination: call this in fips_test_suite.
4738 * Add support for canonical generation of DSA parameter 'g'. See
4741 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4742 POST to handle HMAC cases.
4746 * Add functions FIPS_module_version() and FIPS_module_version_text()
4747 to return numerical and string versions of the FIPS module number.
4751 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4752 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4753 outside the validated module in the FIPS capable OpenSSL.
4757 * Minor change to DRBG entropy callback semantics. In some cases
4758 there is no multiple of the block length between min_len and
4759 max_len. Allow the callback to return more than max_len bytes
4760 of entropy but discard any extra: it is the callback's responsibility
4761 to ensure that the extra data discarded does not impact the
4762 requested amount of entropy.
4766 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4767 information in FIPS186-3, SP800-57 and SP800-131A.
4771 * CCM support via EVP. Interface is very similar to GCM case except we
4772 must supply all data in one chunk (i.e. no update, final) and the
4773 message length must be supplied if AAD is used. Add algorithm test
4778 * Initial version of POST overhaul. Add POST callback to allow the status
4779 of POST to be monitored and/or failures induced. Modify fips_test_suite
4780 to use callback. Always run all selftests even if one fails.
4784 * XTS support including algorithm test driver in the fips_gcmtest program.
4785 Note: this does increase the maximum key length from 32 to 64 bytes but
4786 there should be no binary compatibility issues as existing applications
4787 will never use XTS mode.
4791 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4792 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4793 performs algorithm blocking for unapproved PRNG types. Also do not
4794 set PRNG type in FIPS_mode_set(): leave this to the application.
4795 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4796 the standard OpenSSL PRNG: set additional data to a date time vector.
4800 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4801 This shouldn't present any incompatibility problems because applications
4802 shouldn't be using these directly and any that are will need to rethink
4803 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4807 * Extensive self tests and health checking required by SP800-90 DRBG.
4808 Remove strength parameter from FIPS_drbg_instantiate and always
4809 instantiate at maximum supported strength.
4813 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4817 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4821 * New function DH_compute_key_padded() to compute a DH key and pad with
4822 leading zeroes if needed: this complies with SP800-56A et al.
4826 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4827 anything, incomplete, subject to change and largely untested at present.
4831 * Modify fipscanisteronly build option to only build the necessary object
4832 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4836 * Add experimental option FIPSSYMS to give all symbols in
4837 fipscanister.o and FIPS or fips prefix. This will avoid
4838 conflicts with future versions of OpenSSL. Add perl script
4839 util/fipsas.pl to preprocess assembly language source files
4840 and rename any affected symbols.
4844 * Add selftest checks and algorithm block of non-fips algorithms in
4845 FIPS mode. Remove DES2 from selftests.
4849 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4850 return internal method without any ENGINE dependencies. Add new
4851 tiny fips sign and verify functions.
4855 * New build option no-ec2m to disable characteristic 2 code.
4859 * New build option "fipscanisteronly". This only builds fipscanister.o
4860 and (currently) associated fips utilities. Uses the file Makefile.fips
4861 instead of Makefile.org as the prototype.
4865 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4866 Update fips_gcmtest to use IV generator.
4870 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4871 setting output buffer to NULL. The `*Final` function must be
4872 called although it will not retrieve any additional data. The tag
4873 can be set or retrieved with a ctrl. The IV length is by default 12
4874 bytes (96 bits) but can be set to an alternative value. If the IV
4875 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4880 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4881 underlying do_cipher function handles all cipher semantics itself
4882 including padding and finalisation. This is useful if (for example)
4883 an ENGINE cipher handles block padding itself. The behaviour of
4884 do_cipher is subtly changed if this flag is set: the return value
4885 is the number of characters written to the output buffer (zero is
4886 no longer an error code) or a negative error code. Also if the
4887 input buffer is NULL and length 0 finalisation should be performed.
4891 * If a candidate issuer certificate is already part of the constructed
4892 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4896 * Improve forward-security support: add functions
4898 void SSL_CTX_set_not_resumable_session_callback(
4899 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4900 void SSL_set_not_resumable_session_callback(
4901 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4903 for use by SSL/TLS servers; the callback function will be called whenever a
4904 new session is created, and gets to decide whether the session may be
4905 cached to make it resumable (return 0) or not (return 1). (As by the
4906 SSL/TLS protocol specifications, the session_id sent by the server will be
4907 empty to indicate that the session is not resumable; also, the server will
4908 not generate RFC 4507 (RFC 5077) session tickets.)
4910 A simple reasonable callback implementation is to return is_forward_secure.
4911 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4912 by the SSL/TLS server library, indicating whether it can provide forward
4915 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4917 * New -verify_name option in command line utilities to set verification
4922 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4923 Add CMAC pkey methods.
4927 * Experimental renegotiation in s_server -www mode. If the client
4928 browses /reneg connection is renegotiated. If /renegcert it is
4929 renegotiated requesting a certificate.
4933 * Add an "external" session cache for debugging purposes to s_server. This
4934 should help trace issues which normally are only apparent in deployed
4935 multi-process servers.
4939 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4940 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4941 BIO_set_cipher() and some obscure PEM functions were changed so they
4942 can now return an error. The RAND changes required a change to the
4943 RAND_METHOD structure.
4947 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4948 a gcc attribute to warn if the result of a function is ignored. This
4949 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4950 whose return value is often ignored.
4954 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4955 These allow SCTs (signed certificate timestamps) to be requested and
4956 validated when establishing a connection.
4958 *Rob Percival <robpercival@google.com>*
4963 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4965 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4966 used even when parsing explicit parameters, when loading a encoded key
4967 or calling `EC_GROUP_new_from_ecpkparameters()`/
4968 `EC_GROUP_new_from_ecparameters()`.
4969 This prevents bypass of security hardening and performance gains,
4970 especially for curves with specialized EC_METHODs.
4971 By default, if a key encoded with explicit parameters is loaded and later
4972 encoded, the output is still encoded with explicit parameters, even if
4973 internally a "named" EC_GROUP is used for computation.
4977 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4978 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4979 NULL. After this change, only the cofactor parameter can be NULL. It also
4980 does some minimal sanity checks on the passed order.
4985 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4986 An attack is simple, if the first CMS_recipientInfo is valid but the
4987 second CMS_recipientInfo is chosen ciphertext. If the second
4988 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4989 encryption key will be replaced by garbage, and the message cannot be
4990 decoded, but if the RSA decryption fails, the correct encryption key is
4991 used and the recipient will not notice the attack.
4992 As a work around for this potential attack the length of the decrypted
4993 key must be equal to the cipher default key length, in case the
4994 certificate is not given and all recipientInfo are tried out.
4995 The old behaviour can be re-enabled in the CMS code by setting the
4996 CMS_DEBUG_DECRYPT flag.
5001 * Document issue with installation paths in diverse Windows builds
5003 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5004 binaries and run-time config file.
5009 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5011 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5012 This changes the size when using the `genpkey` command when no size is given.
5013 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5014 generation commands to use 2048 bits by default.
5018 * Add FIPS support for Android Arm 64-bit
5020 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5021 Module in Version 2.0.10. For some reason, the corresponding target
5022 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5023 built with FIPS support on Android Arm 64-bit. This omission has been
5026 *Matthias St. Pierre*
5028 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5030 * 0-byte record padding oracle
5032 If an application encounters a fatal protocol error and then calls
5033 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5034 then OpenSSL can respond differently to the calling application if a 0 byte
5035 record is received with invalid padding compared to if a 0 byte record is
5036 received with an invalid MAC. If the application then behaves differently
5037 based on that in a way that is detectable to the remote peer, then this
5038 amounts to a padding oracle that could be used to decrypt data.
5040 In order for this to be exploitable "non-stitched" ciphersuites must be in
5041 use. Stitched ciphersuites are optimised implementations of certain
5042 commonly used ciphersuites. Also the application must call SSL_shutdown()
5043 twice even if a protocol error has occurred (applications should not do
5044 this but some do anyway).
5046 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5047 Aviram, with additional investigation by Steven Collison and Andrew
5048 Hourselt. It was reported to OpenSSL on 10th December 2018.
5053 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5057 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5059 * Microarchitecture timing vulnerability in ECC scalar multiplication
5061 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5062 shown to be vulnerable to a microarchitecture timing side channel attack.
5063 An attacker with sufficient access to mount local timing attacks during
5064 ECDSA signature generation could recover the private key.
5066 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5067 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5073 * Timing vulnerability in DSA signature generation
5075 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5076 timing side channel attack. An attacker could use variations in the signing
5077 algorithm to recover the private key.
5079 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5084 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5085 Module, accidentally introduced while backporting security fixes from the
5086 development branch and hindering the use of ECC in FIPS mode.
5090 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5092 * Client DoS due to large DH parameter
5094 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5095 malicious server can send a very large prime value to the client. This will
5096 cause the client to spend an unreasonably long period of time generating a
5097 key for this prime resulting in a hang until the client has finished. This
5098 could be exploited in a Denial Of Service attack.
5100 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5105 * Cache timing vulnerability in RSA Key Generation
5107 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5108 a cache timing side channel attack. An attacker with sufficient access to
5109 mount cache timing attacks during the RSA key generation process could
5110 recover the private key.
5112 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5113 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5118 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5119 parameter is no longer accepted, as it leads to a corrupt table. NULL
5120 pem_str is reserved for alias entries only.
5124 * Revert blinding in ECDSA sign and instead make problematic addition
5125 length-invariant. Switch even to fixed-length Montgomery multiplication.
5129 * Change generating and checking of primes so that the error rate of not
5130 being prime depends on the intended use based on the size of the input.
5131 For larger primes this will result in more rounds of Miller-Rabin.
5132 The maximal error rate for primes with more than 1080 bits is lowered
5135 *Kurt Roeckx, Annie Yousar*
5137 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5141 * Add blinding to ECDSA and DSA signatures to protect against side channel
5142 attacks discovered by Keegan Ryan (NCC Group).
5146 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5147 now allow empty (zero character) pass phrases.
5151 * Certificate time validation (X509_cmp_time) enforces stricter
5152 compliance with RFC 5280. Fractional seconds and timezone offsets
5153 are no longer allowed.
5157 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5159 * Constructed ASN.1 types with a recursive definition could exceed the stack
5161 Constructed ASN.1 types with a recursive definition (such as can be found
5162 in PKCS7) could eventually exceed the stack given malicious input with
5163 excessive recursion. This could result in a Denial Of Service attack. There
5164 are no such structures used within SSL/TLS that come from untrusted sources
5165 so this is considered safe.
5167 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5173 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5175 * Read/write after SSL object in error state
5177 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5178 mechanism. The intent was that if a fatal error occurred during a handshake
5179 then OpenSSL would move into the error state and would immediately fail if
5180 you attempted to continue the handshake. This works as designed for the
5181 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5182 SSL_connect()), however due to a bug it does not work correctly if
5183 SSL_read() or SSL_write() is called directly. In that scenario, if the
5184 handshake fails then a fatal error will be returned in the initial function
5185 call. If SSL_read()/SSL_write() is subsequently called by the application
5186 for the same SSL object then it will succeed and the data is passed without
5187 being decrypted/encrypted directly from the SSL/TLS record layer.
5189 In order to exploit this issue an application bug would have to be present
5190 that resulted in a call to SSL_read()/SSL_write() being issued after having
5191 already received a fatal error.
5193 This issue was reported to OpenSSL by David Benjamin (Google).
5198 * rsaz_1024_mul_avx2 overflow bug on x86_64
5200 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5201 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5202 Analysis suggests that attacks against RSA and DSA as a result of this
5203 defect would be very difficult to perform and are not believed likely.
5204 Attacks against DH1024 are considered just feasible, because most of the
5205 work necessary to deduce information about a private key may be performed
5206 offline. The amount of resources required for such an attack would be
5207 significant. However, for an attack on TLS to be meaningful, the server
5208 would have to share the DH1024 private key among multiple clients, which is
5209 no longer an option since CVE-2016-0701.
5211 This only affects processors that support the AVX2 but not ADX extensions
5212 like Intel Haswell (4th generation).
5214 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5215 was originally found via the OSS-Fuzz project.
5220 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5222 * bn_sqrx8x_internal carry bug on x86_64
5224 There is a carry propagating bug in the x86_64 Montgomery squaring
5225 procedure. No EC algorithms are affected. Analysis suggests that attacks
5226 against RSA and DSA as a result of this defect would be very difficult to
5227 perform and are not believed likely. Attacks against DH are considered just
5228 feasible (although very difficult) because most of the work necessary to
5229 deduce information about a private key may be performed offline. The amount
5230 of resources required for such an attack would be very significant and
5231 likely only accessible to a limited number of attackers. An attacker would
5232 additionally need online access to an unpatched system using the target
5233 private key in a scenario with persistent DH parameters and a private
5234 key that is shared between multiple clients.
5236 This only affects processors that support the BMI1, BMI2 and ADX extensions
5237 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5239 This issue was reported to OpenSSL by the OSS-Fuzz project.
5244 * Malformed X.509 IPAddressFamily could cause OOB read
5246 If an X.509 certificate has a malformed IPAddressFamily extension,
5247 OpenSSL could do a one-byte buffer overread. The most likely result
5248 would be an erroneous display of the certificate in text format.
5250 This issue was reported to OpenSSL by the OSS-Fuzz project.
5254 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5256 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5257 platform rather than 'mingw'.
5261 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5263 * Truncated packet could crash via OOB read
5265 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5266 cipher is being used, then a truncated packet can cause that host to
5267 perform an out-of-bounds read, usually resulting in a crash.
5269 This issue was reported to OpenSSL by Robert Święcki of Google.
5274 * BN_mod_exp may produce incorrect results on x86_64
5276 There is a carry propagating bug in the x86_64 Montgomery squaring
5277 procedure. No EC algorithms are affected. Analysis suggests that attacks
5278 against RSA and DSA as a result of this defect would be very difficult to
5279 perform and are not believed likely. Attacks against DH are considered just
5280 feasible (although very difficult) because most of the work necessary to
5281 deduce information about a private key may be performed offline. The amount
5282 of resources required for such an attack would be very significant and
5283 likely only accessible to a limited number of attackers. An attacker would
5284 additionally need online access to an unpatched system using the target
5285 private key in a scenario with persistent DH parameters and a private
5286 key that is shared between multiple clients. For example this can occur by
5287 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5288 similar to CVE-2015-3193 but must be treated as a separate problem.
5290 This issue was reported to OpenSSL by the OSS-Fuzz project.
5295 * Montgomery multiplication may produce incorrect results
5297 There is a carry propagating bug in the Broadwell-specific Montgomery
5298 multiplication procedure that handles input lengths divisible by, but
5299 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5300 and DH private keys are impossible. This is because the subroutine in
5301 question is not used in operations with the private key itself and an input
5302 of the attacker's direct choice. Otherwise the bug can manifest itself as
5303 transient authentication and key negotiation failures or reproducible
5304 erroneous outcome of public-key operations with specially crafted input.
5305 Among EC algorithms only Brainpool P-512 curves are affected and one
5306 presumably can attack ECDH key negotiation. Impact was not analyzed in
5307 detail, because pre-requisites for attack are considered unlikely. Namely
5308 multiple clients have to choose the curve in question and the server has to
5309 share the private key among them, neither of which is default behaviour.
5310 Even then only clients that chose the curve will be affected.
5312 This issue was publicly reported as transient failures and was not
5313 initially recognized as a security issue. Thanks to Richard Morgan for
5314 providing reproducible case.
5319 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5320 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5321 prevent issues where no progress is being made and the peer continually
5322 sends unrecognised record types, using up resources processing them.
5326 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5328 * Missing CRL sanity check
5330 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5331 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5332 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5334 This issue only affects the OpenSSL 1.0.2i
5339 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5341 * OCSP Status Request extension unbounded memory growth
5343 A malicious client can send an excessively large OCSP Status Request
5344 extension. If that client continually requests renegotiation, sending a
5345 large OCSP Status Request extension each time, then there will be unbounded
5346 memory growth on the server. This will eventually lead to a Denial Of
5347 Service attack through memory exhaustion. Servers with a default
5348 configuration are vulnerable even if they do not support OCSP. Builds using
5349 the "no-ocsp" build time option are not affected.
5351 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5356 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5359 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5365 * OOB write in MDC2_Update()
5367 An overflow can occur in MDC2_Update() either if called directly or
5368 through the EVP_DigestUpdate() function using MDC2. If an attacker
5369 is able to supply very large amounts of input data after a previous
5370 call to EVP_EncryptUpdate() with a partial block then a length check
5371 can overflow resulting in a heap corruption.
5373 The amount of data needed is comparable to SIZE_MAX which is impractical
5376 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5381 * Malformed SHA512 ticket DoS
5383 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5384 DoS attack where a malformed ticket will result in an OOB read which will
5387 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5388 a custom server callback and ticket lookup mechanism.
5390 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5395 * OOB write in BN_bn2dec()
5397 The function BN_bn2dec() does not check the return value of BN_div_word().
5398 This can cause an OOB write if an application uses this function with an
5399 overly large BIGNUM. This could be a problem if an overly large certificate
5400 or CRL is printed out from an untrusted source. TLS is not affected because
5401 record limits will reject an oversized certificate before it is parsed.
5403 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5408 * OOB read in TS_OBJ_print_bio()
5410 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5411 the total length the OID text representation would use and not the amount
5412 of data written. This will result in OOB reads when large OIDs are
5415 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5420 * Pointer arithmetic undefined behaviour
5422 Avoid some undefined pointer arithmetic
5424 A common idiom in the codebase is to check limits in the following manner:
5427 Where "p" points to some malloc'd data of SIZE bytes and
5430 "len" here could be from some externally supplied data (e.g. from a TLS
5433 The rules of C pointer arithmetic are such that "p + len" is only well
5434 defined where len <= SIZE. Therefore the above idiom is actually
5435 undefined behaviour.
5437 For example this could cause problems if some malloc implementation
5438 provides an address for "p" such that "p + len" actually overflows for
5439 values of len that are too big and therefore p + len < limit.
5441 This issue was reported to OpenSSL by Guido Vranken
5446 * Constant time flag not preserved in DSA signing
5448 Operations in the DSA signing algorithm should run in constant time in
5449 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5450 implementation means that a non-constant time codepath is followed for
5451 certain operations. This has been demonstrated through a cache-timing
5452 attack to be sufficient for an attacker to recover the private DSA key.
5454 This issue was reported by César Pereida (Aalto University), Billy Brumley
5455 (Tampere University of Technology), and Yuval Yarom (The University of
5456 Adelaide and NICTA).
5461 * DTLS buffered message DoS
5463 In a DTLS connection where handshake messages are delivered out-of-order
5464 those messages that OpenSSL is not yet ready to process will be buffered
5465 for later use. Under certain circumstances, a flaw in the logic means that
5466 those messages do not get removed from the buffer even though the handshake
5467 has been completed. An attacker could force up to approx. 15 messages to
5468 remain in the buffer when they are no longer required. These messages will
5469 be cleared when the DTLS connection is closed. The default maximum size for
5470 a message is 100k. Therefore, the attacker could force an additional 1500k
5471 to be consumed per connection. By opening many simultaneous connections an
5472 attacker could cause a DoS attack through memory exhaustion.
5474 This issue was reported to OpenSSL by Quan Luo.
5479 * DTLS replay protection DoS
5481 A flaw in the DTLS replay attack protection mechanism means that records
5482 that arrive for future epochs update the replay protection "window" before
5483 the MAC for the record has been validated. This could be exploited by an
5484 attacker by sending a record for the next epoch (which does not have to
5485 decrypt or have a valid MAC), with a very large sequence number. This means
5486 that all subsequent legitimate packets are dropped causing a denial of
5487 service for a specific DTLS connection.
5489 This issue was reported to OpenSSL by the OCAP audit team.
5494 * Certificate message OOB reads
5496 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5497 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5498 theoretical DoS risk but this has not been observed in practice on common
5501 The messages affected are client certificate, client certificate request
5502 and server certificate. As a result the attack can only be performed
5503 against a client or a server which enables client authentication.
5505 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5510 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5512 * Prevent padding oracle in AES-NI CBC MAC check
5514 A MITM attacker can use a padding oracle attack to decrypt traffic
5515 when the connection uses an AES CBC cipher and the server support
5518 This issue was introduced as part of the fix for Lucky 13 padding
5519 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5520 constant time by making sure that always the same bytes are read and
5521 compared against either the MAC or padding bytes. But it no longer
5522 checked that there was enough data to have both the MAC and padding
5525 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5529 * Fix EVP_EncodeUpdate overflow
5531 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5532 Base64 encoding of binary data. If an attacker is able to supply very large
5533 amounts of input data then a length check can overflow resulting in a heap
5536 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5537 the `PEM_write_bio*` family of functions. These are mainly used within the
5538 OpenSSL command line applications, so any application which processes data
5539 from an untrusted source and outputs it as a PEM file should be considered
5540 vulnerable to this issue. User applications that call these APIs directly
5541 with large amounts of untrusted data may also be vulnerable.
5543 This issue was reported by Guido Vranken.
5548 * Fix EVP_EncryptUpdate overflow
5550 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5551 is able to supply very large amounts of input data after a previous call to
5552 EVP_EncryptUpdate() with a partial block then a length check can overflow
5553 resulting in a heap corruption. Following an analysis of all OpenSSL
5554 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5555 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5556 the first called function after an EVP_EncryptInit(), and therefore that
5557 specific call must be safe. The second form is where the length passed to
5558 EVP_EncryptUpdate() can be seen from the code to be some small value and
5559 therefore there is no possibility of an overflow. Since all instances are
5560 one of these two forms, it is believed that there can be no overflows in
5561 internal code due to this problem. It should be noted that
5562 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5563 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5564 of these calls have also been analysed too and it is believed there are no
5565 instances in internal usage where an overflow could occur.
5567 This issue was reported by Guido Vranken.
5572 * Prevent ASN.1 BIO excessive memory allocation
5574 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5575 a short invalid encoding can cause allocation of large amounts of memory
5576 potentially consuming excessive resources or exhausting memory.
5578 Any application parsing untrusted data through d2i BIO functions is
5579 affected. The memory based functions such as d2i_X509() are *not* affected.
5580 Since the memory based functions are used by the TLS library, TLS
5581 applications are not affected.
5583 This issue was reported by Brian Carpenter.
5590 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5591 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5592 in arbitrary stack data being returned in the buffer.
5594 This issue was reported by Guido Vranken.
5599 * Modify behavior of ALPN to invoke callback after SNI/servername
5600 callback, such that updates to the SSL_CTX affect ALPN.
5604 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5609 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5610 methods are enabled and ssl2 is disabled the methods return NULL.
5614 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5616 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5617 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5618 provide any "EXPORT" or "LOW" strength ciphers.
5622 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5623 is by default disabled at build-time. Builds that are not configured with
5624 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5625 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5626 will need to explicitly call either of:
5628 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5630 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5632 as appropriate. Even if either of those is used, or the application
5633 explicitly uses the version-specific SSLv2_method() or its client and
5634 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5635 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5636 ciphers, and SSLv2 56-bit DES are no longer available.
5641 * Fix a double-free in DSA code
5643 A double free bug was discovered when OpenSSL parses malformed DSA private
5644 keys and could lead to a DoS attack or memory corruption for applications
5645 that receive DSA private keys from untrusted sources. This scenario is
5648 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5654 * Disable SRP fake user seed to address a server memory leak.
5656 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5658 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5659 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5660 was changed to ignore the "fake user" SRP seed, even if the seed
5663 Users should use SRP_VBASE_get1_by_user instead. Note that in
5664 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5665 also that even though configuring the SRP seed attempts to hide
5666 invalid usernames by continuing the handshake with fake
5667 credentials, this behaviour is not constant time and no strong
5668 guarantees are made that the handshake is indistinguishable from
5669 that of a valid user.
5674 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5676 In the BN_hex2bn function the number of hex digits is calculated using an
5677 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5678 large values of `i` this can result in `bn_expand` not allocating any
5679 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5680 field as NULL leading to a subsequent NULL ptr deref. For very large values
5681 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5682 In this case memory is allocated to the internal BIGNUM data field, but it
5683 is insufficiently sized leading to heap corruption. A similar issue exists
5684 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5685 is ever called by user applications with very large untrusted hex/dec data.
5686 This is anticipated to be a rare occurrence.
5688 All OpenSSL internal usage of these functions use data that is not expected
5689 to be untrusted, e.g. config file data or application command line
5690 arguments. If user developed applications generate config file data based
5691 on untrusted data then it is possible that this could also lead to security
5692 consequences. This is also anticipated to be rare.
5694 This issue was reported to OpenSSL by Guido Vranken.
5699 * Fix memory issues in `BIO_*printf` functions
5701 The internal `fmtstr` function used in processing a "%s" format string in
5702 the `BIO_*printf` functions could overflow while calculating the length of a
5703 string and cause an OOB read when printing very long strings.
5705 Additionally the internal `doapr_outch` function can attempt to write to an
5706 OOB memory location (at an offset from the NULL pointer) in the event of a
5707 memory allocation failure. In 1.0.2 and below this could be caused where
5708 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5709 could be in processing a very long "%s" format string. Memory leaks can
5712 The first issue may mask the second issue dependent on compiler behaviour.
5713 These problems could enable attacks where large amounts of untrusted data
5714 is passed to the `BIO_*printf` functions. If applications use these functions
5715 in this way then they could be vulnerable. OpenSSL itself uses these
5716 functions when printing out human-readable dumps of ASN.1 data. Therefore
5717 applications that print this data could be vulnerable if the data is from
5718 untrusted sources. OpenSSL command line applications could also be
5719 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5720 as command line arguments.
5722 Libssl is not considered directly vulnerable. Additionally certificates etc
5723 received via remote connections via libssl are also unlikely to be able to
5724 trigger these issues because of message size limits enforced within libssl.
5726 This issue was reported to OpenSSL Guido Vranken.
5731 * Side channel attack on modular exponentiation
5733 A side-channel attack was found which makes use of cache-bank conflicts on
5734 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5735 of RSA keys. The ability to exploit this issue is limited as it relies on
5736 an attacker who has control of code in a thread running on the same
5737 hyper-threaded core as the victim thread which is performing decryptions.
5739 This issue was reported to OpenSSL by Yuval Yarom, The University of
5740 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5741 Nadia Heninger, University of Pennsylvania with more information at
5742 <http://cachebleed.info>.
5747 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5748 if no keysize is specified with default_bits. This fixes an
5749 omission in an earlier change that changed all RSA/DSA key generation
5750 commands to use 2048 bits by default.
5754 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5756 * DH small subgroups
5758 Historically OpenSSL only ever generated DH parameters based on "safe"
5759 primes. More recently (in version 1.0.2) support was provided for
5760 generating X9.42 style parameter files such as those required for RFC 5114
5761 support. The primes used in such files may not be "safe". Where an
5762 application is using DH configured with parameters based on primes that are
5763 not "safe" then an attacker could use this fact to find a peer's private
5764 DH exponent. This attack requires that the attacker complete multiple
5765 handshakes in which the peer uses the same private DH exponent. For example
5766 this could be used to discover a TLS server's private DH exponent if it's
5767 reusing the private DH exponent or it's using a static DH ciphersuite.
5769 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5770 TLS. It is not on by default. If the option is not set then the server
5771 reuses the same private DH exponent for the life of the server process and
5772 would be vulnerable to this attack. It is believed that many popular
5773 applications do set this option and would therefore not be at risk.
5775 The fix for this issue adds an additional check where a "q" parameter is
5776 available (as is the case in X9.42 based parameters). This detects the
5777 only known attack, and is the only possible defense for static DH
5778 ciphersuites. This could have some performance impact.
5780 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5781 default and cannot be disabled. This could have some performance impact.
5783 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5788 * SSLv2 doesn't block disabled ciphers
5790 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5791 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5792 been disabled, provided that the SSLv2 protocol was not also disabled via
5795 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5796 and Sebastian Schinzel.
5801 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5803 * BN_mod_exp may produce incorrect results on x86_64
5805 There is a carry propagating bug in the x86_64 Montgomery squaring
5806 procedure. No EC algorithms are affected. Analysis suggests that attacks
5807 against RSA and DSA as a result of this defect would be very difficult to
5808 perform and are not believed likely. Attacks against DH are considered just
5809 feasible (although very difficult) because most of the work necessary to
5810 deduce information about a private key may be performed offline. The amount
5811 of resources required for such an attack would be very significant and
5812 likely only accessible to a limited number of attackers. An attacker would
5813 additionally need online access to an unpatched system using the target
5814 private key in a scenario with persistent DH parameters and a private
5815 key that is shared between multiple clients. For example this can occur by
5816 default in OpenSSL DHE based SSL/TLS ciphersuites.
5818 This issue was reported to OpenSSL by Hanno Böck.
5823 * Certificate verify crash with missing PSS parameter
5825 The signature verification routines will crash with a NULL pointer
5826 dereference if presented with an ASN.1 signature using the RSA PSS
5827 algorithm and absent mask generation function parameter. Since these
5828 routines are used to verify certificate signature algorithms this can be
5829 used to crash any certificate verification operation and exploited in a
5830 DoS attack. Any application which performs certificate verification is
5831 vulnerable including OpenSSL clients and servers which enable client
5834 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5839 * X509_ATTRIBUTE memory leak
5841 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5842 memory. This structure is used by the PKCS#7 and CMS routines so any
5843 application which reads PKCS#7 or CMS data from untrusted sources is
5844 affected. SSL/TLS is not affected.
5846 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5852 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5853 This changes the decoding behaviour for some invalid messages,
5854 though the change is mostly in the more lenient direction, and
5855 legacy behaviour is preserved as much as possible.
5859 * In DSA_generate_parameters_ex, if the provided seed is too short,
5862 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5864 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5866 * Alternate chains certificate forgery
5868 During certificate verification, OpenSSL will attempt to find an
5869 alternative certificate chain if the first attempt to build such a chain
5870 fails. An error in the implementation of this logic can mean that an
5871 attacker could cause certain checks on untrusted certificates to be
5872 bypassed, such as the CA flag, enabling them to use a valid leaf
5873 certificate to act as a CA and "issue" an invalid certificate.
5875 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5880 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5882 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5883 incompatibility in the handling of HMAC. The previous ABI has now been
5888 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5890 * Malformed ECParameters causes infinite loop
5892 When processing an ECParameters structure OpenSSL enters an infinite loop
5893 if the curve specified is over a specially malformed binary polynomial
5896 This can be used to perform denial of service against any
5897 system which processes public keys, certificate requests or
5898 certificates. This includes TLS clients and TLS servers with
5899 client authentication enabled.
5901 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5906 * Exploitable out-of-bounds read in X509_cmp_time
5908 X509_cmp_time does not properly check the length of the ASN1_TIME
5909 string and can read a few bytes out of bounds. In addition,
5910 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5913 An attacker can use this to craft malformed certificates and CRLs of
5914 various sizes and potentially cause a segmentation fault, resulting in
5915 a DoS on applications that verify certificates or CRLs. TLS clients
5916 that verify CRLs are affected. TLS clients and servers with client
5917 authentication enabled may be affected if they use custom verification
5920 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5921 independently by Hanno Böck.
5926 * PKCS7 crash with missing EnvelopedContent
5928 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5929 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5930 with missing content and trigger a NULL pointer dereference on parsing.
5932 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5933 structures from untrusted sources are affected. OpenSSL clients and
5934 servers are not affected.
5936 This issue was reported to OpenSSL by Michal Zalewski (Google).
5941 * CMS verify infinite loop with unknown hash function
5943 When verifying a signedData message the CMS code can enter an infinite loop
5944 if presented with an unknown hash function OID. This can be used to perform
5945 denial of service against any system which verifies signedData messages using
5947 This issue was reported to OpenSSL by Johannes Bauer.
5952 * Race condition handling NewSessionTicket
5954 If a NewSessionTicket is received by a multi-threaded client when attempting to
5955 reuse a previous ticket then a race condition can occur potentially leading to
5956 a double free of the ticket data.
5961 * Only support 256-bit or stronger elliptic curves with the
5962 'ecdh_auto' setting (server) or by default (client). Of supported
5963 curves, prefer P-256 (both).
5967 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5969 * ClientHello sigalgs DoS fix
5971 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5972 invalid signature algorithms extension a NULL pointer dereference will
5973 occur. This can be exploited in a DoS attack against the server.
5975 This issue was was reported to OpenSSL by David Ramos of Stanford
5979 *Stephen Henson and Matt Caswell*
5981 * Multiblock corrupted pointer fix
5983 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5984 feature only applies on 64 bit x86 architecture platforms that support AES
5985 NI instructions. A defect in the implementation of "multiblock" can cause
5986 OpenSSL's internal write buffer to become incorrectly set to NULL when
5987 using non-blocking IO. Typically, when the user application is using a
5988 socket BIO for writing, this will only result in a failed connection.
5989 However if some other BIO is used then it is likely that a segmentation
5990 fault will be triggered, thus enabling a potential DoS attack.
5992 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5997 * Segmentation fault in DTLSv1_listen fix
5999 The DTLSv1_listen function is intended to be stateless and processes the
6000 initial ClientHello from many peers. It is common for user code to loop
6001 over the call to DTLSv1_listen until a valid ClientHello is received with
6002 an associated cookie. A defect in the implementation of DTLSv1_listen means
6003 that state is preserved in the SSL object from one invocation to the next
6004 that can lead to a segmentation fault. Errors processing the initial
6005 ClientHello can trigger this scenario. An example of such an error could be
6006 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6009 This issue was reported to OpenSSL by Per Allansson.
6014 * Segmentation fault in ASN1_TYPE_cmp fix
6016 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6017 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6018 certificate signature algorithm consistency this can be used to crash any
6019 certificate verification operation and exploited in a DoS attack. Any
6020 application which performs certificate verification is vulnerable including
6021 OpenSSL clients and servers which enable client authentication.
6026 * Segmentation fault for invalid PSS parameters fix
6028 The signature verification routines will crash with a NULL pointer
6029 dereference if presented with an ASN.1 signature using the RSA PSS
6030 algorithm and invalid parameters. Since these routines are used to verify
6031 certificate signature algorithms this can be used to crash any
6032 certificate verification operation and exploited in a DoS attack. Any
6033 application which performs certificate verification is vulnerable including
6034 OpenSSL clients and servers which enable client authentication.
6036 This issue was was reported to OpenSSL by Brian Carpenter.
6041 * ASN.1 structure reuse memory corruption fix
6043 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6044 memory corruption via an invalid write. Such reuse is and has been
6045 strongly discouraged and is believed to be rare.
6047 Applications that parse structures containing CHOICE or ANY DEFINED BY
6048 components may be affected. Certificate parsing (d2i_X509 and related
6049 functions) are however not affected. OpenSSL clients and servers are
6055 * PKCS7 NULL pointer dereferences fix
6057 The PKCS#7 parsing code does not handle missing outer ContentInfo
6058 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6059 missing content and trigger a NULL pointer dereference on parsing.
6061 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6062 otherwise parse PKCS#7 structures from untrusted sources are
6063 affected. OpenSSL clients and servers are not affected.
6065 This issue was reported to OpenSSL by Michal Zalewski (Google).
6070 * DoS via reachable assert in SSLv2 servers fix
6072 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6073 servers that both support SSLv2 and enable export cipher suites by sending
6074 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6076 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6077 (OpenSSL development team).
6082 * Empty CKE with client auth and DHE fix
6084 If client auth is used then a server can seg fault in the event of a DHE
6085 ciphersuite being selected and a zero length ClientKeyExchange message
6086 being sent by the client. This could be exploited in a DoS attack.
6091 * Handshake with unseeded PRNG fix
6093 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6094 with an unseeded PRNG. The conditions are:
6095 - The client is on a platform where the PRNG has not been seeded
6096 automatically, and the user has not seeded manually
6097 - A protocol specific client method version has been used (i.e. not
6098 SSL_client_methodv23)
6099 - A ciphersuite is used that does not require additional random data from
6100 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6102 If the handshake succeeds then the client random that has been used will
6103 have been generated from a PRNG with insufficient entropy and therefore the
6104 output may be predictable.
6106 For example using the following command with an unseeded openssl will
6107 succeed on an unpatched platform:
6109 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6114 * Use After Free following d2i_ECPrivatekey error fix
6116 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6117 could cause a use after free condition. This, in turn, could cause a double
6118 free in several private key parsing functions (such as d2i_PrivateKey
6119 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6120 for applications that receive EC private keys from untrusted
6121 sources. This scenario is considered rare.
6123 This issue was discovered by the BoringSSL project and fixed in their
6129 * X509_to_X509_REQ NULL pointer deref fix
6131 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6132 the certificate key is invalid. This function is rarely used in practice.
6134 This issue was discovered by Brian Carpenter.
6139 * Removed the export ciphers from the DEFAULT ciphers
6143 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6145 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6146 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6147 So far those who have to target multiple platforms would compromise
6148 and argue that binary targeting say ARMv5 would still execute on
6149 ARMv8. "Universal" build resolves this compromise by providing
6150 near-optimal performance even on newer platforms.
6154 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6155 (other platforms pending).
6157 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6159 * Add support for the SignedCertificateTimestampList certificate and
6160 OCSP response extensions from RFC6962.
6164 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6165 for corner cases. (Certain input points at infinity could lead to
6166 bogus results, with non-infinity inputs mapped to infinity too.)
6170 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6171 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6172 common cases are optimized and there still is room for further
6173 improvements. Vector Permutation AES for Altivec is also added.
6177 * Add support for little-endian ppc64 Linux target.
6179 *Marcelo Cerri (IBM)*
6181 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6182 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6183 are optimized and there still is room for further improvements.
6184 Both 32- and 64-bit modes are supported.
6186 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6188 * Improved ARMv7 NEON support.
6192 * Support for SPARC Architecture 2011 crypto extensions, first
6193 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6194 SHA256/512, MD5, GHASH and modular exponentiation.
6196 *Andy Polyakov, David Miller*
6198 * Accelerated modular exponentiation for Intel processors, a.k.a.
6201 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6203 * Support for new and upcoming Intel processors, including AVX2,
6204 BMI and SHA ISA extensions. This includes additional "stitched"
6205 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6208 This work was sponsored by Intel Corp.
6212 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6213 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6214 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6218 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6219 this fixes a limitation in previous versions of OpenSSL.
6223 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6224 MGF1 digest and OAEP label.
6228 * Add EVP support for key wrapping algorithms, to avoid problems with
6229 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6230 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6231 algorithms and include tests cases.
6235 * Add functions to allocate and set the fields of an ECDSA_METHOD
6238 *Douglas E. Engert, Steve Henson*
6240 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6241 difference in days and seconds between two tm or ASN1_TIME structures.
6245 * Add -rev test option to s_server to just reverse order of characters
6246 received by client and send back to server. Also prints an abbreviated
6247 summary of the connection parameters.
6251 * New option -brief for s_client and s_server to print out a brief summary
6252 of connection parameters.
6256 * Add callbacks for arbitrary TLS extensions.
6258 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6260 * New option -crl_download in several openssl utilities to download CRLs
6261 from CRLDP extension in certificates.
6265 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6269 * New function X509_CRL_diff to generate a delta CRL from the difference
6270 of two full CRLs. Add support to "crl" utility.
6274 * New functions to set lookup_crls function and to retrieve
6275 X509_STORE from X509_STORE_CTX.
6279 * Print out deprecated issuer and subject unique ID fields in
6284 * Extend OCSP I/O functions so they can be used for simple general purpose
6285 HTTP as well as OCSP. New wrapper function which can be used to download
6286 CRLs using the OCSP API.
6290 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6294 * `SSL_CONF*` functions. These provide a common framework for application
6295 configuration using configuration files or command lines.
6299 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6300 message callback and prints the results. Needs compile time option
6301 "enable-ssl-trace". New options to s_client and s_server to enable
6306 * New ctrl and macro to retrieve supported points extensions.
6307 Print out extension in s_server and s_client.
6311 * New functions to retrieve certificate signature and signature
6316 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6321 * New Suite B modes for TLS code. These use and enforce the requirements
6322 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6323 only use Suite B curves. The Suite B modes can be set by using the
6324 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6328 * New chain verification flags for Suite B levels of security. Check
6329 algorithms are acceptable when flags are set in X509_verify_cert.
6333 * Make tls1_check_chain return a set of flags indicating checks passed
6334 by a certificate chain. Add additional tests to handle client
6335 certificates: checks for matching certificate type and issuer name
6340 * If an attempt is made to use a signature algorithm not in the peer
6341 preference list abort the handshake. If client has no suitable
6342 signature algorithms in response to a certificate request do not
6343 use the certificate.
6347 * If server EC tmp key is not in client preference list abort handshake.
6351 * Add support for certificate stores in CERT structure. This makes it
6352 possible to have different stores per SSL structure or one store in
6353 the parent SSL_CTX. Include distinct stores for certificate chain
6354 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6355 to build and store a certificate chain in CERT structure: returning
6356 an error if the chain cannot be built: this will allow applications
6357 to test if a chain is correctly configured.
6359 Note: if the CERT based stores are not set then the parent SSL_CTX
6360 store is used to retain compatibility with existing behaviour.
6364 * New function ssl_set_client_disabled to set a ciphersuite disabled
6365 mask based on the current session, check mask when sending client
6366 hello and checking the requested ciphersuite.
6370 * New ctrls to retrieve and set certificate types in a certificate
6371 request message. Print out received values in s_client. If certificate
6372 types is not set with custom values set sensible values based on
6373 supported signature algorithms.
6377 * Support for distinct client and server supported signature algorithms.
6381 * Add certificate callback. If set this is called whenever a certificate
6382 is required by client or server. An application can decide which
6383 certificate chain to present based on arbitrary criteria: for example
6384 supported signature algorithms. Add very simple example to s_server.
6385 This fixes many of the problems and restrictions of the existing client
6386 certificate callback: for example you can now clear an existing
6387 certificate and specify the whole chain.
6391 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6392 the certificate can be used for (if anything). Set valid_flags field
6393 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6394 to have similar checks in it.
6396 Add new "cert_flags" field to CERT structure and include a "strict mode".
6397 This enforces some TLS certificate requirements (such as only permitting
6398 certificate signature algorithms contained in the supported algorithms
6399 extension) which some implementations ignore: this option should be used
6400 with caution as it could cause interoperability issues.
6404 * Update and tidy signature algorithm extension processing. Work out
6405 shared signature algorithms based on preferences and peer algorithms
6406 and print them out in s_client and s_server. Abort handshake if no
6407 shared signature algorithms.
6411 * Add new functions to allow customised supported signature algorithms
6412 for SSL and SSL_CTX structures. Add options to s_client and s_server
6417 * New function SSL_certs_clear() to delete all references to certificates
6418 from an SSL structure. Before this once a certificate had been added
6419 it couldn't be removed.
6423 * Integrate hostname, email address and IP address checking with certificate
6424 verification. New verify options supporting checking in openssl utility.
6428 * Fixes and wildcard matching support to hostname and email checking
6429 functions. Add manual page.
6431 *Florian Weimer (Red Hat Product Security Team)*
6433 * New functions to check a hostname email or IP address against a
6434 certificate. Add options x509 utility to print results of checks against
6439 * Fix OCSP checking.
6441 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6443 * Initial experimental support for explicitly trusted non-root CAs.
6444 OpenSSL still tries to build a complete chain to a root but if an
6445 intermediate CA has a trust setting included that is used. The first
6446 setting is used: whether to trust (e.g., -addtrust option to the x509
6451 * Add -trusted_first option which attempts to find certificates in the
6452 trusted store even if an untrusted chain is also supplied.
6456 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6457 platform support for Linux and Android.
6461 * Support for linux-x32, ILP32 environment in x86_64 framework.
6465 * Experimental multi-implementation support for FIPS capable OpenSSL.
6466 When in FIPS mode the approved implementations are used as normal,
6467 when not in FIPS mode the internal unapproved versions are used instead.
6468 This means that the FIPS capable OpenSSL isn't forced to use the
6469 (often lower performance) FIPS implementations outside FIPS mode.
6473 * Transparently support X9.42 DH parameters when calling
6474 PEM_read_bio_DHparameters. This means existing applications can handle
6475 the new parameter format automatically.
6479 * Initial experimental support for X9.42 DH parameter format: mainly
6480 to support use of 'q' parameter for RFC5114 parameters.
6484 * Add DH parameters from RFC5114 including test data to dhtest.
6488 * Support for automatic EC temporary key parameter selection. If enabled
6489 the most preferred EC parameters are automatically used instead of
6490 hardcoded fixed parameters. Now a server just has to call:
6491 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6492 support ECDH and use the most appropriate parameters.
6496 * Enhance and tidy EC curve and point format TLS extension code. Use
6497 static structures instead of allocation if default values are used.
6498 New ctrls to set curves we wish to support and to retrieve shared curves.
6499 Print out shared curves in s_server. New options to s_server and s_client
6500 to set list of supported curves.
6504 * New ctrls to retrieve supported signature algorithms and
6505 supported curve values as an array of NIDs. Extend openssl utility
6506 to print out received values.
6510 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6511 between NIDs and the more common NIST names such as "P-256". Enhance
6512 ecparam utility and ECC method to recognise the NIST names for curves.
6516 * Enhance SSL/TLS certificate chain handling to support different
6517 chains for each certificate instead of one chain in the parent SSL_CTX.
6521 * Support for fixed DH ciphersuite client authentication: where both
6522 server and client use DH certificates with common parameters.
6526 * Support for fixed DH ciphersuites: those requiring DH server
6531 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6533 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6534 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6535 X509_CINF_get_signature were reverted post internal team review.
6540 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6542 * OCSP Status Request extension unbounded memory growth
6544 A malicious client can send an excessively large OCSP Status Request
6545 extension. If that client continually requests renegotiation, sending a
6546 large OCSP Status Request extension each time, then there will be unbounded
6547 memory growth on the server. This will eventually lead to a Denial Of
6548 Service attack through memory exhaustion. Servers with a default
6549 configuration are vulnerable even if they do not support OCSP. Builds using
6550 the "no-ocsp" build time option are not affected.
6552 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6557 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6560 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6566 * OOB write in MDC2_Update()
6568 An overflow can occur in MDC2_Update() either if called directly or
6569 through the EVP_DigestUpdate() function using MDC2. If an attacker
6570 is able to supply very large amounts of input data after a previous
6571 call to EVP_EncryptUpdate() with a partial block then a length check
6572 can overflow resulting in a heap corruption.
6574 The amount of data needed is comparable to SIZE_MAX which is impractical
6577 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6582 * Malformed SHA512 ticket DoS
6584 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6585 DoS attack where a malformed ticket will result in an OOB read which will
6588 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6589 a custom server callback and ticket lookup mechanism.
6591 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6596 * OOB write in BN_bn2dec()
6598 The function BN_bn2dec() does not check the return value of BN_div_word().
6599 This can cause an OOB write if an application uses this function with an
6600 overly large BIGNUM. This could be a problem if an overly large certificate
6601 or CRL is printed out from an untrusted source. TLS is not affected because
6602 record limits will reject an oversized certificate before it is parsed.
6604 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6609 * OOB read in TS_OBJ_print_bio()
6611 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6612 the total length the OID text representation would use and not the amount
6613 of data written. This will result in OOB reads when large OIDs are
6616 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6621 * Pointer arithmetic undefined behaviour
6623 Avoid some undefined pointer arithmetic
6625 A common idiom in the codebase is to check limits in the following manner:
6628 Where "p" points to some malloc'd data of SIZE bytes and
6631 "len" here could be from some externally supplied data (e.g. from a TLS
6634 The rules of C pointer arithmetic are such that "p + len" is only well
6635 defined where len <= SIZE. Therefore, the above idiom is actually
6636 undefined behaviour.
6638 For example this could cause problems if some malloc implementation
6639 provides an address for "p" such that "p + len" actually overflows for
6640 values of len that are too big and therefore p + len < limit.
6642 This issue was reported to OpenSSL by Guido Vranken
6647 * Constant time flag not preserved in DSA signing
6649 Operations in the DSA signing algorithm should run in constant time in
6650 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6651 implementation means that a non-constant time codepath is followed for
6652 certain operations. This has been demonstrated through a cache-timing
6653 attack to be sufficient for an attacker to recover the private DSA key.
6655 This issue was reported by César Pereida (Aalto University), Billy Brumley
6656 (Tampere University of Technology), and Yuval Yarom (The University of
6657 Adelaide and NICTA).
6662 * DTLS buffered message DoS
6664 In a DTLS connection where handshake messages are delivered out-of-order
6665 those messages that OpenSSL is not yet ready to process will be buffered
6666 for later use. Under certain circumstances, a flaw in the logic means that
6667 those messages do not get removed from the buffer even though the handshake
6668 has been completed. An attacker could force up to approx. 15 messages to
6669 remain in the buffer when they are no longer required. These messages will
6670 be cleared when the DTLS connection is closed. The default maximum size for
6671 a message is 100k. Therefore, the attacker could force an additional 1500k
6672 to be consumed per connection. By opening many simultaneous connections an
6673 attacker could cause a DoS attack through memory exhaustion.
6675 This issue was reported to OpenSSL by Quan Luo.
6680 * DTLS replay protection DoS
6682 A flaw in the DTLS replay attack protection mechanism means that records
6683 that arrive for future epochs update the replay protection "window" before
6684 the MAC for the record has been validated. This could be exploited by an
6685 attacker by sending a record for the next epoch (which does not have to
6686 decrypt or have a valid MAC), with a very large sequence number. This means
6687 that all subsequent legitimate packets are dropped causing a denial of
6688 service for a specific DTLS connection.
6690 This issue was reported to OpenSSL by the OCAP audit team.
6695 * Certificate message OOB reads
6697 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6698 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6699 theoretical DoS risk but this has not been observed in practice on common
6702 The messages affected are client certificate, client certificate request
6703 and server certificate. As a result the attack can only be performed
6704 against a client or a server which enables client authentication.
6706 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6711 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6713 * Prevent padding oracle in AES-NI CBC MAC check
6715 A MITM attacker can use a padding oracle attack to decrypt traffic
6716 when the connection uses an AES CBC cipher and the server support
6719 This issue was introduced as part of the fix for Lucky 13 padding
6720 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6721 constant time by making sure that always the same bytes are read and
6722 compared against either the MAC or padding bytes. But it no longer
6723 checked that there was enough data to have both the MAC and padding
6726 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6731 * Fix EVP_EncodeUpdate overflow
6733 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6734 Base64 encoding of binary data. If an attacker is able to supply very large
6735 amounts of input data then a length check can overflow resulting in a heap
6738 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6739 the `PEM_write_bio*` family of functions. These are mainly used within the
6740 OpenSSL command line applications, so any application which processes data
6741 from an untrusted source and outputs it as a PEM file should be considered
6742 vulnerable to this issue. User applications that call these APIs directly
6743 with large amounts of untrusted data may also be vulnerable.
6745 This issue was reported by Guido Vranken.
6750 * Fix EVP_EncryptUpdate overflow
6752 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6753 is able to supply very large amounts of input data after a previous call to
6754 EVP_EncryptUpdate() with a partial block then a length check can overflow
6755 resulting in a heap corruption. Following an analysis of all OpenSSL
6756 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6757 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6758 the first called function after an EVP_EncryptInit(), and therefore that
6759 specific call must be safe. The second form is where the length passed to
6760 EVP_EncryptUpdate() can be seen from the code to be some small value and
6761 therefore there is no possibility of an overflow. Since all instances are
6762 one of these two forms, it is believed that there can be no overflows in
6763 internal code due to this problem. It should be noted that
6764 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6765 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6766 of these calls have also been analysed too and it is believed there are no
6767 instances in internal usage where an overflow could occur.
6769 This issue was reported by Guido Vranken.
6774 * Prevent ASN.1 BIO excessive memory allocation
6776 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6777 a short invalid encoding can casuse allocation of large amounts of memory
6778 potentially consuming excessive resources or exhausting memory.
6780 Any application parsing untrusted data through d2i BIO functions is
6781 affected. The memory based functions such as d2i_X509() are *not* affected.
6782 Since the memory based functions are used by the TLS library, TLS
6783 applications are not affected.
6785 This issue was reported by Brian Carpenter.
6792 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6793 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6794 in arbitrary stack data being returned in the buffer.
6796 This issue was reported by Guido Vranken.
6801 * Modify behavior of ALPN to invoke callback after SNI/servername
6802 callback, such that updates to the SSL_CTX affect ALPN.
6806 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6811 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6812 methods are enabled and ssl2 is disabled the methods return NULL.
6816 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6818 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6819 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6820 provide any "EXPORT" or "LOW" strength ciphers.
6824 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6825 is by default disabled at build-time. Builds that are not configured with
6826 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6827 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6828 will need to explicitly call either of:
6830 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6832 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6834 as appropriate. Even if either of those is used, or the application
6835 explicitly uses the version-specific SSLv2_method() or its client and
6836 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6837 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6838 ciphers, and SSLv2 56-bit DES are no longer available.
6843 * Fix a double-free in DSA code
6845 A double free bug was discovered when OpenSSL parses malformed DSA private
6846 keys and could lead to a DoS attack or memory corruption for applications
6847 that receive DSA private keys from untrusted sources. This scenario is
6850 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6856 * Disable SRP fake user seed to address a server memory leak.
6858 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6860 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6861 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6862 was changed to ignore the "fake user" SRP seed, even if the seed
6865 Users should use SRP_VBASE_get1_by_user instead. Note that in
6866 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6867 also that even though configuring the SRP seed attempts to hide
6868 invalid usernames by continuing the handshake with fake
6869 credentials, this behaviour is not constant time and no strong
6870 guarantees are made that the handshake is indistinguishable from
6871 that of a valid user.
6876 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6878 In the BN_hex2bn function the number of hex digits is calculated using an
6879 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6880 large values of `i` this can result in `bn_expand` not allocating any
6881 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6882 field as NULL leading to a subsequent NULL ptr deref. For very large values
6883 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6884 In this case memory is allocated to the internal BIGNUM data field, but it
6885 is insufficiently sized leading to heap corruption. A similar issue exists
6886 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6887 is ever called by user applications with very large untrusted hex/dec data.
6888 This is anticipated to be a rare occurrence.
6890 All OpenSSL internal usage of these functions use data that is not expected
6891 to be untrusted, e.g. config file data or application command line
6892 arguments. If user developed applications generate config file data based
6893 on untrusted data then it is possible that this could also lead to security
6894 consequences. This is also anticipated to be rare.
6896 This issue was reported to OpenSSL by Guido Vranken.
6901 * Fix memory issues in `BIO_*printf` functions
6903 The internal `fmtstr` function used in processing a "%s" format string in
6904 the `BIO_*printf` functions could overflow while calculating the length of a
6905 string and cause an OOB read when printing very long strings.
6907 Additionally the internal `doapr_outch` function can attempt to write to an
6908 OOB memory location (at an offset from the NULL pointer) in the event of a
6909 memory allocation failure. In 1.0.2 and below this could be caused where
6910 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6911 could be in processing a very long "%s" format string. Memory leaks can
6914 The first issue may mask the second issue dependent on compiler behaviour.
6915 These problems could enable attacks where large amounts of untrusted data
6916 is passed to the `BIO_*printf` functions. If applications use these functions
6917 in this way then they could be vulnerable. OpenSSL itself uses these
6918 functions when printing out human-readable dumps of ASN.1 data. Therefore
6919 applications that print this data could be vulnerable if the data is from
6920 untrusted sources. OpenSSL command line applications could also be
6921 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6922 as command line arguments.
6924 Libssl is not considered directly vulnerable. Additionally certificates etc
6925 received via remote connections via libssl are also unlikely to be able to
6926 trigger these issues because of message size limits enforced within libssl.
6928 This issue was reported to OpenSSL Guido Vranken.
6933 * Side channel attack on modular exponentiation
6935 A side-channel attack was found which makes use of cache-bank conflicts on
6936 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6937 of RSA keys. The ability to exploit this issue is limited as it relies on
6938 an attacker who has control of code in a thread running on the same
6939 hyper-threaded core as the victim thread which is performing decryptions.
6941 This issue was reported to OpenSSL by Yuval Yarom, The University of
6942 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6943 Nadia Heninger, University of Pennsylvania with more information at
6944 <http://cachebleed.info>.
6949 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6950 if no keysize is specified with default_bits. This fixes an
6951 omission in an earlier change that changed all RSA/DSA key generation
6952 commands to use 2048 bits by default.
6956 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6958 * Protection for DH small subgroup attacks
6960 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6961 switched on by default and cannot be disabled. This could have some
6966 * SSLv2 doesn't block disabled ciphers
6968 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6969 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6970 been disabled, provided that the SSLv2 protocol was not also disabled via
6973 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6974 and Sebastian Schinzel.
6979 * Reject DH handshakes with parameters shorter than 1024 bits.
6983 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6985 * Certificate verify crash with missing PSS parameter
6987 The signature verification routines will crash with a NULL pointer
6988 dereference if presented with an ASN.1 signature using the RSA PSS
6989 algorithm and absent mask generation function parameter. Since these
6990 routines are used to verify certificate signature algorithms this can be
6991 used to crash any certificate verification operation and exploited in a
6992 DoS attack. Any application which performs certificate verification is
6993 vulnerable including OpenSSL clients and servers which enable client
6996 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7001 * X509_ATTRIBUTE memory leak
7003 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7004 memory. This structure is used by the PKCS#7 and CMS routines so any
7005 application which reads PKCS#7 or CMS data from untrusted sources is
7006 affected. SSL/TLS is not affected.
7008 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7014 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7015 This changes the decoding behaviour for some invalid messages,
7016 though the change is mostly in the more lenient direction, and
7017 legacy behaviour is preserved as much as possible.
7021 * In DSA_generate_parameters_ex, if the provided seed is too short,
7022 use a random seed, as already documented.
7024 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7026 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7028 * Alternate chains certificate forgery
7030 During certificate verfification, OpenSSL will attempt to find an
7031 alternative certificate chain if the first attempt to build such a chain
7032 fails. An error in the implementation of this logic can mean that an
7033 attacker could cause certain checks on untrusted certificates to be
7034 bypassed, such as the CA flag, enabling them to use a valid leaf
7035 certificate to act as a CA and "issue" an invalid certificate.
7037 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7043 * Race condition handling PSK identify hint
7045 If PSK identity hints are received by a multi-threaded client then
7046 the values are wrongly updated in the parent SSL_CTX structure. This can
7047 result in a race condition potentially leading to a double free of the
7053 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7055 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7056 incompatibility in the handling of HMAC. The previous ABI has now been
7059 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7061 * Malformed ECParameters causes infinite loop
7063 When processing an ECParameters structure OpenSSL enters an infinite loop
7064 if the curve specified is over a specially malformed binary polynomial
7067 This can be used to perform denial of service against any
7068 system which processes public keys, certificate requests or
7069 certificates. This includes TLS clients and TLS servers with
7070 client authentication enabled.
7072 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7077 * Exploitable out-of-bounds read in X509_cmp_time
7079 X509_cmp_time does not properly check the length of the ASN1_TIME
7080 string and can read a few bytes out of bounds. In addition,
7081 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7084 An attacker can use this to craft malformed certificates and CRLs of
7085 various sizes and potentially cause a segmentation fault, resulting in
7086 a DoS on applications that verify certificates or CRLs. TLS clients
7087 that verify CRLs are affected. TLS clients and servers with client
7088 authentication enabled may be affected if they use custom verification
7091 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7092 independently by Hanno Böck.
7097 * PKCS7 crash with missing EnvelopedContent
7099 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7100 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7101 with missing content and trigger a NULL pointer dereference on parsing.
7103 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7104 structures from untrusted sources are affected. OpenSSL clients and
7105 servers are not affected.
7107 This issue was reported to OpenSSL by Michal Zalewski (Google).
7112 * CMS verify infinite loop with unknown hash function
7114 When verifying a signedData message the CMS code can enter an infinite loop
7115 if presented with an unknown hash function OID. This can be used to perform
7116 denial of service against any system which verifies signedData messages using
7118 This issue was reported to OpenSSL by Johannes Bauer.
7123 * Race condition handling NewSessionTicket
7125 If a NewSessionTicket is received by a multi-threaded client when attempting to
7126 reuse a previous ticket then a race condition can occur potentially leading to
7127 a double free of the ticket data.
7132 * Reject DH handshakes with parameters shorter than 768 bits.
7134 *Kurt Roeckx and Emilia Kasper*
7136 * dhparam: generate 2048-bit parameters by default.
7138 *Kurt Roeckx and Emilia Kasper*
7140 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7142 * Segmentation fault in ASN1_TYPE_cmp fix
7144 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7145 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7146 certificate signature algorithm consistency this can be used to crash any
7147 certificate verification operation and exploited in a DoS attack. Any
7148 application which performs certificate verification is vulnerable including
7149 OpenSSL clients and servers which enable client authentication.
7154 * ASN.1 structure reuse memory corruption fix
7156 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7157 memory corruption via an invalid write. Such reuse is and has been
7158 strongly discouraged and is believed to be rare.
7160 Applications that parse structures containing CHOICE or ANY DEFINED BY
7161 components may be affected. Certificate parsing (d2i_X509 and related
7162 functions) are however not affected. OpenSSL clients and servers are
7168 * PKCS7 NULL pointer dereferences fix
7170 The PKCS#7 parsing code does not handle missing outer ContentInfo
7171 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7172 missing content and trigger a NULL pointer dereference on parsing.
7174 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7175 otherwise parse PKCS#7 structures from untrusted sources are
7176 affected. OpenSSL clients and servers are not affected.
7178 This issue was reported to OpenSSL by Michal Zalewski (Google).
7183 * DoS via reachable assert in SSLv2 servers fix
7185 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7186 servers that both support SSLv2 and enable export cipher suites by sending
7187 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7189 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7190 (OpenSSL development team).
7195 * Use After Free following d2i_ECPrivatekey error fix
7197 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7198 could cause a use after free condition. This, in turn, could cause a double
7199 free in several private key parsing functions (such as d2i_PrivateKey
7200 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7201 for applications that receive EC private keys from untrusted
7202 sources. This scenario is considered rare.
7204 This issue was discovered by the BoringSSL project and fixed in their
7210 * X509_to_X509_REQ NULL pointer deref fix
7212 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7213 the certificate key is invalid. This function is rarely used in practice.
7215 This issue was discovered by Brian Carpenter.
7220 * Removed the export ciphers from the DEFAULT ciphers
7224 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7226 * Build fixes for the Windows and OpenVMS platforms
7228 *Matt Caswell and Richard Levitte*
7230 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7232 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7233 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7234 dereference. This could lead to a Denial Of Service attack. Thanks to
7235 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7240 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7241 dtls1_buffer_record function under certain conditions. In particular this
7242 could occur if an attacker sent repeated DTLS records with the same
7243 sequence number but for the next epoch. The memory leak could be exploited
7244 by an attacker in a Denial of Service attack through memory exhaustion.
7245 Thanks to Chris Mueller for reporting this issue.
7250 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7251 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7252 method would be set to NULL which could later result in a NULL pointer
7253 dereference. Thanks to Frank Schmirler for reporting this issue.
7258 * Abort handshake if server key exchange message is omitted for ephemeral
7261 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7262 reporting this issue.
7267 * Remove non-export ephemeral RSA code on client and server. This code
7268 violated the TLS standard by allowing the use of temporary RSA keys in
7269 non-export ciphersuites and could be used by a server to effectively
7270 downgrade the RSA key length used to a value smaller than the server
7271 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7272 INRIA or reporting this issue.
7277 * Fixed issue where DH client certificates are accepted without verification.
7278 An OpenSSL server will accept a DH certificate for client authentication
7279 without the certificate verify message. This effectively allows a client to
7280 authenticate without the use of a private key. This only affects servers
7281 which trust a client certificate authority which issues certificates
7282 containing DH keys: these are extremely rare and hardly ever encountered.
7283 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7289 * Ensure that the session ID context of an SSL is updated when its
7290 SSL_CTX is updated via SSL_set_SSL_CTX.
7292 The session ID context is typically set from the parent SSL_CTX,
7293 and can vary with the CTX.
7297 * Fix various certificate fingerprint issues.
7299 By using non-DER or invalid encodings outside the signed portion of a
7300 certificate the fingerprint can be changed without breaking the signature.
7301 Although no details of the signed portion of the certificate can be changed
7302 this can cause problems with some applications: e.g. those using the
7303 certificate fingerprint for blacklists.
7305 1. Reject signatures with non zero unused bits.
7307 If the BIT STRING containing the signature has non zero unused bits reject
7308 the signature. All current signature algorithms require zero unused bits.
7310 2. Check certificate algorithm consistency.
7312 Check the AlgorithmIdentifier inside TBS matches the one in the
7313 certificate signature. NB: this will result in signature failure
7314 errors for some broken certificates.
7316 Thanks to Konrad Kraszewski from Google for reporting this issue.
7318 3. Check DSA/ECDSA signatures use DER.
7320 Re-encode DSA/ECDSA signatures and compare with the original received
7321 signature. Return an error if there is a mismatch.
7323 This will reject various cases including garbage after signature
7324 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7325 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7326 (negative or with leading zeroes).
7328 Further analysis was conducted and fixes were developed by Stephen Henson
7329 of the OpenSSL core team.
7335 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7336 results on some platforms, including x86_64. This bug occurs at random
7337 with a very low probability, and is not known to be exploitable in any
7338 way, though its exact impact is difficult to determine. Thanks to Pieter
7339 Wuille (Blockstream) who reported this issue and also suggested an initial
7340 fix. Further analysis was conducted by the OpenSSL development team and
7341 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7342 the OpenSSL core team.
7347 * Do not resume sessions on the server if the negotiated protocol
7348 version does not match the session's version. Resuming with a different
7349 version, while not strictly forbidden by the RFC, is of questionable
7350 sanity and breaks all known clients.
7352 *David Benjamin, Emilia Käsper*
7354 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7355 early CCS messages during renegotiation. (Note that because
7356 renegotiation is encrypted, this early CCS was not exploitable.)
7360 * Tighten client-side session ticket handling during renegotiation:
7361 ensure that the client only accepts a session ticket if the server sends
7362 the extension anew in the ServerHello. Previously, a TLS client would
7363 reuse the old extension state and thus accept a session ticket if one was
7364 announced in the initial ServerHello.
7366 Similarly, ensure that the client requires a session ticket if one
7367 was advertised in the ServerHello. Previously, a TLS client would
7368 ignore a missing NewSessionTicket message.
7372 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7376 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7377 sends a carefully crafted handshake message, to cause OpenSSL to fail
7378 to free up to 64k of memory causing a memory leak. This could be
7379 exploited in a Denial Of Service attack. This issue affects OpenSSL
7380 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7381 whether SRTP is used or configured. Implementations of OpenSSL that
7382 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7384 The fix was developed by the OpenSSL team.
7389 * Session Ticket Memory Leak.
7391 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7392 integrity of that ticket is first verified. In the event of a session
7393 ticket integrity check failing, OpenSSL will fail to free memory
7394 causing a memory leak. By sending a large number of invalid session
7395 tickets an attacker could exploit this issue in a Denial Of Service
7401 * Build option no-ssl3 is incomplete.
7403 When OpenSSL is configured with "no-ssl3" as a build option, servers
7404 could accept and complete an SSL 3.0 handshake, and clients could be
7405 configured to send them.
7408 *Akamai and the OpenSSL team*
7410 * Add support for TLS_FALLBACK_SCSV.
7411 Client applications doing fallback retries should call
7412 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7415 *Adam Langley, Bodo Moeller*
7417 * Add additional DigestInfo checks.
7419 Re-encode DigestInto in DER and check against the original when
7420 verifying RSA signature: this will reject any improperly encoded
7421 DigestInfo structures.
7423 Note: this is a precautionary measure and no attacks are currently known.
7427 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7429 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7430 SRP code can be overrun an internal buffer. Add sanity check that
7431 g, A, B < N to SRP code.
7433 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7434 Group for discovering this issue.
7439 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7440 TLS 1.0 instead of higher protocol versions when the ClientHello message
7441 is badly fragmented. This allows a man-in-the-middle attacker to force a
7442 downgrade to TLS 1.0 even if both the server and the client support a
7443 higher protocol version, by modifying the client's TLS records.
7445 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7446 researching this issue.
7451 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7452 to a denial of service attack. A malicious server can crash the client
7453 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7454 ciphersuite and sending carefully crafted handshake messages.
7456 Thanks to Felix Gröbert (Google) for discovering and researching this
7462 * By sending carefully crafted DTLS packets an attacker could cause openssl
7463 to leak memory. This can be exploited through a Denial of Service attack.
7464 Thanks to Adam Langley for discovering and researching this issue.
7469 * An attacker can force openssl to consume large amounts of memory whilst
7470 processing DTLS handshake messages. This can be exploited through a
7471 Denial of Service attack.
7472 Thanks to Adam Langley for discovering and researching this issue.
7477 * An attacker can force an error condition which causes openssl to crash
7478 whilst processing DTLS packets due to memory being freed twice. This
7479 can be exploited through a Denial of Service attack.
7480 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7486 * If a multithreaded client connects to a malicious server using a resumed
7487 session and the server sends an ec point format extension it could write
7488 up to 255 bytes to freed memory.
7490 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7496 * A malicious server can crash an OpenSSL client with a null pointer
7497 dereference (read) by specifying an SRP ciphersuite even though it was not
7498 properly negotiated with the client. This can be exploited through a
7499 Denial of Service attack.
7501 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7502 discovering and researching this issue.
7507 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7508 X509_name_oneline, X509_name_print_ex et al. to leak some information
7509 from the stack. Applications may be affected if they echo pretty printing
7510 output to the attacker.
7512 Thanks to Ivan Fratric (Google) for discovering this issue.
7515 *Emilia Käsper, and Steve Henson*
7517 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7518 for corner cases. (Certain input points at infinity could lead to
7519 bogus results, with non-infinity inputs mapped to infinity too.)
7523 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7525 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7526 handshake can force the use of weak keying material in OpenSSL
7527 SSL/TLS clients and servers.
7529 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7530 researching this issue. ([CVE-2014-0224])
7532 *KIKUCHI Masashi, Steve Henson*
7534 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7535 OpenSSL DTLS client the code can be made to recurse eventually crashing
7538 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7541 *Imre Rad, Steve Henson*
7543 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7544 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7545 client or server. This is potentially exploitable to run arbitrary
7546 code on a vulnerable client or server.
7548 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7550 *Jüri Aedla, Steve Henson*
7552 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7553 are subject to a denial of service attack.
7555 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7556 this issue. ([CVE-2014-3470])
7558 *Felix Gröbert, Ivan Fratric, Steve Henson*
7560 * Harmonize version and its documentation. -f flag is used to display
7563 *mancha <mancha1@zoho.com>*
7565 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7566 in i2d_ECPrivateKey.
7568 *mancha <mancha1@zoho.com>*
7570 * Fix some double frees. These are not thought to be exploitable.
7572 *mancha <mancha1@zoho.com>*
7574 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7576 * A missing bounds check in the handling of the TLS heartbeat extension
7577 can be used to reveal up to 64k of memory to a connected client or
7580 Thanks for Neel Mehta of Google Security for discovering this bug and to
7581 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7582 preparing the fix ([CVE-2014-0160])
7584 *Adam Langley, Bodo Moeller*
7586 * Fix for the attack described in the paper "Recovering OpenSSL
7587 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7588 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7589 <http://eprint.iacr.org/2014/140>
7591 Thanks to Yuval Yarom and Naomi Benger for discovering this
7592 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7594 *Yuval Yarom and Naomi Benger*
7596 * TLS pad extension: draft-agl-tls-padding-03
7598 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7599 TLS client Hello record length value would otherwise be > 255 and
7600 less that 512 pad with a dummy extension containing zeroes so it
7601 is at least 512 bytes long.
7603 *Adam Langley, Steve Henson*
7605 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7607 * Fix for TLS record tampering bug. A carefully crafted invalid
7608 handshake could crash OpenSSL with a NULL pointer exception.
7609 Thanks to Anton Johansson for reporting this issues.
7612 * Keep original DTLS digest and encryption contexts in retransmission
7613 structures so we can use the previous session parameters if they need
7614 to be resent. ([CVE-2013-6450])
7618 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7619 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7620 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7621 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7622 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7623 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7625 *Rob Stradling, Adam Langley*
7627 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7629 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7630 supporting platforms or when small records were transferred.
7632 *Andy Polyakov, Steve Henson*
7634 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7636 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7638 This addresses the flaw in CBC record processing discovered by
7639 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7640 at: <http://www.isg.rhul.ac.uk/tls/>
7642 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7643 Security Group at Royal Holloway, University of London
7644 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7645 Emilia Käsper for the initial patch.
7648 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7650 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7651 ciphersuites which can be exploited in a denial of service attack.
7652 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7653 and detecting this bug and to Wolfgang Ettlinger
7654 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7659 * Return an error when checking OCSP signatures when key is NULL.
7660 This fixes a DoS attack. ([CVE-2013-0166])
7664 * Make openssl verify return errors.
7666 *Chris Palmer <palmer@google.com> and Ben Laurie*
7668 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7669 the right response is stapled. Also change SSL_get_certificate()
7670 so it returns the certificate actually sent.
7671 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7673 *Rob Stradling <rob.stradling@comodo.com>*
7675 * Fix possible deadlock when decoding public keys.
7679 * Don't use TLS 1.0 record version number in initial client hello
7684 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7686 * Sanity check record length before skipping explicit IV in TLS
7687 1.2, 1.1 and DTLS to fix DoS attack.
7689 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7690 fuzzing as a service testing platform.
7695 * Initialise tkeylen properly when encrypting CMS messages.
7696 Thanks to Solar Designer of Openwall for reporting this issue.
7700 * In FIPS mode don't try to use composite ciphers as they are not
7705 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7707 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7708 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7709 mean any application compiled against OpenSSL 1.0.0 headers setting
7710 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7711 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7712 0x10000000L Any application which was previously compiled against
7713 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7714 will need to be recompiled as a result. Letting be results in
7715 inability to disable specifically TLS 1.1 and in client context,
7716 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7720 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7721 disable just protocol X, but all protocols above X *if* there are
7722 protocols *below* X still enabled. In more practical terms it means
7723 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7724 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7725 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7730 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7732 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7733 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7734 in CRYPTO_realloc_clean.
7736 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7737 issue and to Adam Langley <agl@chromium.org> for fixing it.
7740 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7742 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7746 * Workarounds for some broken servers that "hang" if a client hello
7747 record length exceeds 255 bytes.
7749 1. Do not use record version number > TLS 1.0 in initial client
7750 hello: some (but not all) hanging servers will now work.
7751 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7752 the number of ciphers sent in the client hello. This should be
7753 set to an even number, such as 50, for example by passing:
7754 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7755 Most broken servers should now work.
7756 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7757 TLS 1.2 client support entirely.
7761 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7765 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7767 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7768 STRING form instead of a DigestInfo.
7772 * The format used for MDC2 RSA signatures is inconsistent between EVP
7773 and the RSA_sign/RSA_verify functions. This was made more apparent when
7774 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7775 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7776 the correct format in RSA_verify so both forms transparently work.
7780 * Some servers which support TLS 1.0 can choke if we initially indicate
7781 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7782 encrypted premaster secret. As a workaround use the maximum permitted
7783 client version in client hello, this should keep such servers happy
7784 and still work with previous versions of OpenSSL.
7788 * Add support for TLS/DTLS heartbeats.
7790 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7792 * Add support for SCTP.
7794 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7796 * Improved PRNG seeding for VOS.
7798 *Paul Green <Paul.Green@stratus.com>*
7800 * Extensive assembler packs updates, most notably:
7802 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7803 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7804 - x86_64: bit-sliced AES implementation;
7805 - ARM: NEON support, contemporary platforms optimizations;
7806 - s390x: z196 support;
7807 - `*`: GHASH and GF(2^m) multiplication implementations;
7811 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7812 (removal of unnecessary code)
7814 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7816 * Add TLS key material exporter from RFC 5705.
7820 * Add DTLS-SRTP negotiation from RFC 5764.
7824 * Add Next Protocol Negotiation,
7825 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7826 disabled with a no-npn flag to config or Configure. Code donated
7829 *Adam Langley <agl@google.com> and Ben Laurie*
7831 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7832 NIST-P256, NIST-P521, with constant-time single point multiplication on
7833 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7834 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7835 Code made available under Apache License version 2.0.
7837 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7838 line to include this in your build of OpenSSL, and run "make depend" (or
7839 "make update"). This enables the following EC_METHODs:
7841 EC_GFp_nistp224_method()
7842 EC_GFp_nistp256_method()
7843 EC_GFp_nistp521_method()
7845 EC_GROUP_new_by_curve_name() will automatically use these (while
7846 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7849 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7851 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7852 all platforms. Move ssize_t definition from e_os.h to the public
7853 header file e_os2.h as it now appears in public header file cms.h
7857 * New -sigopt option to the ca, req and x509 utilities. Additional
7858 signature parameters can be passed using this option and in
7863 * Add RSA PSS signing function. This will generate and set the
7864 appropriate AlgorithmIdentifiers for PSS based on those in the
7865 corresponding EVP_MD_CTX structure. No application support yet.
7869 * Support for companion algorithm specific ASN1 signing routines.
7870 New function ASN1_item_sign_ctx() signs a pre-initialised
7871 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7872 the appropriate parameters.
7876 * Add new algorithm specific ASN1 verification initialisation function
7877 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7878 handling will be the same no matter what EVP_PKEY_METHOD is used.
7879 Add a PSS handler to support verification of PSS signatures: checked
7880 against a number of sample certificates.
7884 * Add signature printing for PSS. Add PSS OIDs.
7886 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7888 * Add algorithm specific signature printing. An individual ASN1 method
7889 can now print out signatures instead of the standard hex dump.
7891 More complex signatures (e.g. PSS) can print out more meaningful
7892 information. Include DSA version that prints out the signature
7897 * Password based recipient info support for CMS library: implementing
7902 * Split password based encryption into PBES2 and PBKDF2 functions. This
7903 neatly separates the code into cipher and PBE sections and is required
7904 for some algorithms that split PBES2 into separate pieces (such as
7905 password based CMS).
7909 * Session-handling fixes:
7910 - Fix handling of connections that are resuming with a session ID,
7911 but also support Session Tickets.
7912 - Fix a bug that suppressed issuing of a new ticket if the client
7913 presented a ticket with an expired session.
7914 - Try to set the ticket lifetime hint to something reasonable.
7915 - Make tickets shorter by excluding irrelevant information.
7916 - On the client side, don't ignore renewed tickets.
7918 *Adam Langley, Bodo Moeller (Google)*
7920 * Fix PSK session representation.
7924 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7926 This work was sponsored by Intel.
7930 * Add GCM support to TLS library. Some custom code is needed to split
7931 the IV between the fixed (from PRF) and explicit (from TLS record)
7932 portions. This adds all GCM ciphersuites supported by RFC5288 and
7933 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7934 add a special AESGCM string for GCM only.
7938 * Expand range of ctrls for AES GCM. Permit setting invocation
7939 field on decrypt and retrieval of invocation field only on encrypt.
7943 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7944 As required by RFC5289 these ciphersuites cannot be used if for
7945 versions of TLS earlier than 1.2.
7949 * For FIPS capable OpenSSL interpret a NULL default public key method
7950 as unset and return the appropriate default but do *not* set the default.
7951 This means we can return the appropriate method in applications that
7952 switch between FIPS and non-FIPS modes.
7956 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7957 ENGINE is used then we cannot handle that in the FIPS module so we
7958 keep original code iff non-FIPS operations are allowed.
7962 * Add -attime option to openssl utilities.
7964 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7966 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7970 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7971 FIPS EC methods unconditionally for now.
7975 * New build option no-ec2m to disable characteristic 2 code.
7979 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7980 all cases can be covered as some introduce binary incompatibilities.
7984 * Redirect RSA operations to FIPS module including keygen,
7985 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7989 * Add similar low-level API blocking to ciphers.
7993 * low-level digest APIs are not approved in FIPS mode: any attempt
7994 to use these will cause a fatal error. Applications that *really* want
7995 to use them can use the `private_*` version instead.
7999 * Redirect cipher operations to FIPS module for FIPS builds.
8003 * Redirect digest operations to FIPS module for FIPS builds.
8007 * Update build system to add "fips" flag which will link in fipscanister.o
8008 for static and shared library builds embedding a signature if needed.
8012 * Output TLS supported curves in preference order instead of numerical
8013 order. This is currently hardcoded for the highest order curves first.
8014 This should be configurable so applications can judge speed vs strength.
8018 * Add TLS v1.2 server support for client authentication.
8022 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8027 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8028 FIPS modules versions.
8032 * Add TLS v1.2 client side support for client authentication. Keep cache
8033 of handshake records longer as we don't know the hash algorithm to use
8034 until after the certificate request message is received.
8038 * Initial TLS v1.2 client support. Add a default signature algorithms
8039 extension including all the algorithms we support. Parse new signature
8040 format in client key exchange. Relax some ECC signing restrictions for
8041 TLS v1.2 as indicated in RFC5246.
8045 * Add server support for TLS v1.2 signature algorithms extension. Switch
8046 to new signature format when needed using client digest preference.
8047 All server ciphersuites should now work correctly in TLS v1.2. No client
8048 support yet and no support for client certificates.
8052 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8053 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8054 ciphersuites. At present only RSA key exchange ciphersuites work with
8055 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8056 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8057 and version checking.
8061 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8062 with this defined it will not be affected by any changes to ssl internal
8063 structures. Add several utility functions to allow openssl application
8064 to work with OPENSSL_NO_SSL_INTERN defined.
8068 * A long standing patch to add support for SRP from EdelWeb (Peter
8069 Sylvester and Christophe Renou) was integrated.
8070 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8071 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8074 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8078 * Permit abbreviated handshakes when renegotiating using the function
8079 SSL_renegotiate_abbreviated().
8081 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8083 * Add call to ENGINE_register_all_complete() to
8084 ENGINE_load_builtin_engines(), so some implementations get used
8085 automatically instead of needing explicit application support.
8089 * Add support for TLS key exporter as described in RFC5705.
8091 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8093 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8094 a few changes are required:
8096 Add SSL_OP_NO_TLSv1_1 flag.
8097 Add TLSv1_1 methods.
8098 Update version checking logic to handle version 1.1.
8099 Add explicit IV handling (ported from DTLS code).
8100 Add command line options to s_client/s_server.
8107 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8109 * X509_ATTRIBUTE memory leak
8111 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8112 memory. This structure is used by the PKCS#7 and CMS routines so any
8113 application which reads PKCS#7 or CMS data from untrusted sources is
8114 affected. SSL/TLS is not affected.
8116 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8122 * Race condition handling PSK identify hint
8124 If PSK identity hints are received by a multi-threaded client then
8125 the values are wrongly updated in the parent SSL_CTX structure. This can
8126 result in a race condition potentially leading to a double free of the
8132 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8134 * Malformed ECParameters causes infinite loop
8136 When processing an ECParameters structure OpenSSL enters an infinite loop
8137 if the curve specified is over a specially malformed binary polynomial
8140 This can be used to perform denial of service against any
8141 system which processes public keys, certificate requests or
8142 certificates. This includes TLS clients and TLS servers with
8143 client authentication enabled.
8145 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8150 * Exploitable out-of-bounds read in X509_cmp_time
8152 X509_cmp_time does not properly check the length of the ASN1_TIME
8153 string and can read a few bytes out of bounds. In addition,
8154 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8157 An attacker can use this to craft malformed certificates and CRLs of
8158 various sizes and potentially cause a segmentation fault, resulting in
8159 a DoS on applications that verify certificates or CRLs. TLS clients
8160 that verify CRLs are affected. TLS clients and servers with client
8161 authentication enabled may be affected if they use custom verification
8164 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8165 independently by Hanno Böck.
8170 * PKCS7 crash with missing EnvelopedContent
8172 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8173 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8174 with missing content and trigger a NULL pointer dereference on parsing.
8176 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8177 structures from untrusted sources are affected. OpenSSL clients and
8178 servers are not affected.
8180 This issue was reported to OpenSSL by Michal Zalewski (Google).
8185 * CMS verify infinite loop with unknown hash function
8187 When verifying a signedData message the CMS code can enter an infinite loop
8188 if presented with an unknown hash function OID. This can be used to perform
8189 denial of service against any system which verifies signedData messages using
8191 This issue was reported to OpenSSL by Johannes Bauer.
8196 * Race condition handling NewSessionTicket
8198 If a NewSessionTicket is received by a multi-threaded client when attempting to
8199 reuse a previous ticket then a race condition can occur potentially leading to
8200 a double free of the ticket data.
8205 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8207 * Segmentation fault in ASN1_TYPE_cmp fix
8209 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8210 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8211 certificate signature algorithm consistency this can be used to crash any
8212 certificate verification operation and exploited in a DoS attack. Any
8213 application which performs certificate verification is vulnerable including
8214 OpenSSL clients and servers which enable client authentication.
8219 * ASN.1 structure reuse memory corruption fix
8221 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8222 memory corruption via an invalid write. Such reuse is and has been
8223 strongly discouraged and is believed to be rare.
8225 Applications that parse structures containing CHOICE or ANY DEFINED BY
8226 components may be affected. Certificate parsing (d2i_X509 and related
8227 functions) are however not affected. OpenSSL clients and servers are
8233 * PKCS7 NULL pointer dereferences fix
8235 The PKCS#7 parsing code does not handle missing outer ContentInfo
8236 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8237 missing content and trigger a NULL pointer dereference on parsing.
8239 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8240 otherwise parse PKCS#7 structures from untrusted sources are
8241 affected. OpenSSL clients and servers are not affected.
8243 This issue was reported to OpenSSL by Michal Zalewski (Google).
8248 * DoS via reachable assert in SSLv2 servers fix
8250 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8251 servers that both support SSLv2 and enable export cipher suites by sending
8252 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8254 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8255 (OpenSSL development team).
8260 * Use After Free following d2i_ECPrivatekey error fix
8262 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8263 could cause a use after free condition. This, in turn, could cause a double
8264 free in several private key parsing functions (such as d2i_PrivateKey
8265 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8266 for applications that receive EC private keys from untrusted
8267 sources. This scenario is considered rare.
8269 This issue was discovered by the BoringSSL project and fixed in their
8275 * X509_to_X509_REQ NULL pointer deref fix
8277 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8278 the certificate key is invalid. This function is rarely used in practice.
8280 This issue was discovered by Brian Carpenter.
8285 * Removed the export ciphers from the DEFAULT ciphers
8289 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8291 * Build fixes for the Windows and OpenVMS platforms
8293 *Matt Caswell and Richard Levitte*
8295 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8297 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8298 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8299 dereference. This could lead to a Denial Of Service attack. Thanks to
8300 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8305 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8306 dtls1_buffer_record function under certain conditions. In particular this
8307 could occur if an attacker sent repeated DTLS records with the same
8308 sequence number but for the next epoch. The memory leak could be exploited
8309 by an attacker in a Denial of Service attack through memory exhaustion.
8310 Thanks to Chris Mueller for reporting this issue.
8315 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8316 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8317 method would be set to NULL which could later result in a NULL pointer
8318 dereference. Thanks to Frank Schmirler for reporting this issue.
8323 * Abort handshake if server key exchange message is omitted for ephemeral
8326 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8327 reporting this issue.
8332 * Remove non-export ephemeral RSA code on client and server. This code
8333 violated the TLS standard by allowing the use of temporary RSA keys in
8334 non-export ciphersuites and could be used by a server to effectively
8335 downgrade the RSA key length used to a value smaller than the server
8336 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8337 INRIA or reporting this issue.
8342 * Fixed issue where DH client certificates are accepted without verification.
8343 An OpenSSL server will accept a DH certificate for client authentication
8344 without the certificate verify message. This effectively allows a client to
8345 authenticate without the use of a private key. This only affects servers
8346 which trust a client certificate authority which issues certificates
8347 containing DH keys: these are extremely rare and hardly ever encountered.
8348 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8354 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8355 results on some platforms, including x86_64. This bug occurs at random
8356 with a very low probability, and is not known to be exploitable in any
8357 way, though its exact impact is difficult to determine. Thanks to Pieter
8358 Wuille (Blockstream) who reported this issue and also suggested an initial
8359 fix. Further analysis was conducted by the OpenSSL development team and
8360 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8361 the OpenSSL core team.
8366 * Fix various certificate fingerprint issues.
8368 By using non-DER or invalid encodings outside the signed portion of a
8369 certificate the fingerprint can be changed without breaking the signature.
8370 Although no details of the signed portion of the certificate can be changed
8371 this can cause problems with some applications: e.g. those using the
8372 certificate fingerprint for blacklists.
8374 1. Reject signatures with non zero unused bits.
8376 If the BIT STRING containing the signature has non zero unused bits reject
8377 the signature. All current signature algorithms require zero unused bits.
8379 2. Check certificate algorithm consistency.
8381 Check the AlgorithmIdentifier inside TBS matches the one in the
8382 certificate signature. NB: this will result in signature failure
8383 errors for some broken certificates.
8385 Thanks to Konrad Kraszewski from Google for reporting this issue.
8387 3. Check DSA/ECDSA signatures use DER.
8389 Reencode DSA/ECDSA signatures and compare with the original received
8390 signature. Return an error if there is a mismatch.
8392 This will reject various cases including garbage after signature
8393 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8394 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8395 (negative or with leading zeroes).
8397 Further analysis was conducted and fixes were developed by Stephen Henson
8398 of the OpenSSL core team.
8404 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8406 * Session Ticket Memory Leak.
8408 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8409 integrity of that ticket is first verified. In the event of a session
8410 ticket integrity check failing, OpenSSL will fail to free memory
8411 causing a memory leak. By sending a large number of invalid session
8412 tickets an attacker could exploit this issue in a Denial Of Service
8418 * Build option no-ssl3 is incomplete.
8420 When OpenSSL is configured with "no-ssl3" as a build option, servers
8421 could accept and complete an SSL 3.0 handshake, and clients could be
8422 configured to send them.
8425 *Akamai and the OpenSSL team*
8427 * Add support for TLS_FALLBACK_SCSV.
8428 Client applications doing fallback retries should call
8429 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8432 *Adam Langley, Bodo Moeller*
8434 * Add additional DigestInfo checks.
8436 Reencode DigestInto in DER and check against the original when
8437 verifying RSA signature: this will reject any improperly encoded
8438 DigestInfo structures.
8440 Note: this is a precautionary measure and no attacks are currently known.
8444 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8446 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8447 to a denial of service attack. A malicious server can crash the client
8448 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8449 ciphersuite and sending carefully crafted handshake messages.
8451 Thanks to Felix Gröbert (Google) for discovering and researching this
8457 * By sending carefully crafted DTLS packets an attacker could cause openssl
8458 to leak memory. This can be exploited through a Denial of Service attack.
8459 Thanks to Adam Langley for discovering and researching this issue.
8464 * An attacker can force openssl to consume large amounts of memory whilst
8465 processing DTLS handshake messages. This can be exploited through a
8466 Denial of Service attack.
8467 Thanks to Adam Langley for discovering and researching this issue.
8472 * An attacker can force an error condition which causes openssl to crash
8473 whilst processing DTLS packets due to memory being freed twice. This
8474 can be exploited through a Denial of Service attack.
8475 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8481 * If a multithreaded client connects to a malicious server using a resumed
8482 session and the server sends an ec point format extension it could write
8483 up to 255 bytes to freed memory.
8485 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8491 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8492 X509_name_oneline, X509_name_print_ex et al. to leak some information
8493 from the stack. Applications may be affected if they echo pretty printing
8494 output to the attacker.
8496 Thanks to Ivan Fratric (Google) for discovering this issue.
8499 *Emilia Käsper, and Steve Henson*
8501 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8502 for corner cases. (Certain input points at infinity could lead to
8503 bogus results, with non-infinity inputs mapped to infinity too.)
8507 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8509 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8510 handshake can force the use of weak keying material in OpenSSL
8511 SSL/TLS clients and servers.
8513 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8514 researching this issue. ([CVE-2014-0224])
8516 *KIKUCHI Masashi, Steve Henson*
8518 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8519 OpenSSL DTLS client the code can be made to recurse eventually crashing
8522 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8525 *Imre Rad, Steve Henson*
8527 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8528 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8529 client or server. This is potentially exploitable to run arbitrary
8530 code on a vulnerable client or server.
8532 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8534 *Jüri Aedla, Steve Henson*
8536 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8537 are subject to a denial of service attack.
8539 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8540 this issue. ([CVE-2014-3470])
8542 *Felix Gröbert, Ivan Fratric, Steve Henson*
8544 * Harmonize version and its documentation. -f flag is used to display
8547 *mancha <mancha1@zoho.com>*
8549 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8550 in i2d_ECPrivateKey.
8552 *mancha <mancha1@zoho.com>*
8554 * Fix some double frees. These are not thought to be exploitable.
8556 *mancha <mancha1@zoho.com>*
8558 * Fix for the attack described in the paper "Recovering OpenSSL
8559 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8560 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8561 <http://eprint.iacr.org/2014/140>
8563 Thanks to Yuval Yarom and Naomi Benger for discovering this
8564 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8566 *Yuval Yarom and Naomi Benger*
8568 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8570 * Keep original DTLS digest and encryption contexts in retransmission
8571 structures so we can use the previous session parameters if they need
8572 to be resent. ([CVE-2013-6450])
8576 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8577 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8578 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8579 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8580 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8581 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8583 *Rob Stradling, Adam Langley*
8585 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8587 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8589 This addresses the flaw in CBC record processing discovered by
8590 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8591 at: <http://www.isg.rhul.ac.uk/tls/>
8593 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8594 Security Group at Royal Holloway, University of London
8595 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8596 Emilia Käsper for the initial patch.
8599 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8601 * Return an error when checking OCSP signatures when key is NULL.
8602 This fixes a DoS attack. ([CVE-2013-0166])
8606 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8607 the right response is stapled. Also change SSL_get_certificate()
8608 so it returns the certificate actually sent.
8609 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8610 (This is a backport)
8612 *Rob Stradling <rob.stradling@comodo.com>*
8614 * Fix possible deadlock when decoding public keys.
8618 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8620 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8623 * Sanity check record length before skipping explicit IV in DTLS
8626 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8627 fuzzing as a service testing platform.
8632 * Initialise tkeylen properly when encrypting CMS messages.
8633 Thanks to Solar Designer of Openwall for reporting this issue.
8637 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8639 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8640 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8641 in CRYPTO_realloc_clean.
8643 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8644 issue and to Adam Langley <agl@chromium.org> for fixing it.
8647 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8649 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8651 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8652 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8653 content decryption and always return the same error. Note: this attack
8654 needs on average 2^20 messages so it only affects automated senders. The
8655 old behaviour can be re-enabled in the CMS code by setting the
8656 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8657 an MMA defence is not necessary.
8658 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8659 this issue. ([CVE-2012-0884])
8663 * Fix CVE-2011-4619: make sure we really are receiving a
8664 client hello before rejecting multiple SGC restarts. Thanks to
8665 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8669 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8671 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8672 Thanks to Antonio Martin, Enterprise Secure Access Research and
8673 Development, Cisco Systems, Inc. for discovering this bug and
8674 preparing a fix. ([CVE-2012-0050])
8678 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8680 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8681 of the Vaudenay padding oracle attack on CBC mode encryption
8682 which enables an efficient plaintext recovery attack against
8683 the OpenSSL implementation of DTLS. Their attack exploits timing
8684 differences arising during decryption processing. A research
8685 paper describing this attack can be found at:
8686 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8687 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8688 Security Group at Royal Holloway, University of London
8689 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8690 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8691 for preparing the fix. ([CVE-2011-4108])
8693 *Robin Seggelmann, Michael Tuexen*
8695 * Clear bytes used for block padding of SSL 3.0 records.
8698 *Adam Langley (Google)*
8700 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8701 Kadianakis <desnacked@gmail.com> for discovering this issue and
8702 Adam Langley for preparing the fix. ([CVE-2011-4619])
8704 *Adam Langley (Google)*
8706 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8708 *Andrey Kulikov <amdeich@gmail.com>*
8710 * Prevent malformed RFC3779 data triggering an assertion failure.
8711 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8712 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8714 *Rob Austein <sra@hactrn.net>*
8716 * Improved PRNG seeding for VOS.
8718 *Paul Green <Paul.Green@stratus.com>*
8720 * Fix ssl_ciph.c set-up race.
8722 *Adam Langley (Google)*
8724 * Fix spurious failures in ecdsatest.c.
8726 *Emilia Käsper (Google)*
8728 * Fix the BIO_f_buffer() implementation (which was mixing different
8729 interpretations of the `..._len` fields).
8731 *Adam Langley (Google)*
8733 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8734 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8735 threads won't reuse the same blinding coefficients.
8737 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8738 lock to call BN_BLINDING_invert_ex, and avoids one use of
8739 BN_BLINDING_update for each BN_BLINDING structure (previously,
8740 the last update always remained unused).
8742 *Emilia Käsper (Google)*
8744 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8746 *Bob Buckholz (Google)*
8748 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8750 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8751 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8753 *Kaspar Brand <ossl@velox.ch>*
8755 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8756 for multi-threaded use of ECDH. ([CVE-2011-3210])
8758 *Adam Langley (Google)*
8760 * Fix x509_name_ex_d2i memory leak on bad inputs.
8764 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8765 signature public key algorithm by using OID xref utilities instead.
8766 Before this you could only use some ECC ciphersuites with SHA1 only.
8770 * Add protection against ECDSA timing attacks as mentioned in the paper
8771 by Billy Bob Brumley and Nicola Tuveri, see:
8772 <http://eprint.iacr.org/2011/232.pdf>
8774 *Billy Bob Brumley and Nicola Tuveri*
8776 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8778 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8780 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8782 * Fix bug in string printing code: if *any* escaping is enabled we must
8783 escape the escape character (backslash) or the resulting string is
8788 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8790 * Disable code workaround for ancient and obsolete Netscape browsers
8791 and servers: an attacker can use it in a ciphersuite downgrade attack.
8792 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8796 * Fixed J-PAKE implementation error, originally discovered by
8797 Sebastien Martini, further info and confirmation from Stefan
8798 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8802 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8804 * Fix extension code to avoid race conditions which can result in a buffer
8805 overrun vulnerability: resumed sessions must not be modified as they can
8806 be shared by multiple threads. CVE-2010-3864
8810 * Fix WIN32 build system to correctly link an ENGINE directory into
8815 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8817 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8820 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8822 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8824 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8825 context. The operation can be customised via the ctrl mechanism in
8826 case ENGINEs want to include additional functionality.
8830 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8834 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8835 output hashes compatible with older versions of OpenSSL.
8837 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8839 * Fix compression algorithm handling: if resuming a session use the
8840 compression algorithm of the resumed session instead of determining
8841 it from client hello again. Don't allow server to change algorithm.
8845 * Add load_crls() function to commands tidying load_certs() too. Add option
8846 to verify utility to allow additional CRLs to be included.
8850 * Update OCSP request code to permit adding custom headers to the request:
8851 some responders need this.
8855 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8858 *Julia Lawall <julia@diku.dk>*
8860 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8861 needlessly dereferenced structures, used obsolete functions and
8862 didn't handle all updated verify codes correctly.
8866 * Disable MD2 in the default configuration.
8870 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8871 indicate the initial BIO being pushed or popped. This makes it possible
8872 to determine whether the BIO is the one explicitly called or as a result
8873 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8874 it handles reference counts correctly and doesn't zero out the I/O bio
8875 when it is not being explicitly popped. WARNING: applications which
8876 included workarounds for the old buggy behaviour will need to be modified
8877 or they could free up already freed BIOs.
8881 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8882 renaming to all platforms (within the 0.9.8 branch, this was
8883 done conditionally on Netware platforms to avoid a name clash).
8885 *Guenter <lists@gknw.net>*
8887 * Add ECDHE and PSK support to DTLS.
8889 *Michael Tuexen <tuexen@fh-muenster.de>*
8891 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8896 * Add "missing" function EVP_MD_flags() (without this the only way to
8897 retrieve a digest flags is by accessing the structure directly. Update
8898 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8899 or cipher is registered as in the "from" argument. Print out all
8900 registered digests in the dgst usage message instead of manually
8901 attempting to work them out.
8905 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8906 this allows the use of compression and extensions. Change default cipher
8907 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8908 by default unless an application cipher string requests it.
8912 * Alter match criteria in PKCS12_parse(). It used to try to use local
8913 key ids to find matching certificates and keys but some PKCS#12 files
8914 don't follow the (somewhat unwritten) rules and this strategy fails.
8915 Now just gather all certificates together and the first private key
8916 then look for the first certificate that matches the key.
8920 * Support use of registered digest and cipher names for dgst and cipher
8921 commands instead of having to add each one as a special case. So now
8928 openssl dgst -sha256 foo
8930 and this works for ENGINE based algorithms too.
8934 * Update Gost ENGINE to support parameter files.
8936 *Victor B. Wagner <vitus@cryptocom.ru>*
8938 * Support GeneralizedTime in ca utility.
8940 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8942 * Enhance the hash format used for certificate directory links. The new
8943 form uses the canonical encoding (meaning equivalent names will work
8944 even if they aren't identical) and uses SHA1 instead of MD5. This form
8945 is incompatible with the older format and as a result c_rehash should
8946 be used to rebuild symbolic links.
8950 * Make PKCS#8 the default write format for private keys, replacing the
8951 traditional format. This form is standardised, more secure and doesn't
8952 include an implicit MD5 dependency.
8956 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8957 committed to OpenSSL should pass this lot as a minimum.
8961 * Add session ticket override functionality for use by EAP-FAST.
8963 *Jouni Malinen <j@w1.fi>*
8965 * Modify HMAC functions to return a value. Since these can be implemented
8966 in an ENGINE errors can occur.
8970 * Type-checked OBJ_bsearch_ex.
8974 * Type-checked OBJ_bsearch. Also some constification necessitated
8975 by type-checking. Still to come: TXT_DB, bsearch(?),
8976 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8981 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8982 seconds to a tm structure directly, instead of going through OS
8983 specific date routines. This avoids any issues with OS routines such
8984 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8985 and X509_time_adj_ex() to cover the extended range. The existing
8986 X509_time_adj() is still usable and will no longer have any date issues.
8990 * Delta CRL support. New use deltas option which will attempt to locate
8991 and search any appropriate delta CRLs available.
8993 This work was sponsored by Google.
8997 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8998 code and add additional score elements. Validate alternate CRL paths
8999 as part of the CRL checking and indicate a new error "CRL path validation
9000 error" in this case. Applications wanting additional details can use
9001 the verify callback and check the new "parent" field. If this is not
9002 NULL CRL path validation is taking place. Existing applications won't
9003 see this because it requires extended CRL support which is off by
9006 This work was sponsored by Google.
9010 * Support for freshest CRL extension.
9012 This work was sponsored by Google.
9016 * Initial indirect CRL support. Currently only supported in the CRLs
9017 passed directly and not via lookup. Process certificate issuer
9018 CRL entry extension and lookup CRL entries by bother issuer name
9019 and serial number. Check and process CRL issuer entry in IDP extension.
9021 This work was sponsored by Google.
9025 * Add support for distinct certificate and CRL paths. The CRL issuer
9026 certificate is validated separately in this case. Only enabled if
9027 an extended CRL support flag is set: this flag will enable additional
9028 CRL functionality in future.
9030 This work was sponsored by Google.
9034 * Add support for policy mappings extension.
9036 This work was sponsored by Google.
9040 * Fixes to pathlength constraint, self issued certificate handling,
9041 policy processing to align with RFC3280 and PKITS tests.
9043 This work was sponsored by Google.
9047 * Support for name constraints certificate extension. DN, email, DNS
9048 and URI types are currently supported.
9050 This work was sponsored by Google.
9054 * To cater for systems that provide a pointer-based thread ID rather
9055 than numeric, deprecate the current numeric thread ID mechanism and
9056 replace it with a structure and associated callback type. This
9057 mechanism allows a numeric "hash" to be extracted from a thread ID in
9058 either case, and on platforms where pointers are larger than 'long',
9059 mixing is done to help ensure the numeric 'hash' is usable even if it
9060 can't be guaranteed unique. The default mechanism is to use "&errno"
9061 as a pointer-based thread ID to distinguish between threads.
9063 Applications that want to provide their own thread IDs should now use
9064 CRYPTO_THREADID_set_callback() to register a callback that will call
9065 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9067 Note that ERR_remove_state() is now deprecated, because it is tied
9068 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9069 to free the current thread's error state should be replaced by
9070 ERR_remove_thread_state(NULL).
9072 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9073 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9074 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9075 application was previously providing a numeric thread callback that
9076 was inappropriate for distinguishing threads, then uniqueness might
9077 have been obtained with &errno that happened immediately in the
9078 intermediate development versions of OpenSSL; this is no longer the
9079 case, the numeric thread callback will now override the automatic use
9082 *Geoff Thorpe, with help from Bodo Moeller*
9084 * Initial support for different CRL issuing certificates. This covers a
9085 simple case where the self issued certificates in the chain exist and
9086 the real CRL issuer is higher in the existing chain.
9088 This work was sponsored by Google.
9092 * Removed effectively defunct crypto/store from the build.
9096 * Revamp of STACK to provide stronger type-checking. Still to come:
9097 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9098 ASN1_STRING, CONF_VALUE.
9102 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9103 RAM on SSL connections. This option can save about 34k per idle SSL.
9107 * Revamp of LHASH to provide stronger type-checking. Still to come:
9108 STACK, TXT_DB, bsearch, qsort.
9112 * Initial support for Cryptographic Message Syntax (aka CMS) based
9113 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9114 support for data, signedData, compressedData, digestedData and
9115 encryptedData, envelopedData types included. Scripts to check against
9116 RFC4134 examples draft and interop and consistency checks of many
9117 content types and variants.
9121 * Add options to enc utility to support use of zlib compression BIO.
9125 * Extend mk1mf to support importing of options and assembly language
9126 files from Configure script, currently only included in VC-WIN32.
9127 The assembly language rules can now optionally generate the source
9128 files from the associated perl scripts.
9132 * Implement remaining functionality needed to support GOST ciphersuites.
9133 Interop testing has been performed using CryptoPro implementations.
9135 *Victor B. Wagner <vitus@cryptocom.ru>*
9137 * s390x assembler pack.
9141 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9146 * Implement Opaque PRF Input TLS extension as specified in
9147 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9148 official specification yet and no extension type assignment by
9149 IANA exists, this extension (for now) will have to be explicitly
9150 enabled when building OpenSSL by providing the extension number
9151 to use. For example, specify an option
9153 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9155 to the "config" or "Configure" script to enable the extension,
9156 assuming extension number 0x9527 (which is a completely arbitrary
9157 and unofficial assignment based on the MD5 hash of the Internet
9158 Draft). Note that by doing so, you potentially lose
9159 interoperability with other TLS implementations since these might
9160 be using the same extension number for other purposes.
9162 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9163 opaque PRF input value to use in the handshake. This will create
9164 an internal copy of the length-'len' string at 'src', and will
9165 return non-zero for success.
9167 To get more control and flexibility, provide a callback function
9170 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9171 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9175 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9178 Callback function 'cb' will be called in handshakes, and is
9179 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9180 Argument 'arg' is for application purposes (the value as given to
9181 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9182 be provided to the callback function). The callback function
9183 has to return non-zero to report success: usually 1 to use opaque
9184 PRF input just if possible, or 2 to enforce use of the opaque PRF
9185 input. In the latter case, the library will abort the handshake
9186 if opaque PRF input is not successfully negotiated.
9188 Arguments 'peerinput' and 'len' given to the callback function
9189 will always be NULL and 0 in the case of a client. A server will
9190 see the client's opaque PRF input through these variables if
9191 available (NULL and 0 otherwise). Note that if the server
9192 provides an opaque PRF input, the length must be the same as the
9193 length of the client's opaque PRF input.
9195 Note that the callback function will only be called when creating
9196 a new session (session resumption can resume whatever was
9197 previously negotiated), and will not be called in SSL 2.0
9198 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9199 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9200 for applications that need to enforce opaque PRF input.
9204 * Update ssl code to support digests other than SHA1+MD5 for handshake
9207 *Victor B. Wagner <vitus@cryptocom.ru>*
9209 * Add RFC4507 support to OpenSSL. This includes the corrections in
9210 RFC4507bis. The encrypted ticket format is an encrypted encoded
9211 SSL_SESSION structure, that way new session features are automatically
9214 If a client application caches session in an SSL_SESSION structure
9215 support is transparent because tickets are now stored in the encoded
9218 The SSL_CTX structure automatically generates keys for ticket
9219 protection in servers so again support should be possible
9220 with no application modification.
9222 If a client or server wishes to disable RFC4507 support then the option
9223 SSL_OP_NO_TICKET can be set.
9225 Add a TLS extension debugging callback to allow the contents of any client
9226 or server extensions to be examined.
9228 This work was sponsored by Google.
9232 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9233 OpenSSL should now compile cleanly on gcc 4.2
9235 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9237 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9238 support including streaming MAC support: this is required for GOST
9239 ciphersuite support.
9241 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9243 * Add option -stream to use PKCS#7 streaming in smime utility. New
9244 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9245 to output in BER and PEM format.
9249 * Experimental support for use of HMAC via EVP_PKEY interface. This
9250 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9251 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9252 ENGINE support for HMAC keys which are unextractable. New -mac and
9253 -macopt options to dgst utility.
9257 * New option -sigopt to dgst utility. Update dgst to use
9258 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9259 alternative signing parameters such as X9.31 or PSS in the dgst
9264 * Change ssl_cipher_apply_rule(), the internal function that does
9265 the work each time a ciphersuite string requests enabling
9266 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9267 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9268 the order of disabled ciphersuites such that those ciphersuites
9269 that most recently went from enabled to disabled not only stay
9270 in order with respect to each other, but also have higher priority
9271 than other disabled ciphersuites the next time ciphersuites are
9274 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9275 the same ciphersuites as with "HIGH" alone, but in a specific
9276 order where the PSK ciphersuites come first (since they are the
9277 most recently disabled ciphersuites when "HIGH" is parsed).
9279 Also, change ssl_create_cipher_list() (using this new
9280 functionality) such that between otherwise identical
9281 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9286 * Change ssl_create_cipher_list() so that it automatically
9287 arranges the ciphersuites in reasonable order before starting
9288 to process the rule string. Thus, the definition for "DEFAULT"
9289 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9290 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9291 This makes it much easier to arrive at a reasonable default order
9292 in applications for which anonymous ciphers are OK (meaning
9293 that you can't actually use DEFAULT).
9295 *Bodo Moeller; suggested by Victor Duchovni*
9297 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9298 processing) into multiple integers instead of setting
9299 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9300 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9301 (These masks as well as the individual bit definitions are hidden
9302 away into the non-exported interface ssl/ssl_locl.h, so this
9303 change to the definition of the SSL_CIPHER structure shouldn't
9304 affect applications.) This give us more bits for each of these
9305 categories, so there is no longer a need to coagulate AES128 and
9306 AES256 into a single algorithm bit, and to coagulate Camellia128
9307 and Camellia256 into a single algorithm bit, which has led to all
9310 Thus, among other things, the kludge introduced in 0.9.7m and
9311 0.9.8e for masking out AES256 independently of AES128 or masking
9312 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9314 With the change, we also introduce new ciphersuite aliases that
9315 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9320 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9321 Use the leftmost N bytes of the signature input if the input is
9322 larger than the prime q (with N being the size in bytes of q).
9326 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9327 it yet and it is largely untested.
9331 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9335 * Initial incomplete changes to avoid need for function casts in OpenSSL
9336 some compilers (gcc 4.2 and later) reject their use. Safestack is
9337 reimplemented. Update ASN1 to avoid use of legacy functions.
9341 * Win32/64 targets are linked with Winsock2.
9345 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9346 to external functions. This can be used to increase CRL handling
9347 efficiency especially when CRLs are very large by (for example) storing
9348 the CRL revoked certificates in a database.
9352 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9353 new CRLs added to a directory can be used. New command line option
9354 -verify_return_error to s_client and s_server. This causes real errors
9355 to be returned by the verify callback instead of carrying on no matter
9356 what. This reflects the way a "real world" verify callback would behave.
9360 * GOST engine, supporting several GOST algorithms and public key formats.
9361 Kindly donated by Cryptocom.
9365 * Partial support for Issuing Distribution Point CRL extension. CRLs
9366 partitioned by DP are handled but no indirect CRL or reason partitioning
9367 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9368 selected via a scoring technique which handles IDP and AKID in CRLs.
9372 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9373 will ultimately be used for all verify operations: this will remove the
9374 X509_STORE dependency on certificate verification and allow alternative
9375 lookup methods. X509_STORE based implementations of these two callbacks.
9379 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9380 Modify get_crl() to find a valid (unexpired) CRL if possible.
9384 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9385 this would be called X509_CRL_cmp() but that name is already used by
9386 a function that just compares CRL issuer names. Cache several CRL
9387 extensions in X509_CRL structure and cache CRLDP in X509.
9391 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9392 this maps equivalent X509_NAME structures into a consistent structure.
9393 Name comparison can then be performed rapidly using memcmp().
9397 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9402 * Allow digests to supply their own micalg string for S/MIME type using
9403 the ctrl EVP_MD_CTRL_MICALG.
9407 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9408 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9409 ctrl. It can then customise the structure before and/or after signing
9414 * New function OBJ_add_sigid() to allow application defined signature OIDs
9415 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9416 to free up any added signature OIDs.
9420 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9421 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9422 digest and cipher tables. New options added to openssl utility:
9423 list-message-digest-algorithms and list-cipher-algorithms.
9427 * Change the array representation of binary polynomials: the list
9428 of degrees of non-zero coefficients is now terminated with -1.
9429 Previously it was terminated with 0, which was also part of the
9430 value; thus, the array representation was not applicable to
9431 polynomials where t^0 has coefficient zero. This change makes
9432 the array representation useful in a more general context.
9436 * Various modifications and fixes to SSL/TLS cipher string
9437 handling. For ECC, the code now distinguishes between fixed ECDH
9438 with RSA certificates on the one hand and with ECDSA certificates
9439 on the other hand, since these are separate ciphersuites. The
9440 unused code for Fortezza ciphersuites has been removed.
9442 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9443 (not "ECDHE"). For consistency with the code for DH
9444 certificates, use of ECDH certificates is now considered ECDH
9445 authentication, not RSA or ECDSA authentication (the latter is
9446 merely the CA's signing algorithm and not actively used in the
9449 The temporary ciphersuite alias "ECCdraft" is no longer
9450 available, and ECC ciphersuites are no longer excluded from "ALL"
9451 and "DEFAULT". The following aliases now exist for RFC 4492
9452 ciphersuites, most of these by analogy with the DH case:
9454 kECDHr - ECDH cert, signed with RSA
9455 kECDHe - ECDH cert, signed with ECDSA
9456 kECDH - ECDH cert (signed with either RSA or ECDSA)
9457 kEECDH - ephemeral ECDH
9458 ECDH - ECDH cert or ephemeral ECDH
9464 AECDH - anonymous ECDH
9465 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9469 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9470 Use correct micalg parameters depending on digest(s) in signed message.
9474 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9475 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9479 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9480 an engine to register a method. Add ENGINE lookups for methods and
9481 functional reference processing.
9485 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9486 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9491 * New -resign option to smime utility. This adds one or more signers
9492 to an existing PKCS#7 signedData structure. Also -md option to use an
9493 alternative message digest algorithm for signing.
9497 * Tidy up PKCS#7 routines and add new functions to make it easier to
9498 create PKCS7 structures containing multiple signers. Update smime
9499 application to support multiple signers.
9503 * New -macalg option to pkcs12 utility to allow setting of an alternative
9508 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9509 Reorganize PBE internals to lookup from a static table using NIDs,
9510 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9511 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9512 PRF which will be automatically used with PBES2.
9516 * Replace the algorithm specific calls to generate keys in "req" with the
9521 * Update PKCS#7 enveloped data routines to use new API. This is now
9522 supported by any public key method supporting the encrypt operation. A
9523 ctrl is added to allow the public key algorithm to examine or modify
9524 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9529 * Add a ctrl to asn1 method to allow a public key algorithm to express
9530 a default digest type to use. In most cases this will be SHA1 but some
9531 algorithms (such as GOST) need to specify an alternative digest. The
9532 return value indicates how strong the preference is 1 means optional and
9533 2 is mandatory (that is it is the only supported type). Modify
9534 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9535 use the default md. Update openssl utilities to use the default digest
9536 type for signing if it is not explicitly indicated.
9540 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9541 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9542 signing method from the key type. This effectively removes the link
9543 between digests and public key types.
9547 * Add an OID cross reference table and utility functions. Its purpose is to
9548 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9549 rsaEncryption. This will allow some of the algorithm specific hackery
9550 needed to use the correct OID to be removed.
9554 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9555 structures for PKCS7_sign(). They are now set up by the relevant public
9560 * Add provisional EC pkey method with support for ECDSA and ECDH.
9564 * Add support for key derivation (agreement) in the API, DH method and
9569 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9570 public and private key formats. As a side effect these add additional
9571 command line functionality not previously available: DSA signatures can be
9572 generated and verified using pkeyutl and DH key support and generation in
9579 *Oliver Tappe <zooey@hirschkaefer.de>*
9581 * New make target "install_html_docs" installs HTML renditions of the
9584 *Oliver Tappe <zooey@hirschkaefer.de>*
9586 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9587 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9588 support key and parameter generation and add initial key generation
9589 functionality for RSA.
9593 * Add functions for main EVP_PKEY_method operations. The undocumented
9594 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9595 `EVP_PKEY_{encrypt,decrypt}_old`.
9599 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9600 key API, doesn't do much yet.
9604 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9605 public key algorithms. New option to openssl utility:
9606 "list-public-key-algorithms" to print out info.
9610 * Implement the Supported Elliptic Curves Extension for
9611 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9615 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9616 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9620 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9621 utilities such as rsa, dsa, dsaparam etc except they process any key
9626 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9627 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9628 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9633 * Initial support for pluggable public key ASN1.
9634 De-spaghettify the public key ASN1 handling. Move public and private
9635 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9636 algorithm specific handling to a single module within the relevant
9637 algorithm directory. Add functions to allow (near) opaque processing
9638 of public and private key structures.
9642 * Implement the Supported Point Formats Extension for
9643 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9647 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9648 for the psk identity [hint] and the psk callback functions to the
9649 SSL_SESSION, SSL and SSL_CTX structure.
9652 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9656 SSL_CTX_use_psk_identity_hint
9657 SSL_get_psk_identity_hint
9658 SSL_get_psk_identity
9659 SSL_use_psk_identity_hint
9661 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9663 * Add RFC 3161 compliant time stamp request creation, response generation
9664 and response verification functionality.
9666 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9668 * Add initial support for TLS extensions, specifically for the server_name
9669 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9670 have new members for a hostname. The SSL data structure has an
9671 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9672 stored in that context to allow for session resumption, even after the
9673 SSL has been switched to a new SSL_CTX in reaction to a client's
9674 server_name extension.
9676 New functions (subject to change):
9678 SSL_get_servername()
9679 SSL_get_servername_type()
9682 New CTRL codes and macros (subject to change):
9684 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9685 - SSL_CTX_set_tlsext_servername_callback()
9686 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9687 - SSL_CTX_set_tlsext_servername_arg()
9688 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9690 openssl s_client has a new '-servername ...' option.
9692 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9693 '-key2 ...', '-servername_fatal' (subject to change). This allows
9694 testing the HostName extension for a specific single hostname ('-cert'
9695 and '-key' remain fallbacks for handshakes without HostName
9696 negotiation). If the unrecognized_name alert has to be sent, this by
9697 default is a warning; it becomes fatal with the '-servername_fatal'
9700 *Peter Sylvester, Remy Allais, Christophe Renou*
9702 * Whirlpool hash implementation is added.
9706 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9707 bn(64,32). Because of instruction set limitations it doesn't have
9708 any negative impact on performance. This was done mostly in order
9709 to make it possible to share assembler modules, such as bn_mul_mont
9710 implementations, between 32- and 64-bit builds without hassle.
9714 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9715 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9720 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9721 dedicated Montgomery multiplication procedure, is introduced.
9722 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9723 "64-bit" performance on certain 32-bit targets.
9727 * New option SSL_OP_NO_COMP to disable use of compression selectively
9728 in SSL structures. New SSL ctrl to set maximum send fragment size.
9729 Save memory by setting the I/O buffer sizes dynamically instead of
9730 using the maximum available value.
9734 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9735 in addition to the text details.
9739 * Very, very preliminary EXPERIMENTAL support for printing of general
9740 ASN1 structures. This currently produces rather ugly output and doesn't
9741 handle several customised structures at all.
9745 * Integrated support for PVK file format and some related formats such
9746 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9747 these in the 'rsa' and 'dsa' utilities.
9751 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9755 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9756 place for the (very old) "NETSCAPE" format certificates which are now
9757 handled using new ASN1 code equivalents.
9761 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9762 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9763 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9767 * Modify CRL distribution points extension code to print out previously
9768 unsupported fields. Enhance extension setting code to allow setting of
9773 * Add print and set support for Issuing Distribution Point CRL extension.
9777 * Change 'Configure' script to enable Camellia by default.
9784 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9786 * When rejecting SSL/TLS records due to an incorrect version number, never
9787 update s->server with a new major version number. As of
9788 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9789 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9790 the previous behavior could result in a read attempt at NULL when
9791 receiving specific incorrect SSL/TLS records once record payload
9792 protection is active. ([CVE-2010-0740])
9794 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9796 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9797 could be crashed if the relevant tables were not present (e.g. chrooted).
9799 *Tomas Hoger <thoger@redhat.com>*
9801 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9803 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9805 *Martin Olsson, Neel Mehta*
9807 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9808 accommodate for stack sorting, always a write lock!).
9812 * On some versions of WIN32 Heap32Next is very slow. This can cause
9813 excessive delays in the RAND_poll(): over a minute. As a workaround
9814 include a time check in the inner Heap32Next loop too.
9818 * The code that handled flushing of data in SSL/TLS originally used the
9819 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9820 the problem outlined in PR#1949. The fix suggested there however can
9821 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9822 of Apache). So instead simplify the code to flush unconditionally.
9823 This should be fine since flushing with no data to flush is a no op.
9827 * Handle TLS versions 2.0 and later properly and correctly use the
9828 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9829 off ancient servers have a habit of sticking around for a while...
9833 * Modify compression code so it frees up structures without using the
9834 ex_data callbacks. This works around a problem where some applications
9835 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9836 restarting) then use compression (e.g. SSL with compression) later.
9837 This results in significant per-connection memory leaks and
9838 has caused some security issues including CVE-2008-1678 and
9843 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9844 change when encrypting or decrypting.
9848 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9849 connect and renegotiate with servers which do not support RI.
9850 Until RI is more widely deployed this option is enabled by default.
9854 * Add "missing" ssl ctrls to clear options and mode.
9858 * If client attempts to renegotiate and doesn't support RI respond with
9859 a no_renegotiation alert as required by RFC5746. Some renegotiating
9860 TLS clients will continue a connection gracefully when they receive
9861 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9862 waiting for a server hello which it will never receive. Now we treat a
9863 received no_renegotiation alert as a fatal error. This is because
9864 applications requesting a renegotiation might well expect it to succeed
9865 and would have no code in place to handle the server denying it so the
9866 only safe thing to do is to terminate the connection.
9870 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9871 peer supports secure renegotiation and 0 otherwise. Print out peer
9872 renegotiation support in s_client/s_server.
9876 * Replace the highly broken and deprecated SPKAC certification method with
9877 the updated NID creation version. This should correctly handle UTF8.
9881 * Implement RFC5746. Re-enable renegotiation but require the extension
9882 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9883 turns out to be a bad idea. It has been replaced by
9884 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9885 SSL_CTX_set_options(). This is really not recommended unless you
9886 know what you are doing.
9888 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9890 * Fixes to stateless session resumption handling. Use initial_ctx when
9891 issuing and attempting to decrypt tickets in case it has changed during
9892 servername handling. Use a non-zero length session ID when attempting
9893 stateless session resumption: this makes it possible to determine if
9894 a resumption has occurred immediately after receiving server hello
9895 (several places in OpenSSL subtly assume this) instead of later in
9900 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9901 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9902 fixes for a few places where the return code is not checked
9905 *Julia Lawall <julia@diku.dk>*
9907 * Add --strict-warnings option to Configure script to include devteam
9908 warnings in other configurations.
9912 * Add support for --libdir option and LIBDIR variable in makefiles. This
9913 makes it possible to install openssl libraries in locations which
9914 have names other than "lib", for example "/usr/lib64" which some
9917 *Steve Henson, based on patch from Jeremy Utley*
9919 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9920 X690 8.9.12 and can produce some misleading textual output of OIDs.
9922 *Steve Henson, reported by Dan Kaminsky*
9924 * Delete MD2 from algorithm tables. This follows the recommendation in
9925 several standards that it is not used in new applications due to
9926 several cryptographic weaknesses. For binary compatibility reasons
9927 the MD2 API is still compiled in by default.
9931 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9936 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9937 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9940 *Guenter <lists@gknw.net>*
9942 * Fix the server certificate chain building code to use X509_verify_cert(),
9943 it used to have an ad-hoc builder which was unable to cope with anything
9944 other than a simple chain.
9946 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9948 * Don't check self signed certificate signatures in X509_verify_cert()
9949 by default (a flag can override this): it just wastes time without
9950 adding any security. As a useful side effect self signed root CAs
9951 with non-FIPS digests are now usable in FIPS mode.
9955 * In dtls1_process_out_of_seq_message() the check if the current message
9956 is already buffered was missing. For every new message was memory
9957 allocated, allowing an attacker to perform an denial of service attack
9958 with sending out of seq handshake messages until there is no memory
9959 left. Additionally every future message was buffered, even if the
9960 sequence number made no sense and would be part of another handshake.
9961 So only messages with sequence numbers less than 10 in advance will be
9962 buffered. ([CVE-2009-1378])
9964 *Robin Seggelmann, discovered by Daniel Mentz*
9966 * Records are buffered if they arrive with a future epoch to be
9967 processed after finishing the corresponding handshake. There is
9968 currently no limitation to this buffer allowing an attacker to perform
9969 a DOS attack with sending records with future epochs until there is no
9970 memory left. This patch adds the pqueue_size() function to determine
9971 the size of a buffer and limits the record buffer to 100 entries.
9974 *Robin Seggelmann, discovered by Daniel Mentz*
9976 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9977 parent structure is freed. ([CVE-2009-1379])
9981 * Handle non-blocking I/O properly in SSL_shutdown() call.
9983 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9985 * Add `2.5.4.*` OIDs
9987 *Ilya O. <vrghost@gmail.com>*
9989 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9991 * Disable renegotiation completely - this fixes a severe security
9992 problem ([CVE-2009-3555]) at the cost of breaking all
9993 renegotiation. Renegotiation can be re-enabled by setting
9994 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9995 run-time. This is really not recommended unless you know what
10000 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10002 * Don't set val to NULL when freeing up structures, it is freed up by
10003 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10004 zeroing past the valid field. ([CVE-2009-0789])
10006 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10008 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10009 checked correctly. This would allow some invalid signed attributes to
10010 appear to verify correctly. ([CVE-2009-0591])
10012 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10014 * Reject UniversalString and BMPString types with invalid lengths. This
10015 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10016 a legal length. ([CVE-2009-0590])
10020 * Set S/MIME signing as the default purpose rather than setting it
10021 unconditionally. This allows applications to override it at the store
10026 * Permit restricted recursion of ASN1 strings. This is needed in practice
10027 to handle some structures.
10031 * Improve efficiency of mem_gets: don't search whole buffer each time
10034 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10036 * New -hex option for openssl rand.
10040 * Print out UTF8String and NumericString when parsing ASN1.
10044 * Support NumericString type for name components.
10048 * Allow CC in the environment to override the automatically chosen
10049 compiler. Note that nothing is done to ensure flags work with the
10054 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10056 * Properly check EVP_VerifyFinal() and similar return values
10059 *Ben Laurie, Bodo Moeller, Google Security Team*
10061 * Enable TLS extensions by default.
10065 * Allow the CHIL engine to be loaded, whether the application is
10066 multithreaded or not. (This does not release the developer from the
10067 obligation to set up the dynamic locking callbacks.)
10069 *Sander Temme <sander@temme.net>*
10071 * Use correct exit code if there is an error in dgst command.
10073 *Steve Henson; problem pointed out by Roland Dirlewanger*
10075 * Tweak Configure so that you need to say "experimental-jpake" to enable
10076 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10080 * Add experimental JPAKE support, including demo authentication in
10081 s_client and s_server.
10085 * Set the comparison function in v3_addr_canonize().
10087 *Rob Austein <sra@hactrn.net>*
10089 * Add support for XMPP STARTTLS in s_client.
10091 *Philip Paeps <philip@freebsd.org>*
10093 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10094 to ensure that even with this option, only ciphersuites in the
10095 server's preference list will be accepted. (Note that the option
10096 applies only when resuming a session, so the earlier behavior was
10097 just about the algorithm choice for symmetric cryptography.)
10101 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10103 * Fix NULL pointer dereference if a DTLS server received
10104 ChangeCipherSpec as first record ([CVE-2009-1386]).
10108 * Fix a state transition in s3_srvr.c and d1_srvr.c
10109 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10111 *Nagendra Modadugu*
10113 * The fix in 0.9.8c that supposedly got rid of unsafe
10114 double-checked locking was incomplete for RSA blinding,
10115 addressing just one layer of what turns out to have been
10116 doubly unsafe triple-checked locking.
10118 So now fix this for real by retiring the MONT_HELPER macro
10119 in crypto/rsa/rsa_eay.c.
10121 *Bodo Moeller; problem pointed out by Marius Schilder*
10123 * Various precautionary measures:
10125 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10127 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10128 (NB: This would require knowledge of the secret session ticket key
10129 to exploit, in which case you'd be SOL either way.)
10131 - Change bn_nist.c so that it will properly handle input BIGNUMs
10132 outside the expected range.
10134 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10137 *Neel Mehta, Bodo Moeller*
10139 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10140 the load fails. Useful for distros.
10142 *Ben Laurie and the FreeBSD team*
10144 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10148 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10152 * Expand ENGINE to support engine supplied SSL client certificate functions.
10154 This work was sponsored by Logica.
10158 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10159 keystores. Support for SSL/TLS client authentication too.
10160 Not compiled unless enable-capieng specified to Configure.
10162 This work was sponsored by Logica.
10166 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10167 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10168 attribute creation routines such as certificate requests and PKCS#12
10173 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10175 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10176 handshake which could lead to a client crash as found using the
10177 Codenomicon TLS test suite ([CVE-2008-1672])
10179 *Steve Henson, Mark Cox*
10181 * Fix double free in TLS server name extensions which could lead to
10182 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10186 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10188 Clear the error queue to ensure that error entries left from
10189 older function calls do not interfere with the correct operation.
10191 *Lutz Jaenicke, Erik de Castro Lopo*
10193 * Remove root CA certificates of commercial CAs:
10195 The OpenSSL project does not recommend any specific CA and does not
10196 have any policy with respect to including or excluding any CA.
10197 Therefore, it does not make any sense to ship an arbitrary selection
10198 of root CA certificates with the OpenSSL software.
10202 * RSA OAEP patches to fix two separate invalid memory reads.
10203 The first one involves inputs when 'lzero' is greater than
10204 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10205 before the beginning of from). The second one involves inputs where
10206 the 'db' section contains nothing but zeroes (there is a one-byte
10207 invalid read after the end of 'db').
10209 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10211 * Partial backport from 0.9.9-dev:
10213 Introduce bn_mul_mont (dedicated Montgomery multiplication
10214 procedure) as a candidate for BIGNUM assembler implementation.
10215 While 0.9.9-dev uses assembler for various architectures, only
10216 x86_64 is available by default here in the 0.9.8 branch, and
10217 32-bit x86 is available through a compile-time setting.
10219 To try the 32-bit x86 assembler implementation, use Configure
10220 option "enable-montasm" (which exists only for this backport).
10222 As "enable-montasm" for 32-bit x86 disclaims code stability
10223 anyway, in this constellation we activate additional code
10224 backported from 0.9.9-dev for further performance improvements,
10225 namely BN_from_montgomery_word. (To enable this otherwise,
10226 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10228 *Andy Polyakov (backport partially by Bodo Moeller)*
10230 * Add TLS session ticket callback. This allows an application to set
10231 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10232 values. This is useful for key rollover for example where several key
10233 sets may exist with different names.
10237 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10238 This was broken until now in 0.9.8 releases, such that the only way
10239 a registered ENGINE could be used (assuming it initialises
10240 successfully on the host) was to explicitly set it as the default
10241 for the relevant algorithms. This is in contradiction with 0.9.7
10242 behaviour and the documentation. With this fix, when an ENGINE is
10243 registered into a given algorithm's table of implementations, the
10244 'uptodate' flag is reset so that auto-discovery will be used next
10245 time a new context for that algorithm attempts to select an
10248 *Ian Lister (tweaked by Geoff Thorpe)*
10250 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10251 implementation in the following ways:
10253 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10256 Lack of BER streaming support means one pass streaming processing is
10257 only supported if data is detached: setting the streaming flag is
10258 ignored for embedded content.
10260 CMS support is disabled by default and must be explicitly enabled
10261 with the enable-cms configuration option.
10265 * Update the GMP engine glue to do direct copies between BIGNUM and
10266 mpz_t when openssl and GMP use the same limb size. Otherwise the
10267 existing "conversion via a text string export" trick is still used.
10269 *Paul Sheer <paulsheer@gmail.com>*
10271 * Zlib compression BIO. This is a filter BIO which compressed and
10272 uncompresses any data passed through it.
10276 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10277 RFC3394 compatible AES key wrapping.
10281 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10282 sets string data without copying. X509_ALGOR_set0() and
10283 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10284 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10285 from an X509_ATTRIBUTE structure optionally checking it occurs only
10286 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10291 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10292 to get the expected BN_FLG_CONSTTIME behavior.
10294 *Bodo Moeller (Google)*
10298 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10299 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10300 - added some more tests to do_tests.pl
10301 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10302 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10303 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10304 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10305 - various changes to netware.pl to enable gcc-cross builds on Win32
10307 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10308 - various changes to fix missing prototype warnings
10309 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10310 - added AES, WHIRLPOOL and CPUID assembler code to build files
10311 - added missing AES assembler make rules to mk1mf.pl
10312 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10314 *Guenter Knauf <eflash@gmx.net>*
10316 * Implement certificate status request TLS extension defined in RFC3546.
10317 A client can set the appropriate parameters and receive the encoded
10318 OCSP response via a callback. A server can query the supplied parameters
10319 and set the encoded OCSP response in the callback. Add simplified examples
10320 to s_client and s_server.
10324 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10326 * Fix various bugs:
10327 + Binary incompatibility of ssl_ctx_st structure
10328 + DTLS interoperation with non-compliant servers
10329 + Don't call get_session_cb() without proposed session
10330 + Fix ia64 assembler code
10332 *Andy Polyakov, Steve Henson*
10334 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10336 * DTLS Handshake overhaul. There were longstanding issues with
10337 OpenSSL DTLS implementation, which were making it impossible for
10338 RFC 4347 compliant client to communicate with OpenSSL server.
10339 Unfortunately just fixing these incompatibilities would "cut off"
10340 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10341 server keeps tolerating non RFC compliant syntax. The opposite is
10342 not true, 0.9.8f client can not communicate with earlier server.
10343 This update even addresses CVE-2007-4995.
10347 * Changes to avoid need for function casts in OpenSSL: some compilers
10348 (gcc 4.2 and later) reject their use.
10349 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10352 * Add RFC4507 support to OpenSSL. This includes the corrections in
10353 RFC4507bis. The encrypted ticket format is an encrypted encoded
10354 SSL_SESSION structure, that way new session features are automatically
10357 If a client application caches session in an SSL_SESSION structure
10358 support is transparent because tickets are now stored in the encoded
10361 The SSL_CTX structure automatically generates keys for ticket
10362 protection in servers so again support should be possible
10363 with no application modification.
10365 If a client or server wishes to disable RFC4507 support then the option
10366 SSL_OP_NO_TICKET can be set.
10368 Add a TLS extension debugging callback to allow the contents of any client
10369 or server extensions to be examined.
10371 This work was sponsored by Google.
10375 * Add initial support for TLS extensions, specifically for the server_name
10376 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10377 have new members for a hostname. The SSL data structure has an
10378 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10379 stored in that context to allow for session resumption, even after the
10380 SSL has been switched to a new SSL_CTX in reaction to a client's
10381 server_name extension.
10383 New functions (subject to change):
10385 SSL_get_servername()
10386 SSL_get_servername_type()
10389 New CTRL codes and macros (subject to change):
10391 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10392 - SSL_CTX_set_tlsext_servername_callback()
10393 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10394 - SSL_CTX_set_tlsext_servername_arg()
10395 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10397 openssl s_client has a new '-servername ...' option.
10399 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10400 '-key2 ...', '-servername_fatal' (subject to change). This allows
10401 testing the HostName extension for a specific single hostname ('-cert'
10402 and '-key' remain fallbacks for handshakes without HostName
10403 negotiation). If the unrecognized_name alert has to be sent, this by
10404 default is a warning; it becomes fatal with the '-servername_fatal'
10407 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10409 * Add AES and SSE2 assembly language support to VC++ build.
10413 * Mitigate attack on final subtraction in Montgomery reduction.
10417 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10418 (which previously caused an internal error).
10422 * Squeeze another 10% out of IGE mode when in != out.
10426 * AES IGE mode speedup.
10428 *Dean Gaudet (Google)*
10430 * Add the Korean symmetric 128-bit cipher SEED (see
10431 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10432 add SEED ciphersuites from RFC 4162:
10434 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10435 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10436 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10437 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10439 To minimize changes between patchlevels in the OpenSSL 0.9.8
10440 series, SEED remains excluded from compilation unless OpenSSL
10441 is configured with 'enable-seed'.
10443 *KISA, Bodo Moeller*
10445 * Mitigate branch prediction attacks, which can be practical if a
10446 single processor is shared, allowing a spy process to extract
10447 information. For detailed background information, see
10448 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10449 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10450 and Necessary Software Countermeasures"). The core of the change
10451 are new versions BN_div_no_branch() and
10452 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10453 respectively, which are slower, but avoid the security-relevant
10454 conditional branches. These are automatically called by BN_div()
10455 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10456 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10457 remove a conditional branch.
10459 BN_FLG_CONSTTIME is the new name for the previous
10460 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10461 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10462 in the exponent causes BN_mod_exp_mont() to use the alternative
10463 implementation in BN_mod_exp_mont_consttime().) The old name
10464 remains as a deprecated alias.
10466 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10467 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10468 constant-time implementations for more than just exponentiation.
10469 Here too the old name is kept as a deprecated alias.
10471 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10472 the BN_BLINDING structure gets an independent copy of the
10473 modulus. This means that the previous `BIGNUM *m` argument to
10474 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10475 essentially becomes `const BIGNUM *m`, although we can't actually
10476 change this in the header file before 0.9.9. It allows
10477 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10478 enable BN_FLG_CONSTTIME.
10480 *Matthew D Wood (Intel Corp)*
10482 * In the SSL/TLS server implementation, be strict about session ID
10483 context matching (which matters if an application uses a single
10484 external cache for different purposes). Previously,
10485 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10486 set. This did ensure strict client verification, but meant that,
10487 with applications using a single external cache for quite
10488 different requirements, clients could circumvent ciphersuite
10489 restrictions for a given session ID context by starting a session
10490 in a different context.
10494 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10495 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10496 authentication-only ciphersuites.
10500 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10501 not complete and could lead to a possible single byte overflow
10502 ([CVE-2007-5135]) [Ben Laurie]
10504 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10506 * Since AES128 and AES256 (and similarly Camellia128 and
10507 Camellia256) share a single mask bit in the logic of
10508 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10509 kludge to work properly if AES128 is available and AES256 isn't
10510 (or if Camellia128 is available and Camellia256 isn't).
10514 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10515 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10516 When a point or a seed is encoded in a BIT STRING, we need to
10517 prevent the removal of trailing zero bits to get the proper DER
10518 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10519 of a NamedBitList, for which trailing 0 bits need to be removed.)
10523 * Have SSL/TLS server implementation tolerate "mismatched" record
10524 protocol version while receiving ClientHello even if the
10525 ClientHello is fragmented. (The server can't insist on the
10526 particular protocol version it has chosen before the ServerHello
10527 message has informed the client about his choice.)
10531 * Add RFC 3779 support.
10533 *Rob Austein for ARIN, Ben Laurie*
10535 * Load error codes if they are not already present instead of using a
10536 static variable. This allows them to be cleanly unloaded and reloaded.
10537 Improve header file function name parsing.
10541 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10542 or CAPABILITY handshake as required by RFCs.
10544 *Goetz Babin-Ebell*
10546 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10548 * Introduce limits to prevent malicious keys being able to
10549 cause a denial of service. ([CVE-2006-2940])
10551 *Steve Henson, Bodo Moeller*
10553 * Fix ASN.1 parsing of certain invalid structures that can result
10554 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10556 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10557 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10559 * Fix SSL client code which could crash if connecting to a
10560 malicious SSLv2 server. ([CVE-2006-4343])
10562 *Tavis Ormandy and Will Drewry, Google Security Team*
10564 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10565 match only those. Before that, "AES256-SHA" would be interpreted
10566 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10567 the same strength classification in 0.9.7h) as we currently only
10568 have a single AES bit in the ciphersuite description bitmap.
10569 That change, however, also applied to ciphersuite strings such as
10570 "RC4-MD5" that intentionally matched multiple ciphersuites --
10571 namely, SSL 2.0 ciphersuites in addition to the more common ones
10572 from SSL 3.0/TLS 1.0.
10574 So we change the selection algorithm again: Naming an explicit
10575 ciphersuite selects this one ciphersuite, and any other similar
10576 ciphersuite (same bitmap) from *other* protocol versions.
10577 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10578 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10580 Since SSL 2.0 does not have any ciphersuites for which the
10581 128/256 bit distinction would be relevant, this works for now.
10582 The proper fix will be to use different bits for AES128 and
10583 AES256, which would have avoided the problems from the beginning;
10584 however, bits are scarce, so we can only do this in a new release
10585 (not just a patchlevel) when we can change the SSL_CIPHER
10586 definition to split the single 'unsigned long mask' bitmap into
10587 multiple values to extend the available space.
10591 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10593 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10594 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10596 * Add AES IGE and biIGE modes.
10600 * Change the Unix randomness entropy gathering to use poll() when
10601 possible instead of select(), since the latter has some
10602 undesirable limitations.
10604 *Darryl Miles via Richard Levitte and Bodo Moeller*
10606 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10607 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10608 cannot be implicitly activated as part of, e.g., the "AES" alias.
10609 However, please upgrade to OpenSSL 0.9.9[-dev] for
10610 non-experimental use of the ECC ciphersuites to get TLS extension
10611 support, which is required for curve and point format negotiation
10612 to avoid potential handshake problems.
10616 * Disable rogue ciphersuites:
10618 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10619 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10620 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10622 The latter two were purportedly from
10623 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10626 Also deactivate the remaining ciphersuites from
10627 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10628 unofficial, and the ID has long expired.
10632 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10633 dual-core machines) and other potential thread-safety issues.
10637 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10638 versions), which is now available for royalty-free use
10639 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10640 Also, add Camellia TLS ciphersuites from RFC 4132.
10642 To minimize changes between patchlevels in the OpenSSL 0.9.8
10643 series, Camellia remains excluded from compilation unless OpenSSL
10644 is configured with 'enable-camellia'.
10648 * Disable the padding bug check when compression is in use. The padding
10649 bug check assumes the first packet is of even length, this is not
10650 necessarily true if compression is enabled and can result in false
10651 positives causing handshake failure. The actual bug test is ancient
10652 code so it is hoped that implementations will either have fixed it by
10653 now or any which still have the bug do not support compression.
10657 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10659 * When applying a cipher rule check to see if string match is an explicit
10660 cipher suite and only match that one cipher suite if it is.
10664 * Link in manifests for VC++ if needed.
10666 *Austin Ziegler <halostatue@gmail.com>*
10668 * Update support for ECC-based TLS ciphersuites according to
10669 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10670 TLS extensions, which are supported starting with the 0.9.9
10671 branch, not in the OpenSSL 0.9.8 branch).
10675 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10676 opaque EVP_CIPHER_CTX handling.
10680 * Fixes and enhancements to zlib compression code. We now only use
10681 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10682 to conform with the standards mentioned here:
10683 <http://www.zlib.net/DLL_FAQ.txt>
10684 Static zlib linking now works on Windows and the new --with-zlib-include
10685 --with-zlib-lib options to Configure can be used to supply the location
10686 of the headers and library. Gracefully handle case where zlib library
10691 * Several fixes and enhancements to the OID generation code. The old code
10692 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10693 handle numbers larger than ULONG_MAX, truncated printing and had a
10694 non standard OBJ_obj2txt() behaviour.
10698 * Add support for building of engines under engine/ as shared libraries
10699 under VC++ build system.
10703 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10704 Hopefully, we will not see any false combination of paths any more.
10708 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10710 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10711 (part of SSL_OP_ALL). This option used to disable the
10712 countermeasure against man-in-the-middle protocol-version
10713 rollback in the SSL 2.0 server implementation, which is a bad
10714 idea. ([CVE-2005-2969])
10716 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10717 for Information Security, National Institute of Advanced Industrial
10718 Science and Technology [AIST], Japan)*
10720 * Add two function to clear and return the verify parameter flags.
10724 * Keep cipherlists sorted in the source instead of sorting them at
10725 runtime, thus removing the need for a lock.
10729 * Avoid some small subgroup attacks in Diffie-Hellman.
10731 *Nick Mathewson and Ben Laurie*
10733 * Add functions for well-known primes.
10737 * Extended Windows CE support.
10739 *Satoshi Nakamura and Andy Polyakov*
10741 * Initialize SSL_METHOD structures at compile time instead of during
10742 runtime, thus removing the need for a lock.
10746 * Make PKCS7_decrypt() work even if no certificate is supplied by
10747 attempting to decrypt each encrypted key in turn. Add support to
10752 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10754 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10757 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10761 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10762 key into the same file any more.
10766 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10770 * Add -utf8 command line and config file option to 'ca'.
10772 *Stefan <stf@udoma.org*
10774 * Removed the macro des_crypt(), as it seems to conflict with some
10775 libraries. Use DES_crypt().
10779 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10780 involves renaming the source and generated shared-libs for
10781 both. The engines will accept the corrected or legacy ids
10782 ('ncipher' and '4758_cca' respectively) when binding. NB,
10783 this only applies when building 'shared'.
10785 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10787 * Add attribute functions to EVP_PKEY structure. Modify
10788 PKCS12_create() to recognize a CSP name attribute and
10789 use it. Make -CSP option work again in pkcs12 utility.
10793 * Add new functionality to the bn blinding code:
10794 - automatic re-creation of the BN_BLINDING parameters after
10795 a fixed number of uses (currently 32)
10796 - add new function for parameter creation
10797 - introduce flags to control the update behaviour of the
10798 BN_BLINDING parameters
10799 - hide BN_BLINDING structure
10800 Add a second BN_BLINDING slot to the RSA structure to improve
10801 performance when a single RSA object is shared among several
10806 * Add support for DTLS.
10808 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10810 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10811 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10815 * Remove buggy and incomplete DH cert support from
10816 ssl/ssl_rsa.c and ssl/s3_both.c
10820 * Use SHA-1 instead of MD5 as the default digest algorithm for
10821 the `apps/openssl` commands.
10825 * Compile clean with "-Wall -Wmissing-prototypes
10826 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10827 DEBUG_SAFESTACK must also be set.
10831 * Change ./Configure so that certain algorithms can be disabled by default.
10832 The new counterpiece to "no-xxx" is "enable-xxx".
10834 The patented RC5 and MDC2 algorithms will now be disabled unless
10835 "enable-rc5" and "enable-mdc2", respectively, are specified.
10837 (IDEA remains enabled despite being patented. This is because IDEA
10838 is frequently required for interoperability, and there is no license
10839 fee for non-commercial use. As before, "no-idea" can be used to
10840 avoid this algorithm.)
10844 * Add processing of proxy certificates (see RFC 3820). This work was
10845 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10846 EGEE (Enabling Grids for E-science in Europe).
10850 * RC4 performance overhaul on modern architectures/implementations, such
10851 as Intel P4, IA-64 and AMD64.
10855 * New utility extract-section.pl. This can be used specify an alternative
10856 section number in a pod file instead of having to treat each file as
10857 a separate case in Makefile. This can be done by adding two lines to the
10860 =for comment openssl_section:XXX
10862 The blank line is mandatory.
10866 * New arguments -certform, -keyform and -pass for s_client and s_server
10867 to allow alternative format key and certificate files and passphrase
10872 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10873 update associated structures and add various utility functions.
10875 Add new policy related verify parameters, include policy checking in
10876 standard verify code. Enhance 'smime' application with extra parameters
10877 to support policy checking and print out.
10881 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10882 Nehemiah processors. These extensions support AES encryption in hardware
10883 as well as RNG (though RNG support is currently disabled).
10885 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10887 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10891 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10893 *Andy Polyakov and a number of other people*
10895 * Improved PowerPC platform support. Most notably BIGNUM assembler
10896 implementation contributed by IBM.
10898 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10900 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10901 exponent rather than 'unsigned long'. There is a corresponding change to
10902 the new 'rsa_keygen' element of the RSA_METHOD structure.
10904 *Jelte Jansen, Geoff Thorpe*
10906 * Functionality for creating the initial serial number file is now
10907 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10909 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10910 number file to 1, which is bound to cause problems. To avoid
10911 the problems while respecting compatibility between different 0.9.7
10912 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10913 CA.pl for serial number initialization. With the new release 0.9.8,
10914 we can fix the problem directly in the 'ca' utility.)
10918 * Reduced header interdependencies by declaring more opaque objects in
10919 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10920 give fewer recursive includes, which could break lazy source code - so
10921 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10922 developers should define this symbol when building and using openssl to
10923 ensure they track the recommended behaviour, interfaces, [etc], but
10924 backwards-compatible behaviour prevails when this isn't defined.
10928 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10932 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10933 This will generate a random key of the appropriate length based on the
10934 cipher context. The EVP_CIPHER can provide its own random key generation
10935 routine to support keys of a specific form. This is used in the des and
10936 3des routines to generate a key of the correct parity. Update S/MIME
10937 code to use new functions and hence generate correct parity DES keys.
10938 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10939 valid (weak or incorrect parity).
10943 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10944 as looking them up. This is useful when the verified structure may contain
10945 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10946 present unless the new PKCS7_NO_CRL flag is asserted.
10950 * Extend ASN1 oid configuration module. It now additionally accepts the
10953 shortName = some long name, 1.2.3.4
10957 * Reimplemented the BN_CTX implementation. There is now no more static
10958 limitation on the number of variables it can handle nor the depth of the
10959 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10960 information can now expand as required, and rather than having a single
10961 static array of bignums, BN_CTX now uses a linked-list of such arrays
10962 allowing it to expand on demand whilst maintaining the usefulness of
10963 BN_CTX's "bundling".
10967 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10968 to allow all RSA operations to function using a single BN_CTX.
10972 * Preliminary support for certificate policy evaluation and checking. This
10973 is initially intended to pass the tests outlined in "Conformance Testing
10974 of Relying Party Client Certificate Path Processing Logic" v1.07.
10978 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10979 remained unused and not that useful. A variety of other little bignum
10980 tweaks and fixes have also been made continuing on from the audit (see
10985 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10986 associated ASN1, EVP and SSL functions and old ASN1 macros.
10990 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10991 and this should never fail. So the return value from the use of
10992 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10993 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10997 * BN_CTX_get() should return zero-valued bignums, providing the same
10998 initialised value as BN_new().
11000 *Geoff Thorpe, suggested by Ulf Möller*
11002 * Support for inhibitAnyPolicy certificate extension.
11006 * An audit of the BIGNUM code is underway, for which debugging code is
11007 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11008 is considered valid when processing BIGNUMs, and causes execution to
11009 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11010 further steps are taken to deliberately pollute unused data in BIGNUM
11011 structures to try and expose faulty code further on. For now, openssl will
11012 (in its default mode of operation) continue to tolerate the inconsistent
11013 forms that it has tolerated in the past, but authors and packagers should
11014 consider trying openssl and their own applications when compiled with
11015 these debugging symbols defined. It will help highlight potential bugs in
11016 their own code, and will improve the test coverage for OpenSSL itself. At
11017 some point, these tighter rules will become openssl's default to improve
11018 maintainability, though the assert()s and other overheads will remain only
11019 in debugging configurations. See bn.h for more details.
11021 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11023 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11024 that can only be obtained through BN_CTX_new() (which implicitly
11025 initialises it). The presence of this function only made it possible
11026 to overwrite an existing structure (and cause memory leaks).
11030 * Because of the callback-based approach for implementing LHASH as a
11031 template type, lh_insert() adds opaque objects to hash-tables and
11032 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11033 to clean up those corresponding objects before destroying the hash table
11034 (and losing the object pointers). So some over-zealous constifications in
11035 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11036 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11037 prototyped to have "const" restrictions on the object pointers they are
11038 given (and so aren't required to cast them away any more).
11042 * The tmdiff.h API was so ugly and minimal that our own timing utility
11043 (speed) prefers to use its own implementation. The two implementations
11044 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11045 its object type properly exposed (MS_TM) instead of casting to/from
11046 `char *`. This may still change yet if someone realises MS_TM and
11048 aren't necessarily the greatest nomenclatures - but this is what was used
11049 internally to the implementation so I've used that for now.
11053 * Ensure that deprecated functions do not get compiled when
11054 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11055 the self-tests were still using deprecated key-generation functions so
11056 these have been updated also.
11060 * Reorganise PKCS#7 code to separate the digest location functionality
11061 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11062 New function PKCS7_set_digest() to set the digest type for PKCS#7
11063 digestedData type. Add additional code to correctly generate the
11064 digestedData type and add support for this type in PKCS7 initialization
11069 * New function PKCS7_set0_type_other() this initializes a PKCS7
11070 structure of type "other".
11074 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11075 sure the loop does correctly stop and breaking ("division by zero")
11076 modulus operations are not performed. The (pre-generated) prime
11077 table crypto/bn/bn_prime.h was already correct, but it could not be
11078 re-generated on some platforms because of the "division by zero"
11079 situation in the script.
11081 *Ralf S. Engelschall*
11083 * Update support for ECC-based TLS ciphersuites according to
11084 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11085 SHA-1 now is only used for "small" curves (where the
11086 representation of a field element takes up to 24 bytes); for
11087 larger curves, the field element resulting from ECDH is directly
11088 used as premaster secret.
11090 *Douglas Stebila (Sun Microsystems Laboratories)*
11092 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11093 curve secp160r1 to the tests.
11095 *Douglas Stebila (Sun Microsystems Laboratories)*
11097 * Add the possibility to load symbols globally with DSO.
11099 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11101 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11102 control of the error stack.
11106 * Add support for STORE in ENGINE.
11110 * Add the STORE type. The intention is to provide a common interface
11111 to certificate and key stores, be they simple file-based stores, or
11112 HSM-type store, or LDAP stores, or...
11113 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11117 * Add a generic structure called OPENSSL_ITEM. This can be used to
11118 pass a list of arguments to any function as well as provide a way
11119 for a function to pass data back to the caller.
11123 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11124 works like BUF_strdup() but can be used to duplicate a portion of
11125 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11130 * Add the function sk_find_ex() which works like sk_find(), but will
11131 return an index to an element even if an exact match couldn't be
11132 found. The index is guaranteed to point at the element where the
11133 searched-for key would be inserted to preserve sorting order.
11137 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11138 takes an extra flags argument for optional functionality. Currently,
11139 the following flags are defined:
11141 OBJ_BSEARCH_VALUE_ON_NOMATCH
11142 This one gets OBJ_bsearch_ex() to return a pointer to the first
11143 element where the comparing function returns a negative or zero
11146 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11147 This one gets OBJ_bsearch_ex() to return a pointer to the first
11148 element where the comparing function returns zero. This is useful
11149 if there are more than one element where the comparing function
11154 * Make it possible to create self-signed certificates with 'openssl ca'
11155 in such a way that the self-signed certificate becomes part of the
11156 CA database and uses the same mechanisms for serial number generation
11157 as all other certificate signing. The new flag '-selfsign' enables
11158 this functionality. Adapt CA.sh and CA.pl.in.
11162 * Add functionality to check the public key of a certificate request
11163 against a given private. This is useful to check that a certificate
11164 request can be signed by that key (self-signing).
11168 * Make it possible to have multiple active certificates with the same
11169 subject in the CA index file. This is done only if the keyword
11170 'unique_subject' is set to 'no' in the main CA section (default
11171 if 'CA_default') of the configuration file. The value is saved
11172 with the database itself in a separate index attribute file,
11173 named like the index file with '.attr' appended to the name.
11177 * Generate multi-valued AVAs using '+' notation in config files for
11182 * Support for nameConstraints certificate extension.
11186 * Support for policyConstraints certificate extension.
11190 * Support for policyMappings certificate extension.
11194 * Make sure the default DSA_METHOD implementation only uses its
11195 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11196 and change its own handlers to be NULL so as to remove unnecessary
11197 indirection. This lets alternative implementations fallback to the
11198 default implementation more easily.
11202 * Support for directoryName in GeneralName related extensions
11207 * Make it possible to link applications using Makefile.shared.
11208 Make that possible even when linking against static libraries!
11212 * Support for single pass processing for S/MIME signing. This now
11213 means that S/MIME signing can be done from a pipe, in addition
11214 cleartext signing (multipart/signed type) is effectively streaming
11215 and the signed data does not need to be all held in memory.
11217 This is done with a new flag PKCS7_STREAM. When this flag is set
11218 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11219 is done after the data is output (and digests calculated) in
11220 SMIME_write_PKCS7().
11224 * Add full support for -rpath/-R, both in shared libraries and
11225 applications, at least on the platforms where it's known how
11230 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11231 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11232 will now compute a table of multiples of the generator that
11233 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11234 faster (notably in the case of a single point multiplication,
11235 scalar * generator).
11237 *Nils Larsch, Bodo Moeller*
11239 * IPv6 support for certificate extensions. The various extensions
11240 which use the IP:a.b.c.d can now take IPv6 addresses using the
11241 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11246 * Added an ENGINE that implements RSA by performing private key
11247 exponentiations with the GMP library. The conversions to and from
11248 GMP's mpz_t format aren't optimised nor are any montgomery forms
11249 cached, and on x86 it appears OpenSSL's own performance has caught up.
11250 However there are likely to be other architectures where GMP could
11251 provide a boost. This ENGINE is not built in by default, but it can be
11252 specified at Configure time and should be accompanied by the necessary
11253 linker additions, eg;
11254 ./config -DOPENSSL_USE_GMP -lgmp
11258 * "openssl engine" will not display ENGINE/DSO load failure errors when
11259 testing availability of engines with "-t" - the old behaviour is
11260 produced by increasing the feature's verbosity with "-tt".
11264 * ECDSA routines: under certain error conditions uninitialized BN objects
11265 could be freed. Solution: make sure initialization is performed early
11266 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11271 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11272 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11273 software implementations. For DSA and DH, parameter generation can
11274 also be overridden by providing the appropriate method callbacks.
11278 * Change the "progress" mechanism used in key-generation and
11279 primality testing to functions that take a new BN_GENCB pointer in
11280 place of callback/argument pairs. The new API functions have `_ex`
11281 postfixes and the older functions are reimplemented as wrappers for
11282 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11283 declarations of the old functions to help (graceful) attempts to
11284 migrate to the new functions. Also, the new key-generation API
11285 functions operate on a caller-supplied key-structure and return
11286 success/failure rather than returning a key or NULL - this is to
11287 help make "keygen" another member function of RSA_METHOD etc.
11289 Example for using the new callback interface:
11291 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11292 void *my_arg = ...;
11295 BN_GENCB_set(&my_cb, my_callback, my_arg);
11297 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11298 /* For the meaning of a, b in calls to my_callback(), see the
11299 * documentation of the function that calls the callback.
11300 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11301 * my_callback should return 1 if it wants BN_is_prime_ex()
11302 * to continue, or 0 to stop.
11307 * Change the ZLIB compression method to be stateful, and make it
11308 available to TLS with the number defined in
11309 draft-ietf-tls-compression-04.txt.
11313 * Add the ASN.1 structures and functions for CertificatePair, which
11314 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11316 CertificatePair ::= SEQUENCE {
11317 forward [0] Certificate OPTIONAL,
11318 reverse [1] Certificate OPTIONAL,
11319 -- at least one of the pair shall be present -- }
11321 Also implement the PEM functions to read and write certificate
11322 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11324 This needed to be defined, mostly for the sake of the LDAP
11325 attribute crossCertificatePair, but may prove useful elsewhere as
11330 * Make it possible to inhibit symlinking of shared libraries in
11331 Makefile.shared, for Cygwin's sake.
11335 * Extend the BIGNUM API by creating a function
11336 void BN_set_negative(BIGNUM *a, int neg);
11337 and a macro that behave like
11338 int BN_is_negative(const BIGNUM *a);
11340 to avoid the need to access 'a->neg' directly in applications.
11344 * Implement fast modular reduction for pseudo-Mersenne primes
11345 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11346 EC_GROUP_new_curve_GFp() will now automatically use this
11349 *Nils Larsch <nla@trustcenter.de>*
11351 * Add new lock type (CRYPTO_LOCK_BN).
11355 * Change the ENGINE framework to automatically load engines
11356 dynamically from specific directories unless they could be
11357 found to already be built in or loaded. Move all the
11358 current engines except for the cryptodev one to a new
11359 directory engines/.
11360 The engines in engines/ are built as shared libraries if
11361 the "shared" options was given to ./Configure or ./config.
11362 Otherwise, they are inserted in libcrypto.a.
11363 /usr/local/ssl/engines is the default directory for dynamic
11364 engines, but that can be overridden at configure time through
11365 the usual use of --prefix and/or --openssldir, and at run
11366 time with the environment variable OPENSSL_ENGINES.
11368 *Geoff Thorpe and Richard Levitte*
11370 * Add Makefile.shared, a helper makefile to build shared
11371 libraries. Adapt Makefile.org.
11375 * Add version info to Win32 DLLs.
11377 *Peter 'Luna' Runestig" <peter@runestig.com>*
11379 * Add new 'medium level' PKCS#12 API. Certificates and keys
11380 can be added using this API to created arbitrary PKCS#12
11381 files while avoiding the low-level API.
11383 New options to PKCS12_create(), key or cert can be NULL and
11384 will then be omitted from the output file. The encryption
11385 algorithm NIDs can be set to -1 for no encryption, the mac
11386 iteration count can be set to 0 to omit the mac.
11388 Enhance pkcs12 utility by making the -nokeys and -nocerts
11389 options work when creating a PKCS#12 file. New option -nomac
11390 to omit the mac, NONE can be set for an encryption algorithm.
11391 New code is modified to use the enhanced PKCS12_create()
11392 instead of the low-level API.
11396 * Extend ASN1 encoder to support indefinite length constructed
11397 encoding. This can output sequences tags and octet strings in
11398 this form. Modify pk7_asn1.c to support indefinite length
11399 encoding. This is experimental and needs additional code to
11400 be useful, such as an ASN1 bio and some enhanced streaming
11403 Extend template encode functionality so that tagging is passed
11404 down to the template encoder.
11408 * Let 'openssl req' fail if an argument to '-newkey' is not
11409 recognized instead of using RSA as a default.
11413 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11414 As these are not official, they are not included in "ALL";
11415 the "ECCdraft" ciphersuite group alias can be used to select them.
11417 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11419 * Add ECDH engine support.
11421 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11423 * Add ECDH in new directory crypto/ecdh/.
11425 *Douglas Stebila (Sun Microsystems Laboratories)*
11427 * Let BN_rand_range() abort with an error after 100 iterations
11428 without success (which indicates a broken PRNG).
11432 * Change BN_mod_sqrt() so that it verifies that the input value
11433 is really the square of the return value. (Previously,
11434 BN_mod_sqrt would show GIGO behaviour.)
11438 * Add named elliptic curves over binary fields from X9.62, SECG,
11439 and WAP/WTLS; add OIDs that were still missing.
11441 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11443 * Extend the EC library for elliptic curves over binary fields
11444 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11447 EC_GF2m_simple_method
11451 EC_GROUP_new_curve_GF2m
11452 EC_GROUP_set_curve_GF2m
11453 EC_GROUP_get_curve_GF2m
11454 EC_POINT_set_affine_coordinates_GF2m
11455 EC_POINT_get_affine_coordinates_GF2m
11456 EC_POINT_set_compressed_coordinates_GF2m
11458 Point compression for binary fields is disabled by default for
11459 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11462 As binary polynomials are represented as BIGNUMs, various members
11463 of the EC_GROUP and EC_POINT data structures can be shared
11464 between the implementations for prime fields and binary fields;
11465 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11466 are essentially identical to their `..._GFp` counterparts.
11467 (For simplicity, the `..._GFp` prefix has been dropped from
11468 various internal method names.)
11470 An internal 'field_div' method (similar to 'field_mul' and
11471 'field_sqr') has been added; this is used only for binary fields.
11473 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11475 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11476 through methods ('mul', 'precompute_mult').
11478 The generic implementations (now internally called 'ec_wNAF_mul'
11479 and 'ec_wNAF_precomputed_mult') remain the default if these
11480 methods are undefined.
11482 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11484 * New function EC_GROUP_get_degree, which is defined through
11485 EC_METHOD. For curves over prime fields, this returns the bit
11486 length of the modulus.
11488 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11490 * New functions EC_GROUP_dup, EC_POINT_dup.
11491 (These simply call ..._new and ..._copy).
11493 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11495 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11496 Polynomials are represented as BIGNUMs (where the sign bit is not
11497 used) in the following functions [macros]:
11500 BN_GF2m_sub [= BN_GF2m_add]
11501 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11502 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11503 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11505 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11506 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11507 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11508 BN_GF2m_cmp [= BN_ucmp]
11510 (Note that only the 'mod' functions are actually for fields GF(2^m).
11511 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11513 For some functions, an the irreducible polynomial defining a
11514 field can be given as an 'unsigned int[]' with strictly
11515 decreasing elements giving the indices of those bits that are set;
11516 i.e., p[] represents the polynomial
11517 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11519 p[0] > p[1] > ... > p[k] = 0.
11520 This applies to the following functions:
11523 BN_GF2m_mod_mul_arr
11524 BN_GF2m_mod_sqr_arr
11525 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11526 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11527 BN_GF2m_mod_exp_arr
11528 BN_GF2m_mod_sqrt_arr
11529 BN_GF2m_mod_solve_quad_arr
11533 Conversion can be performed by the following functions:
11538 bntest.c has additional tests for binary polynomial arithmetic.
11540 Two implementations for BN_GF2m_mod_div() are available.
11541 The default algorithm simply uses BN_GF2m_mod_inv() and
11542 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11543 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11544 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11546 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11548 * Add new error code 'ERR_R_DISABLED' that can be used when some
11549 functionality is disabled at compile-time.
11551 *Douglas Stebila <douglas.stebila@sun.com>*
11553 * Change default behaviour of 'openssl asn1parse' so that more
11554 information is visible when viewing, e.g., a certificate:
11556 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11557 mode the content of non-printable OCTET STRINGs is output in a
11558 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11559 avoid the appearance of a printable string.
11561 *Nils Larsch <nla@trustcenter.de>*
11563 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11565 EC_GROUP_set_asn1_flag()
11566 EC_GROUP_get_asn1_flag()
11567 EC_GROUP_set_point_conversion_form()
11568 EC_GROUP_get_point_conversion_form()
11569 These control ASN1 encoding details:
11570 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11571 has been set to OPENSSL_EC_NAMED_CURVE.
11572 - Points are encoded in uncompressed form by default; options for
11573 asn1_for are as for point2oct, namely
11574 POINT_CONVERSION_COMPRESSED
11575 POINT_CONVERSION_UNCOMPRESSED
11576 POINT_CONVERSION_HYBRID
11578 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11580 EC_GROUP_set_seed()
11581 EC_GROUP_get0_seed()
11582 EC_GROUP_get_seed_len()
11583 This is used only for ASN1 purposes (so far).
11585 *Nils Larsch <nla@trustcenter.de>*
11587 * Add 'field_type' member to EC_METHOD, which holds the NID
11588 of the appropriate field type OID. The new function
11589 EC_METHOD_get_field_type() returns this value.
11591 *Nils Larsch <nla@trustcenter.de>*
11594 EC_POINT_point2bn()
11595 EC_POINT_bn2point()
11596 EC_POINT_point2hex()
11597 EC_POINT_hex2point()
11598 providing useful interfaces to EC_POINT_point2oct() and
11599 EC_POINT_oct2point().
11601 *Nils Larsch <nla@trustcenter.de>*
11603 * Change internals of the EC library so that the functions
11604 EC_GROUP_set_generator()
11605 EC_GROUP_get_generator()
11606 EC_GROUP_get_order()
11607 EC_GROUP_get_cofactor()
11608 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11609 to methods, which would lead to unnecessary code duplication when
11610 adding different types of curves.
11612 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11614 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11615 arithmetic, and such that modified wNAFs are generated
11616 (which avoid length expansion in many cases).
11620 * Add a function EC_GROUP_check_discriminant() (defined via
11621 EC_METHOD) that verifies that the curve discriminant is non-zero.
11623 Add a function EC_GROUP_check() that makes some sanity tests
11624 on a EC_GROUP, its generator and order. This includes
11625 EC_GROUP_check_discriminant().
11627 *Nils Larsch <nla@trustcenter.de>*
11629 * Add ECDSA in new directory crypto/ecdsa/.
11631 Add applications 'openssl ecparam' and 'openssl ecdsa'
11632 (these are based on 'openssl dsaparam' and 'openssl dsa').
11634 ECDSA support is also included in various other files across the
11635 library. Most notably,
11636 - 'openssl req' now has a '-newkey ecdsa:file' option;
11637 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11638 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11639 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11640 them suitable for ECDSA where domain parameters must be
11641 extracted before the specific public key;
11642 - ECDSA engine support has been added.
11644 *Nils Larsch <nla@trustcenter.de>*
11646 * Include some named elliptic curves, and add OIDs from X9.62,
11647 SECG, and WAP/WTLS. Each curve can be obtained from the new
11649 EC_GROUP_new_by_curve_name(),
11650 and the list of available named curves can be obtained with
11651 EC_get_builtin_curves().
11652 Also add a 'curve_name' member to EC_GROUP objects, which can be
11654 EC_GROUP_set_curve_name()
11655 EC_GROUP_get_curve_name()
11657 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11659 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11660 was actually never needed) and in BN_mul(). The removal in BN_mul()
11661 required a small change in bn_mul_part_recursive() and the addition
11662 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11663 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11664 bn_sub_words() and bn_add_words() except they take arrays with
11669 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11671 * Cleanse PEM buffers before freeing them since they may contain
11674 *Benjamin Bennett <ben@psc.edu>*
11676 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11677 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11678 authentication-only ciphersuites.
11682 * Since AES128 and AES256 share a single mask bit in the logic of
11683 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11684 kludge to work properly if AES128 is available and AES256 isn't.
11688 * Expand security boundary to match 1.1.1 module.
11692 * Remove redundant features: hash file source, editing of test vectors
11693 modify fipsld to use external fips_premain.c signature.
11697 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11698 run algorithm test programs.
11702 * Make algorithm test programs more tolerant of whitespace.
11706 * Have SSL/TLS server implementation tolerate "mismatched" record
11707 protocol version while receiving ClientHello even if the
11708 ClientHello is fragmented. (The server can't insist on the
11709 particular protocol version it has chosen before the ServerHello
11710 message has informed the client about his choice.)
11714 * Load error codes if they are not already present instead of using a
11715 static variable. This allows them to be cleanly unloaded and reloaded.
11719 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11721 * Introduce limits to prevent malicious keys being able to
11722 cause a denial of service. ([CVE-2006-2940])
11724 *Steve Henson, Bodo Moeller*
11726 * Fix ASN.1 parsing of certain invalid structures that can result
11727 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11729 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11730 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11732 * Fix SSL client code which could crash if connecting to a
11733 malicious SSLv2 server. ([CVE-2006-4343])
11735 *Tavis Ormandy and Will Drewry, Google Security Team*
11737 * Change ciphersuite string processing so that an explicit
11738 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11739 will no longer include "AES128-SHA"), and any other similar
11740 ciphersuite (same bitmap) from *other* protocol versions (so that
11741 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11742 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11743 changes from 0.9.8b and 0.9.8d.
11747 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11749 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11750 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11752 * Change the Unix randomness entropy gathering to use poll() when
11753 possible instead of select(), since the latter has some
11754 undesirable limitations.
11756 *Darryl Miles via Richard Levitte and Bodo Moeller*
11758 * Disable rogue ciphersuites:
11760 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11761 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11762 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11764 The latter two were purportedly from
11765 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11768 Also deactivate the remaining ciphersuites from
11769 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11770 unofficial, and the ID has long expired.
11774 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11775 dual-core machines) and other potential thread-safety issues.
11779 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11781 * Adapt fipsld and the build system to link against the validated FIPS
11782 module in FIPS mode.
11786 * Fixes for VC++ 2005 build under Windows.
11790 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11791 from a Windows bash shell such as MSYS. It is autodetected from the
11792 "config" script when run from a VC++ environment. Modify standard VC++
11793 build to use fipscanister.o from the GNU make build.
11797 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11799 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11800 The value now differs depending on if you build for FIPS or not.
11801 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11802 safely run with a non-FIPSed libcrypto, as it may crash because of
11803 the difference induced by this change.
11807 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11809 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11810 (part of SSL_OP_ALL). This option used to disable the
11811 countermeasure against man-in-the-middle protocol-version
11812 rollback in the SSL 2.0 server implementation, which is a bad
11813 idea. ([CVE-2005-2969])
11815 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11816 for Information Security, National Institute of Advanced Industrial
11817 Science and Technology [AIST, Japan)]*
11819 * Minimal support for X9.31 signatures and PSS padding modes. This is
11820 mainly for FIPS compliance and not fully integrated at this stage.
11824 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11825 the exponentiation using a fixed-length exponent. (Otherwise,
11826 the information leaked through timing could expose the secret key
11827 after many signatures; cf. Bleichenbacher's attack on DSA with
11832 * Make a new fixed-window mod_exp implementation the default for
11833 RSA, DSA, and DH private-key operations so that the sequence of
11834 squares and multiplies and the memory access pattern are
11835 independent of the particular secret key. This will mitigate
11836 cache-timing and potential related attacks.
11838 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11839 and this is automatically used by BN_mod_exp_mont() if the new flag
11840 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11841 will use this BN flag for private exponents unless the flag
11842 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11843 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11845 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11847 * Change the client implementation for SSLv23_method() and
11848 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11849 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11850 (Previously, the SSL 2.0 backwards compatible Client Hello
11851 message format would be used even with SSL_OP_NO_SSLv2.)
11855 * Add support for smime-type MIME parameter in S/MIME messages which some
11860 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11861 a threadsafe manner. Modify rsa code to use new function and add calls
11862 to dsa and dh code (which had race conditions before).
11866 * Include the fixed error library code in the C error file definitions
11867 instead of fixing them up at runtime. This keeps the error code
11868 structures constant.
11872 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11874 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11877 * Fixes for newer kerberos headers. NB: the casts are needed because
11878 the 'length' field is signed on one version and unsigned on another
11879 with no (?) obvious way to tell the difference, without these VC++
11880 complains. Also the "definition" of FAR (blank) is no longer included
11881 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11882 some needed definitions.
11886 * Undo Cygwin change.
11890 * Added support for proxy certificates according to RFC 3820.
11891 Because they may be a security thread to unaware applications,
11892 they must be explicitly allowed in run-time. See
11893 docs/HOWTO/proxy_certificates.txt for further information.
11897 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11899 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11900 server and client random values. Previously
11901 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11902 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11904 This change has negligible security impact because:
11906 1. Server and client random values still have 24 bytes of pseudo random
11909 2. Server and client random values are sent in the clear in the initial
11912 3. The master secret is derived using the premaster secret (48 bytes in
11913 size for static RSA ciphersuites) as well as client server and random
11916 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11919 *Stephen Henson, reported by UK NISCC*
11921 * Use Windows randomness collection on Cygwin.
11925 * Fix hang in EGD/PRNGD query when communication socket is closed
11926 prematurely by EGD/PRNGD.
11928 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11930 * Prompt for pass phrases when appropriate for PKCS12 input format.
11934 * Back-port of selected performance improvements from development
11935 branch, as well as improved support for PowerPC platforms.
11939 * Add lots of checks for memory allocation failure, error codes to indicate
11940 failure and freeing up memory if a failure occurs.
11942 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11944 * Add new -passin argument to dgst.
11948 * Perform some character comparisons of different types in X509_NAME_cmp:
11949 this is needed for some certificates that re-encode DNs into UTF8Strings
11950 (in violation of RFC3280) and can't or won't issue name rollover
11955 * Make an explicit check during certificate validation to see that
11956 the CA setting in each certificate on the chain is correct. As a
11957 side effect always do the following basic checks on extensions,
11958 not just when there's an associated purpose to the check:
11960 - if there is an unhandled critical extension (unless the user
11961 has chosen to ignore this fault)
11962 - if the path length has been exceeded (if one is set at all)
11963 - that certain extensions fit the associated purpose (if one has
11968 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11970 * Avoid a race condition when CRLs are checked in a multi threaded
11971 environment. This would happen due to the reordering of the revoked
11972 entries during signature checking and serial number lookup. Now the
11973 encoding is cached and the serial number sort performed under a lock.
11974 Add new STACK function sk_is_sorted().
11978 * Add Delta CRL to the extension code.
11982 * Various fixes to s3_pkt.c so alerts are sent properly.
11984 *David Holmes <d.holmes@f5.com>*
11986 * Reduce the chances of duplicate issuer name and serial numbers (in
11987 violation of RFC3280) using the OpenSSL certificate creation utilities.
11988 This is done by creating a random 64 bit value for the initial serial
11989 number when a serial number file is created or when a self signed
11990 certificate is created using 'openssl req -x509'. The initial serial
11991 number file is created using 'openssl x509 -next_serial' in CA.pl
11992 rather than being initialized to 1.
11996 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11998 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11999 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12001 *Joe Orton, Steve Henson*
12003 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12006 *Joe Orton, Steve Henson*
12008 * Make it possible to have multiple active certificates with the same
12009 subject in the CA index file. This is done only if the keyword
12010 'unique_subject' is set to 'no' in the main CA section (default
12011 if 'CA_default') of the configuration file. The value is saved
12012 with the database itself in a separate index attribute file,
12013 named like the index file with '.attr' appended to the name.
12017 * X509 verify fixes. Disable broken certificate workarounds when
12018 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12019 keyUsage extension present. Don't accept CRLs with unhandled critical
12020 extensions: since verify currently doesn't process CRL extensions this
12021 rejects a CRL with *any* critical extensions. Add new verify error codes
12026 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12027 A clarification of RFC2560 will require the use of OCTET STRINGs and
12028 some implementations cannot handle the current raw format. Since OpenSSL
12029 copies and compares OCSP nonces as opaque blobs without any attempt at
12030 parsing them this should not create any compatibility issues.
12034 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12035 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12036 this HMAC (and other) operations are several times slower than OpenSSL
12041 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12043 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12045 * Use the correct content when signing type "other".
12049 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12051 * Fix various bugs revealed by running the NISCC test suite:
12053 Stop out of bounds reads in the ASN1 code when presented with
12054 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12056 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12058 If verify callback ignores invalid public key errors don't try to check
12059 certificate signature with the NULL public key.
12063 * New -ignore_err option in ocsp application to stop the server
12064 exiting on the first error in a request.
12068 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12069 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12074 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12075 extra data after the compression methods not only for TLS 1.0
12076 but also for SSL 3.0 (as required by the specification).
12078 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12080 * Change X509_certificate_type() to mark the key as exported/exportable
12081 when it's 512 *bits* long, not 512 bytes.
12085 * Change AES_cbc_encrypt() so it outputs exact multiple of
12086 blocks during encryption.
12090 * Various fixes to base64 BIO and non blocking I/O. On write
12091 flushes were not handled properly if the BIO retried. On read
12092 data was not being buffered properly and had various logic bugs.
12093 This also affects blocking I/O when the data being decoded is a
12098 * Various S/MIME bugfixes and compatibility changes:
12099 output correct application/pkcs7 MIME type if
12100 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12101 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12102 of files as .eml work). Correctly handle very long lines in MIME
12107 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12109 * Countermeasure against the Klima-Pokorny-Rosa extension of
12110 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12111 a protocol version number mismatch like a decryption error
12112 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12116 * Turn on RSA blinding by default in the default implementation
12117 to avoid a timing attack. Applications that don't want it can call
12118 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12119 They would be ill-advised to do so in most cases.
12121 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12123 * Change RSA blinding code so that it works when the PRNG is not
12124 seeded (in this case, the secret RSA exponent is abused as
12125 an unpredictable seed -- if it is not unpredictable, there
12126 is no point in blinding anyway). Make RSA blinding thread-safe
12127 by remembering the creator's thread ID in rsa->blinding and
12128 having all other threads use local one-time blinding factors
12129 (this requires more computation than sharing rsa->blinding, but
12130 avoids excessive locking; and if an RSA object is not shared
12131 between threads, blinding will still be very fast).
12135 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12136 ENGINE as defaults for all supported algorithms irrespective of
12137 the 'flags' parameter. 'flags' is now honoured, so applications
12138 should make sure they are passing it correctly.
12142 * Target "mingw" now allows native Windows code to be generated in
12143 the Cygwin environment as well as with the MinGW compiler.
12147 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12149 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12150 via timing by performing a MAC computation even if incorrect
12151 block cipher padding has been found. This is a countermeasure
12152 against active attacks where the attacker has to distinguish
12153 between bad padding and a MAC verification error. ([CVE-2003-0078])
12155 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12156 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12157 Martin Vuagnoux (EPFL, Ilion)*
12159 * Make the no-err option work as intended. The intention with no-err
12160 is not to have the whole error stack handling routines removed from
12161 libcrypto, it's only intended to remove all the function name and
12162 reason texts, thereby removing some of the footprint that may not
12163 be interesting if those errors aren't displayed anyway.
12165 NOTE: it's still possible for any application or module to have its
12166 own set of error texts inserted. The routines are there, just not
12167 used by default when no-err is given.
12171 * Add support for FreeBSD on IA64.
12173 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12175 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12176 Kerberos function mit_des_cbc_cksum(). Before this change,
12177 the value returned by DES_cbc_cksum() was like the one from
12178 mit_des_cbc_cksum(), except the bytes were swapped.
12180 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12182 * Allow an application to disable the automatic SSL chain building.
12183 Before this a rather primitive chain build was always performed in
12184 ssl3_output_cert_chain(): an application had no way to send the
12185 correct chain if the automatic operation produced an incorrect result.
12187 Now the chain builder is disabled if either:
12189 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12191 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12193 The reasoning behind this is that an application would not want the
12194 auto chain building to take place if extra chain certificates are
12195 present and it might also want a means of sending no additional
12196 certificates (for example the chain has two certificates and the
12201 * Add the possibility to build without the ENGINE framework.
12203 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12205 * Under Win32 gmtime() can return NULL: check return value in
12206 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12210 * DSA routines: under certain error conditions uninitialized BN objects
12211 could be freed. Solution: make sure initialization is performed early
12212 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12213 Nils Larsch <nla@trustcenter.de> via PR#459)
12217 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12218 checked on reconnect on the client side, therefore session resumption
12219 could still fail with a "ssl session id is different" error. This
12220 behaviour is masked when SSL_OP_ALL is used due to
12221 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12222 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12223 followup to PR #377.
12227 * IA-32 assembler support enhancements: unified ELF targets, support
12228 for SCO/Caldera platforms, fix for Cygwin shared build.
12232 * Add support for FreeBSD on sparc64. As a consequence, support for
12233 FreeBSD on non-x86 processors is separate from x86 processors on
12234 the config script, much like the NetBSD support.
12236 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12238 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12240 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12243 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12244 code (06) was taken as the first octet of the session ID and the last
12245 octet was ignored consequently. As a result SSLv2 client side session
12246 caching could not have worked due to the session ID mismatch between
12248 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12253 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12254 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12259 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12260 seems that in spite of existing for more than a year, many application
12261 author have done nothing to provide the necessary callbacks, which
12262 means that this particular engine will not work properly anywhere.
12263 This is a very unfortunate situation which forces us, in the name
12264 of usability, to give the hw_ncipher.c a static lock, which is part
12266 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12267 appear in 0.9.8 or later. We EXPECT application authors to have
12268 dealt properly with this when 0.9.8 is released (unless we actually
12269 make such changes in the libcrypto locking code that changes will
12270 have to be made anyway).
12274 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12275 octets have been read, EOF or an error occurs. Without this change
12276 some truncated ASN1 structures will not produce an error.
12280 * Disable Heimdal support, since it hasn't been fully implemented.
12281 Still give the possibility to force the use of Heimdal, but with
12282 warnings and a request that patches get sent to openssl-dev.
12286 * Add the VC-CE target, introduce the WINCE sysname, and add
12287 INSTALL.WCE and appropriate conditionals to make it build.
12289 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12291 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12292 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12293 edit numbers of the version.
12295 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12297 * Introduce safe string copy and catenation functions
12298 (BUF_strlcpy() and BUF_strlcat()).
12300 *Ben Laurie (CHATS) and Richard Levitte*
12302 * Avoid using fixed-size buffers for one-line DNs.
12304 *Ben Laurie (CHATS)*
12306 * Add BUF_MEM_grow_clean() to avoid information leakage when
12307 resizing buffers containing secrets, and use where appropriate.
12309 *Ben Laurie (CHATS)*
12311 * Avoid using fixed size buffers for configuration file location.
12313 *Ben Laurie (CHATS)*
12315 * Avoid filename truncation for various CA files.
12317 *Ben Laurie (CHATS)*
12319 * Use sizeof in preference to magic numbers.
12321 *Ben Laurie (CHATS)*
12323 * Avoid filename truncation in cert requests.
12325 *Ben Laurie (CHATS)*
12327 * Add assertions to check for (supposedly impossible) buffer
12330 *Ben Laurie (CHATS)*
12332 * Don't cache truncated DNS entries in the local cache (this could
12333 potentially lead to a spoofing attack).
12335 *Ben Laurie (CHATS)*
12337 * Fix various buffers to be large enough for hex/decimal
12338 representations in a platform independent manner.
12340 *Ben Laurie (CHATS)*
12342 * Add CRYPTO_realloc_clean() to avoid information leakage when
12343 resizing buffers containing secrets, and use where appropriate.
12345 *Ben Laurie (CHATS)*
12347 * Add BIO_indent() to avoid much slightly worrying code to do
12350 *Ben Laurie (CHATS)*
12352 * Convert sprintf()/BIO_puts() to BIO_printf().
12354 *Ben Laurie (CHATS)*
12356 * buffer_gets() could terminate with the buffer only half
12359 *Ben Laurie (CHATS)*
12361 * Add assertions to prevent user-supplied crypto functions from
12362 overflowing internal buffers by having large block sizes, etc.
12364 *Ben Laurie (CHATS)*
12366 * New OPENSSL_assert() macro (similar to assert(), but enabled
12369 *Ben Laurie (CHATS)*
12371 * Eliminate unused copy of key in RC4.
12373 *Ben Laurie (CHATS)*
12375 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12377 *Ben Laurie (CHATS)*
12379 * Fix off-by-one error in EGD path.
12381 *Ben Laurie (CHATS)*
12383 * If RANDFILE path is too long, ignore instead of truncating.
12385 *Ben Laurie (CHATS)*
12387 * Eliminate unused and incorrectly sized X.509 structure
12390 *Ben Laurie (CHATS)*
12392 * Eliminate unused and dangerous function knumber().
12394 *Ben Laurie (CHATS)*
12396 * Eliminate unused and dangerous structure, KSSL_ERR.
12398 *Ben Laurie (CHATS)*
12400 * Protect against overlong session ID context length in an encoded
12401 session object. Since these are local, this does not appear to be
12404 *Ben Laurie (CHATS)*
12406 * Change from security patch (see 0.9.6e below) that did not affect
12407 the 0.9.6 release series:
12409 Remote buffer overflow in SSL3 protocol - an attacker could
12410 supply an oversized master key in Kerberos-enabled versions.
12413 *Ben Laurie (CHATS)*
12415 * Change the SSL kerb5 codes to match RFC 2712.
12419 * Make -nameopt work fully for req and add -reqopt switch.
12421 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12423 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12425 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12427 * Make sure tests can be performed even if the corresponding algorithms
12428 have been removed entirely. This was also the last step to make
12429 OpenSSL compilable with DJGPP under all reasonable conditions.
12431 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12433 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12434 to allow version independent disabling of normally unselected ciphers,
12435 which may be activated as a side-effect of selecting a single cipher.
12437 (E.g., cipher list string "RSA" enables ciphersuites that are left
12438 out of "ALL" because they do not provide symmetric encryption.
12439 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12441 *Lutz Jaenicke, Bodo Moeller*
12443 * Add appropriate support for separate platform-dependent build
12444 directories. The recommended way to make a platform-dependent
12445 build directory is the following (tested on Linux), maybe with
12448 # Place yourself outside of the OpenSSL source tree. In
12449 # this example, the environment variable OPENSSL_SOURCE
12450 # is assumed to contain the absolute OpenSSL source directory.
12451 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12452 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12453 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12454 mkdir -p `dirname $F`
12455 ln -s $OPENSSL_SOURCE/$F $F
12458 To be absolutely sure not to disturb the source tree, a "make clean"
12459 is a good thing. If it isn't successful, don't worry about it,
12460 it probably means the source directory is very clean.
12464 * Make sure any ENGINE control commands make local copies of string
12465 pointers passed to them whenever necessary. Otherwise it is possible
12466 the caller may have overwritten (or deallocated) the original string
12467 data when a later ENGINE operation tries to use the stored values.
12469 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12471 * Improve diagnostics in file reading and command-line digests.
12473 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12475 * Add AES modes CFB and OFB to the object database. Correct an
12476 error in AES-CFB decryption.
12480 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12481 allows existing EVP_CIPHER_CTX structures to be reused after
12482 calling `EVP_*Final()`. This behaviour is used by encryption
12483 BIOs and some applications. This has the side effect that
12484 applications must explicitly clean up cipher contexts with
12485 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12489 * Check the values of dna and dnb in bn_mul_recursive before calling
12490 bn_mul_comba (a non zero value means the a or b arrays do not contain
12491 n2 elements) and fallback to bn_mul_normal if either is not zero.
12495 * Fix escaping of non-ASCII characters when using the -subj option
12496 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12500 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12501 form for "surname", serialNumber has no short form.
12502 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12503 therefore remove "mail" short name for "internet 7".
12504 The OID for unique identifiers in X509 certificates is
12505 x500UniqueIdentifier, not uniqueIdentifier.
12506 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12510 * Add an "init" command to the ENGINE config module and auto initialize
12511 ENGINEs. Without any "init" command the ENGINE will be initialized
12512 after all ctrl commands have been executed on it. If init=1 the
12513 ENGINE is initialized at that point (ctrls before that point are run
12514 on the uninitialized ENGINE and after on the initialized one). If
12515 init=0 then the ENGINE will not be initialized at all.
12519 * Fix the 'app_verify_callback' interface so that the user-defined
12520 argument is actually passed to the callback: In the
12521 SSL_CTX_set_cert_verify_callback() prototype, the callback
12522 declaration has been changed from
12525 int (*cb)(X509_STORE_CTX *,void *);
12526 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12527 i=s->ctx->app_verify_callback(&ctx)
12528 has been changed into
12529 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12531 To update applications using SSL_CTX_set_cert_verify_callback(),
12532 a dummy argument can be added to their callback functions.
12534 *D. K. Smetters <smetters@parc.xerox.com>*
12536 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12538 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12540 * Add and OPENSSL_LOAD_CONF define which will cause
12541 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12542 This allows older applications to transparently support certain
12543 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12544 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12545 load the config file and OPENSSL_add_all_algorithms_conf() which will
12546 always load it have also been added.
12550 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12551 Adjust NIDs and EVP layer.
12553 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12555 * Config modules support in openssl utility.
12557 Most commands now load modules from the config file,
12558 though in a few (such as version) this isn't done
12559 because it couldn't be used for anything.
12561 In the case of ca and req the config file used is
12562 the same as the utility itself: that is the -config
12563 command line option can be used to specify an
12568 * Move default behaviour from OPENSSL_config(). If appname is NULL
12569 use "openssl_conf" if filename is NULL use default openssl config file.
12573 * Add an argument to OPENSSL_config() to allow the use of an alternative
12574 config section name. Add a new flag to tolerate a missing config file
12575 and move code to CONF_modules_load_file().
12579 * Support for crypto accelerator cards from Accelerated Encryption
12580 Processing, www.aep.ie. (Use engine 'aep')
12581 The support was copied from 0.9.6c [engine] and adapted/corrected
12582 to work with the new engine framework.
12584 *AEP Inc. and Richard Levitte*
12586 * Support for SureWare crypto accelerator cards from Baltimore
12587 Technologies. (Use engine 'sureware')
12588 The support was copied from 0.9.6c [engine] and adapted
12589 to work with the new engine framework.
12593 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12594 make the newer ENGINE framework commands for the CHIL engine work.
12596 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12598 * Make it possible to produce shared libraries on ReliantUNIX.
12600 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12602 * Add the configuration target debug-linux-ppro.
12603 Make 'openssl rsa' use the general key loading routines
12604 implemented in `apps.c`, and make those routines able to
12605 handle the key format FORMAT_NETSCAPE and the variant
12608 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12610 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12612 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12614 * Add -keyform to rsautl, and document -engine.
12616 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12618 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12619 BIO_R_NO_SUCH_FILE error code rather than the generic
12620 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12624 * Add new functions
12625 ERR_peek_last_error
12626 ERR_peek_last_error_line
12627 ERR_peek_last_error_line_data.
12628 These are similar to
12630 ERR_peek_error_line
12631 ERR_peek_error_line_data,
12632 but report on the latest error recorded rather than the first one
12633 still in the error queue.
12635 *Ben Laurie, Bodo Moeller*
12637 * default_algorithms option in ENGINE config module. This allows things
12639 default_algorithms = ALL
12640 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12644 * Preliminary ENGINE config module.
12648 * New experimental application configuration code.
12652 * Change the AES code to follow the same name structure as all other
12653 symmetric ciphers, and behave the same way. Move everything to
12654 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12656 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12658 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12660 *Ben Laurie and Theo de Raadt*
12662 * Add option to output public keys in req command.
12664 *Massimiliano Pala madwolf@openca.org*
12666 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12667 (up to about 10% better than before for P-192 and P-224).
12671 * New functions/macros
12673 SSL_CTX_set_msg_callback(ctx, cb)
12674 SSL_CTX_set_msg_callback_arg(ctx, arg)
12675 SSL_set_msg_callback(ssl, cb)
12676 SSL_set_msg_callback_arg(ssl, arg)
12678 to request calling a callback function
12680 void cb(int write_p, int version, int content_type,
12681 const void *buf, size_t len, SSL *ssl, void *arg)
12683 whenever a protocol message has been completely received
12684 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12685 protocol version according to which the SSL library interprets
12686 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12687 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12688 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12689 specification (change_cipher_spec(20), alert(21), handshake(22)).
12690 'buf' and 'len' point to the actual message, 'ssl' to the
12691 SSL object, and 'arg' is the application-defined value set by
12692 SSL[_CTX]_set_msg_callback_arg().
12694 'openssl s_client' and 'openssl s_server' have new '-msg' options
12695 to enable a callback that displays all protocol messages.
12699 * Change the shared library support so shared libraries are built as
12700 soon as the corresponding static library is finished, and thereby get
12701 openssl and the test programs linked against the shared library.
12702 This still only happens when the keyword "shard" has been given to
12703 the configuration scripts.
12705 NOTE: shared library support is still an experimental thing, and
12706 backward binary compatibility is still not guaranteed.
12708 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12710 * Add support for Subject Information Access extension.
12712 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12714 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12715 additional bytes when new memory had to be allocated, not just
12716 when reusing an existing buffer.
12720 * New command line and configuration option 'utf8' for the req command.
12721 This allows field values to be specified as UTF8 strings.
12725 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12726 runs for the former and machine-readable output for the latter.
12730 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12731 of the e-mail address in the DN (i.e., it will go into a certificate
12732 extension only). The new configuration file option 'email_in_dn = no'
12733 has the same effect.
12735 *Massimiliano Pala madwolf@openca.org*
12737 * Change all functions with names starting with `des_` to be starting
12738 with `DES_` instead. Add wrappers that are compatible with libdes,
12739 but are named `_ossl_old_des_*`. Finally, add macros that map the
12740 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12741 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12742 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12745 Since we provide two compatibility mappings, the user needs to
12746 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12747 compatibility is desired. The default (i.e., when that macro
12748 isn't defined) is OpenSSL 0.9.6c compatibility.
12750 There are also macros that enable and disable the support of old
12751 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12752 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12753 are defined, the default will apply: to support the old des routines.
12755 In either case, one must include openssl/des.h to get the correct
12756 definitions. Do not try to just include openssl/des_old.h, that
12759 NOTE: This is a major break of an old API into a new one. Software
12760 authors are encouraged to switch to the `DES_` style functions. Some
12761 time in the future, des_old.h and the libdes compatibility functions
12762 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12763 default), and then completely removed.
12767 * Test for certificates which contain unsupported critical extensions.
12768 If such a certificate is found during a verify operation it is
12769 rejected by default: this behaviour can be overridden by either
12770 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12771 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12772 X509_supported_extension() has also been added which returns 1 if a
12773 particular extension is supported.
12777 * Modify the behaviour of EVP cipher functions in similar way to digests
12778 to retain compatibility with existing code.
12782 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12783 compatibility with existing code. In particular the 'ctx' parameter does
12784 not have to be to be initialized before the call to EVP_DigestInit() and
12785 it is tidied up after a call to EVP_DigestFinal(). New function
12786 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12787 EVP_MD_CTX_copy() changed to not require the destination to be
12788 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12789 requires the destination to be valid.
12791 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12792 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12796 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12797 so that complete 'Handshake' protocol structures are kept in memory
12798 instead of overwriting 'msg_type' and 'length' with 'body' data.
12802 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12804 *Massimo Santin via Richard Levitte*
12806 * Major restructuring to the underlying ENGINE code. This includes
12807 reduction of linker bloat, separation of pure "ENGINE" manipulation
12808 (initialisation, etc) from functionality dealing with implementations
12809 of specific crypto interfaces. This change also introduces integrated
12810 support for symmetric ciphers and digest implementations - so ENGINEs
12811 can now accelerate these by providing EVP_CIPHER and EVP_MD
12812 implementations of their own. This is detailed in
12813 [crypto/engine/README.md](crypto/engine/README.md)
12814 as it couldn't be adequately described here. However, there are a few
12815 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12816 were changed in the original introduction of ENGINE code have now
12817 reverted back - the hooking from this code to ENGINE is now a good
12818 deal more passive and at run-time, operations deal directly with
12819 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12820 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12821 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12822 they were not being used by the framework as there is no concept of a
12823 BIGNUM_METHOD and they could not be generalised to the new
12824 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12825 ENGINE_cpy() has been removed as it cannot be consistently defined in
12830 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12834 * Change mkdef.pl to sort symbols that get the same entry number,
12835 and make sure the automatically generated functions `ERR_load_*`
12836 become part of libeay.num as well.
12840 * New function SSL_renegotiate_pending(). This returns true once
12841 renegotiation has been requested (either SSL_renegotiate() call
12842 or HelloRequest/ClientHello received from the peer) and becomes
12843 false once a handshake has been completed.
12844 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12845 sends a HelloRequest, but does not ensure that a handshake takes
12846 place. SSL_renegotiate_pending() is useful for checking if the
12847 client has followed the request.)
12851 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12852 By default, clients may request session resumption even during
12853 renegotiation (if session ID contexts permit); with this option,
12854 session resumption is possible only in the first handshake.
12856 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12857 more bits available for options that should not be part of
12858 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12862 * Add some demos for certificate and certificate request creation.
12866 * Make maximum certificate chain size accepted from the peer application
12867 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12868 "Douglas E. Engert" <deengert@anl.gov>.
12872 * Add support for shared libraries for Unixware-7
12873 (Boyd Lynn Gerber <gerberb@zenez.com>).
12877 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12878 be done prior to destruction. Use this to unload error strings from
12879 ENGINEs that load their own error strings. NB: This adds two new API
12880 functions to "get" and "set" this destroy handler in an ENGINE.
12884 * Alter all existing ENGINE implementations (except "openssl" and
12885 "openbsd") to dynamically instantiate their own error strings. This
12886 makes them more flexible to be built both as statically-linked ENGINEs
12887 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12888 Also, add stub code to each that makes building them as self-contained
12889 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12893 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12894 implementations into applications that are completely implemented in
12895 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12896 commands that can be used to configure what shared-library to load and
12897 to control aspects of the way it is handled. Also, made an update to
12898 the [README-Engine.md](README-Engine.md) file
12899 that brings its information up-to-date and
12900 provides some information and instructions on the "dynamic" ENGINE
12901 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12905 * Make it possible to unload ranges of ERR strings with a new
12906 "ERR_unload_strings" function.
12910 * Add a copy() function to EVP_MD.
12914 * Make EVP_MD routines take a context pointer instead of just the
12915 md_data void pointer.
12919 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12920 that the digest can only process a single chunk of data
12921 (typically because it is provided by a piece of
12922 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12923 is only going to provide a single chunk of data, and hence the
12924 framework needn't accumulate the data for oneshot drivers.
12928 * As with "ERR", make it possible to replace the underlying "ex_data"
12929 functions. This change also alters the storage and management of global
12930 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12931 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12932 index counters. The API functions that use this state have been changed
12933 to take a "class_index" rather than pointers to the class's local STACK
12934 and counter, and there is now an API function to dynamically create new
12935 classes. This centralisation allows us to (a) plug a lot of the
12936 thread-safety problems that existed, and (b) makes it possible to clean
12937 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12938 such data would previously have always leaked in application code and
12939 workarounds were in place to make the memory debugging turn a blind eye
12940 to it. Application code that doesn't use this new function will still
12941 leak as before, but their memory debugging output will announce it now
12942 rather than letting it slide.
12944 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12945 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12946 has a return value to indicate success or failure.
12950 * Make it possible to replace the underlying "ERR" functions such that the
12951 global state (2 LHASH tables and 2 locks) is only used by the "default"
12952 implementation. This change also adds two functions to "get" and "set"
12953 the implementation prior to it being automatically set the first time
12954 any other ERR function takes place. Ie. an application can call "get",
12955 pass the return value to a module it has just loaded, and that module
12956 can call its own "set" function using that value. This means the
12957 module's "ERR" operations will use (and modify) the error state in the
12958 application and not in its own statically linked copy of OpenSSL code.
12962 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12963 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12964 the operation, and provides a more encapsulated way for external code
12965 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12966 to use these functions rather than manually incrementing the counts.
12968 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12972 * Add EVP test program.
12976 * Add symmetric cipher support to ENGINE. Expect the API to change!
12980 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12981 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12982 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12983 These allow a CRL to be built without having to access X509_CRL fields
12984 directly. Modify 'ca' application to use new functions.
12988 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12989 bug workarounds. Rollback attack detection is a security feature.
12990 The problem will only arise on OpenSSL servers when TLSv1 is not
12991 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12992 Software authors not wanting to support TLSv1 will have special reasons
12993 for their choice and can explicitly enable this option.
12995 *Bodo Moeller, Lutz Jaenicke*
12997 * Rationalise EVP so it can be extended: don't include a union of
12998 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12999 (similar to those existing for EVP_CIPHER_CTX).
13004 EVP_MD_CTX_init(&md); /* new function call */
13005 EVP_DigestInit(&md, EVP_sha1());
13006 EVP_DigestUpdate(&md, in, len);
13007 EVP_DigestFinal(&md, out, NULL);
13008 EVP_MD_CTX_cleanup(&md); /* new function call */
13012 * Make DES key schedule conform to the usual scheme, as well as
13013 correcting its structure. This means that calls to DES functions
13014 now have to pass a pointer to a des_key_schedule instead of a
13015 plain des_key_schedule (which was actually always a pointer
13018 des_key_schedule ks;
13020 des_set_key_checked(..., &ks);
13021 des_ncbc_encrypt(..., &ks, ...);
13023 (Note that a later change renames 'des_...' into 'DES_...'.)
13027 * Initial reduction of linker bloat: the use of some functions, such as
13028 PEM causes large amounts of unused functions to be linked in due to
13029 poor organisation. For example pem_all.c contains every PEM function
13030 which has a knock on effect of linking in large amounts of (unused)
13031 ASN1 code. Grouping together similar functions and splitting unrelated
13032 functions prevents this.
13036 * Cleanup of EVP macros.
13040 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13041 correct `_ecb suffix`.
13045 * Add initial OCSP responder support to ocsp application. The
13046 revocation information is handled using the text based index
13047 use by the ca application. The responder can either handle
13048 requests generated internally, supplied in files (for example
13049 via a CGI script) or using an internal minimal server.
13053 * Add configuration choices to get zlib compression for TLS.
13057 * Changes to Kerberos SSL for RFC 2712 compliance:
13058 1. Implemented real KerberosWrapper, instead of just using
13059 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13060 2. Implemented optional authenticator field of KerberosWrapper.
13062 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13063 and authenticator structs; see crypto/krb5/.
13065 Generalized Kerberos calls to support multiple Kerberos libraries.
13066 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13067 via Richard Levitte*
13069 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13070 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13071 values for each of the key sizes rather than having just
13072 parameters (and 'speed' generating keys each time).
13076 * Speed up EVP routines.
13079 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13080 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13081 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13082 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13084 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13085 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13086 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13089 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13091 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13095 * Added the OS2-EMX target.
13097 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13099 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13100 New functions to support `NCONF` routines in extension code.
13101 New function `CONF_set_nconf()`
13102 to allow functions which take an `NCONF` to also handle the old `LHASH`
13103 structure: this means that the old `CONF` compatible routines can be
13104 retained (in particular w.rt. extensions) without having to duplicate the
13105 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13109 * Enhance the general user interface with mechanisms for inner control
13110 and with possibilities to have yes/no kind of prompts.
13114 * Change all calls to low-level digest routines in the library and
13115 applications to use EVP. Add missing calls to HMAC_cleanup() and
13116 don't assume HMAC_CTX can be copied using memcpy().
13118 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13120 * Add the possibility to control engines through control names but with
13121 arbitrary arguments instead of just a string.
13122 Change the key loaders to take a UI_METHOD instead of a callback
13123 function pointer. NOTE: this breaks binary compatibility with earlier
13124 versions of OpenSSL [engine].
13125 Adapt the nCipher code for these new conditions and add a card insertion
13130 * Enhance the general user interface with mechanisms to better support
13131 dialog box interfaces, application-defined prompts, the possibility
13132 to use defaults (for example default passwords from somewhere else)
13133 and interrupts/cancellations.
13137 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13138 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13142 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13143 tidy up some unnecessarily weird code in 'sk_new()').
13145 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13147 * Change the key loading routines for ENGINEs to use the same kind
13148 callback (pem_password_cb) as all other routines that need this
13153 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13154 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13155 than this minimum value is recommended.
13159 * New random seeder for OpenVMS, using the system process statistics
13160 that are easily reachable.
13164 * Windows apparently can't transparently handle global
13165 variables defined in DLLs. Initialisations such as:
13167 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13169 won't compile. This is used by the any applications that need to
13170 declare their own ASN1 modules. This was fixed by adding the option
13171 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13172 needed for static libraries under Win32.
13176 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13177 setting of purpose and trust fields. New X509_STORE trust and
13178 purpose functions and tidy up setting in other SSL functions.
13182 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13183 structure. These are inherited by X509_STORE_CTX when it is
13184 initialised. This allows various defaults to be set in the
13185 X509_STORE structure (such as flags for CRL checking and custom
13186 purpose or trust settings) for functions which only use X509_STORE_CTX
13187 internally such as S/MIME.
13189 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13190 trust settings if they are not set in X509_STORE. This allows X509_STORE
13191 purposes and trust (in S/MIME for example) to override any set by default.
13193 Add command line options for CRL checking to smime, s_client and s_server
13198 * Initial CRL based revocation checking. If the CRL checking flag(s)
13199 are set then the CRL is looked up in the X509_STORE structure and
13200 its validity and signature checked, then if the certificate is found
13201 in the CRL the verify fails with a revoked error.
13203 Various new CRL related callbacks added to X509_STORE_CTX structure.
13205 Command line options added to 'verify' application to support this.
13207 This needs some additional work, such as being able to handle multiple
13208 CRLs with different times, extension based lookup (rather than just
13209 by subject name) and ultimately more complete V2 CRL extension
13214 * Add a general user interface API (crypto/ui/). This is designed
13215 to replace things like des_read_password and friends (backward
13216 compatibility functions using this new API are provided).
13217 The purpose is to remove prompting functions from the DES code
13218 section as well as provide for prompting through dialog boxes in
13219 a window system and the like.
13223 * Add "ex_data" support to ENGINE so implementations can add state at a
13224 per-structure level rather than having to store it globally.
13228 * Make it possible for ENGINE structures to be copied when retrieved by
13229 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13230 This causes the "original" ENGINE structure to act like a template,
13231 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13232 operational state can be localised to each ENGINE structure, despite the
13233 fact they all share the same "methods". New ENGINE structures returned in
13234 this case have no functional references and the return value is the single
13235 structural reference. This matches the single structural reference returned
13236 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13241 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13242 needs to match any other type at all we need to manually clear the
13247 * Changes to the "openssl engine" utility to include;
13248 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13249 about an ENGINE's available control commands.
13250 - executing control commands from command line arguments using the
13251 '-pre' and '-post' switches. '-post' is only used if '-t' is
13252 specified and the ENGINE is successfully initialised. The syntax for
13253 the individual commands are colon-separated, for example;
13254 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13258 * New dynamic control command support for ENGINEs. ENGINEs can now
13259 declare their own commands (numbers), names (strings), descriptions,
13260 and input types for run-time discovery by calling applications. A
13261 subset of these commands are implicitly classed as "executable"
13262 depending on their input type, and only these can be invoked through
13263 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13264 can be based on user input, config files, etc). The distinction is
13265 that "executable" commands cannot return anything other than a boolean
13266 result and can only support numeric or string input, whereas some
13267 discoverable commands may only be for direct use through
13268 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13269 pointers, or other custom uses. The "executable" commands are to
13270 support parameterisations of ENGINE behaviour that can be
13271 unambiguously defined by ENGINEs and used consistently across any
13272 OpenSSL-based application. Commands have been added to all the
13273 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13274 control over shared-library paths without source code alterations.
13278 * Changed all ENGINE implementations to dynamically allocate their
13279 ENGINEs rather than declaring them statically. Apart from this being
13280 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13281 this also allows the implementations to compile without using the
13282 internal engine_int.h header.
13286 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13287 'const' value. Any code that should be able to modify a RAND_METHOD
13288 should already have non-const pointers to it (ie. they should only
13289 modify their own ones).
13293 * Made a variety of little tweaks to the ENGINE code.
13294 - "atalla" and "ubsec" string definitions were moved from header files
13295 to C code. "nuron" string definitions were placed in variables
13296 rather than hard-coded - allowing parameterisation of these values
13297 later on via ctrl() commands.
13298 - Removed unused "#if 0"'d code.
13299 - Fixed engine list iteration code so it uses ENGINE_free() to release
13300 structural references.
13301 - Constified the RAND_METHOD element of ENGINE structures.
13302 - Constified various get/set functions as appropriate and added
13303 missing functions (including a catch-all ENGINE_cpy that duplicates
13304 all ENGINE values onto a new ENGINE except reference counts/state).
13305 - Removed NULL parameter checks in get/set functions. Setting a method
13306 or function to NULL is a way of cancelling out a previously set
13307 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13308 and doesn't justify the extra error symbols and code.
13309 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13310 flags from engine_int.h to engine.h.
13311 - Changed prototypes for ENGINE handler functions (init(), finish(),
13312 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13316 * Implement binary inversion algorithm for BN_mod_inverse in addition
13317 to the algorithm using long division. The binary algorithm can be
13318 used only if the modulus is odd. On 32-bit systems, it is faster
13319 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13320 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13321 up to 450 bits. In 64-bit environments, the binary algorithm
13322 appears to be advantageous for much longer moduli; here we use it
13323 for moduli up to 2048 bits.
13327 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13328 could not support the combine flag in choice fields.
13332 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13333 extensions from a certificate request to the certificate.
13337 * Allow multiple 'certopt' and 'nameopt' options to be separated
13338 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13339 file: this allows the display of the certificate about to be
13340 signed to be customised, to allow certain fields to be included
13341 or excluded and extension details. The old system didn't display
13342 multicharacter strings properly, omitted fields not in the policy
13343 and couldn't display additional details such as extensions.
13347 * Function EC_POINTs_mul for multiple scalar multiplication
13348 of an arbitrary number of elliptic curve points
13349 \sum scalars[i]*points[i],
13350 optionally including the generator defined for the EC_GROUP:
13351 scalar*generator + \sum scalars[i]*points[i].
13353 EC_POINT_mul is a simple wrapper function for the typical case
13354 that the point list has just one item (besides the optional
13359 * First EC_METHODs for curves over GF(p):
13361 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13362 operations and provides various method functions that can also
13363 operate with faster implementations of modular arithmetic.
13365 EC_GFp_mont_method() reuses most functions that are part of
13366 EC_GFp_simple_method, but uses Montgomery arithmetic.
13368 *Bodo Moeller; point addition and point doubling
13369 implementation directly derived from source code provided by
13370 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13372 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13373 crypto/ec/ec_lib.c):
13375 Curves are EC_GROUP objects (with an optional group generator)
13376 based on EC_METHODs that are built into the library.
13378 Points are EC_POINT objects based on EC_GROUP objects.
13380 Most of the framework would be able to handle curves over arbitrary
13381 finite fields, but as there are no obvious types for fields other
13382 than GF(p), some functions are limited to that for now.
13386 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13387 that the file contains a complete HTTP response.
13391 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13392 change the def and num file printf format specifier from "%-40sXXX"
13393 to "%-39s XXX". The latter will always guarantee a space after the
13394 field while the former will cause them to run together if the field
13395 is 40 of more characters long.
13399 * Constify the cipher and digest 'method' functions and structures
13400 and modify related functions to take constant EVP_MD and EVP_CIPHER
13405 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13406 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13410 * Modify `EVP_Digest*()` routines so they now return values. Although the
13411 internal software routines can never fail additional hardware versions
13416 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13418 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13419 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13422 ERR_R_NESTED_ASN1_ERROR
13424 ERR_R_MISSING_ASN1_EOS
13425 were 4 .. 9, conflicting with
13426 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13428 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13429 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13431 Add new error code 'ERR_R_INTERNAL_ERROR'.
13435 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13440 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13441 sets the subject name for a new request or supersedes the
13442 subject name in a given request. Formats that can be parsed are
13443 'CN=Some Name, OU=myOU, C=IT'
13445 'CN=Some Name/OU=myOU/C=IT'.
13447 Add options '-batch' and '-verbose' to 'openssl req'.
13449 *Massimiliano Pala <madwolf@hackmasters.net>*
13451 * Introduce the possibility to access global variables through
13452 functions on platform were that's the best way to handle exporting
13453 global variables in shared libraries. To enable this functionality,
13454 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13455 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13456 is normally done by Configure or something similar).
13458 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13459 in the source file (foo.c) like this:
13461 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13462 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13464 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13465 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13467 OPENSSL_DECLARE_GLOBAL(int,foo);
13468 #define foo OPENSSL_GLOBAL_REF(foo)
13469 OPENSSL_DECLARE_GLOBAL(double,bar);
13470 #define bar OPENSSL_GLOBAL_REF(bar)
13472 The #defines are very important, and therefore so is including the
13473 header file everywhere where the defined globals are used.
13475 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13476 of ASN.1 items, but that structure is a bit different.
13478 The largest change is in util/mkdef.pl which has been enhanced with
13479 better and easier to understand logic to choose which symbols should
13480 go into the Windows .def files as well as a number of fixes and code
13481 cleanup (among others, algorithm keywords are now sorted
13482 lexicographically to avoid constant rewrites).
13486 * In BN_div() keep a copy of the sign of 'num' before writing the
13487 result to 'rm' because if rm==num the value will be overwritten
13488 and produce the wrong result if 'num' is negative: this caused
13489 problems with BN_mod() and BN_nnmod().
13493 * Function OCSP_request_verify(). This checks the signature on an
13494 OCSP request and verifies the signer certificate. The signer
13495 certificate is just checked for a generic purpose and OCSP request
13500 * Add OCSP_check_validity() function to check the validity of OCSP
13501 responses. OCSP responses are prepared in real time and may only
13502 be a few seconds old. Simply checking that the current time lies
13503 between thisUpdate and nextUpdate max reject otherwise valid responses
13504 caused by either OCSP responder or client clock inaccuracy. Instead
13505 we allow thisUpdate and nextUpdate to fall within a certain period of
13506 the current time. The age of the response can also optionally be
13507 checked. Two new options -validity_period and -status_age added to
13512 * If signature or public key algorithm is unrecognized print out its
13513 OID rather that just UNKNOWN.
13517 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13518 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13519 ID to be generated from the issuer certificate alone which can then be
13520 passed to OCSP_id_issuer_cmp().
13524 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13525 ASN1 modules to export functions returning ASN1_ITEM pointers
13526 instead of the ASN1_ITEM structures themselves. This adds several
13527 new macros which allow the underlying ASN1 function/structure to
13528 be accessed transparently. As a result code should not use ASN1_ITEM
13529 references directly (such as &X509_it) but instead use the relevant
13530 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13531 use of the new ASN1 code on platforms where exporting structures
13532 is problematical (for example in shared libraries) but exporting
13533 functions returning pointers to structures is not.
13537 * Add support for overriding the generation of SSL/TLS session IDs.
13538 These callbacks can be registered either in an SSL_CTX or per SSL.
13539 The purpose of this is to allow applications to control, if they wish,
13540 the arbitrary values chosen for use as session IDs, particularly as it
13541 can be useful for session caching in multiple-server environments. A
13542 command-line switch for testing this (and any client code that wishes
13543 to use such a feature) has been added to "s_server".
13545 *Geoff Thorpe, Lutz Jaenicke*
13547 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13548 of the form `#if defined(...) || defined(...) || ...` and
13549 `#if !defined(...) && !defined(...) && ...`. This also avoids
13550 the growing number of special cases it was previously handling.
13554 * Make all configuration macros available for application by making
13555 sure they are available in opensslconf.h, by giving them names starting
13556 with `OPENSSL_` to avoid conflicts with other packages and by making
13557 sure e_os2.h will cover all platform-specific cases together with
13559 Additionally, it is now possible to define configuration/platform-
13560 specific names (called "system identities"). In the C code, these
13561 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13562 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13563 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13568 * New option -set_serial to 'req' and 'x509' this allows the serial
13569 number to use to be specified on the command line. Previously self
13570 signed certificates were hard coded with serial number 0 and the
13571 CA options of 'x509' had to use a serial number in a file which was
13576 * New options to 'ca' utility to support V2 CRL entry extensions.
13577 Currently CRL reason, invalidity date and hold instruction are
13578 supported. Add new CRL extensions to V3 code and some new objects.
13582 * New function EVP_CIPHER_CTX_set_padding() this is used to
13583 disable standard block padding (aka PKCS#5 padding) in the EVP
13584 API, which was previously mandatory. This means that the data is
13585 not padded in any way and so the total length much be a multiple
13586 of the block size, otherwise an error occurs.
13590 * Initial (incomplete) OCSP SSL support.
13594 * New function OCSP_parse_url(). This splits up a URL into its host,
13595 port and path components: primarily to parse OCSP URLs. New -url
13596 option to ocsp utility.
13600 * New nonce behavior. The return value of OCSP_check_nonce() now
13601 reflects the various checks performed. Applications can decide
13602 whether to tolerate certain situations such as an absent nonce
13603 in a response when one was present in a request: the ocsp application
13604 just prints out a warning. New function OCSP_add1_basic_nonce()
13605 this is to allow responders to include a nonce in a response even if
13606 the request is nonce-less.
13610 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13611 skipped when using openssl x509 multiple times on a single input file,
13612 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13616 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13617 set string type: to handle setting ASN1_TIME structures. Fix ca
13618 utility to correctly initialize revocation date of CRLs.
13622 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13623 the clients preferred ciphersuites and rather use its own preferences.
13624 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13625 Internet Explorer by ensuring unchanged hash method during stepup.
13626 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13630 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13631 to aes and add a new 'exist' option to print out symbols that don't
13636 * Additional options to ocsp utility to allow flags to be set and
13637 additional certificates supplied.
13641 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13642 OCSP client a number of certificate to only verify the response
13647 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13648 handle the new API. Currently only ECB, CBC modes supported. Add new
13651 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13652 Encryption Standard (AES) Ciphersuites for Transport Layer
13653 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13654 not enabled by default and were not part of the "ALL" ciphersuite
13655 alias because they were not yet official; they could be
13656 explicitly requested by specifying the "AESdraft" ciphersuite
13657 group alias. In the final release of OpenSSL 0.9.7, the group
13658 alias is called "AES" and is part of "ALL".)
13660 *Ben Laurie, Steve Henson, Bodo Moeller*
13662 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13663 request to response.
13667 * Functions for OCSP responders. OCSP_request_onereq_count(),
13668 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13669 extract information from a certificate request. OCSP_response_create()
13670 creates a response and optionally adds a basic response structure.
13671 OCSP_basic_add1_status() adds a complete single response to a basic
13672 response and returns the OCSP_SINGLERESP structure just added (to allow
13673 extensions to be included for example). OCSP_basic_add1_cert() adds a
13674 certificate to a basic response and OCSP_basic_sign() signs a basic
13675 response with various flags. New helper functions ASN1_TIME_check()
13676 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13677 (converts ASN1_TIME to GeneralizedTime).
13681 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13682 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13683 structure from a certificate. X509_pubkey_digest() digests the public_key
13684 contents: this is used in various key identifiers.
13688 * Make sk_sort() tolerate a NULL argument.
13690 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13692 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13693 passed by the function are trusted implicitly. If any of them signed the
13694 response then it is assumed to be valid and is not verified.
13698 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13699 to data. This was previously part of the PKCS7 ASN1 code. This
13700 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13701 *Steve Henson, reported by Kenneth R. Robinette
13702 <support@securenetterm.com>*
13704 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13705 routines: without these tracing memory leaks is very painful.
13706 Fix leaks in PKCS12 and PKCS7 routines.
13710 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13711 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13712 effectively meant GeneralizedTime would never be used. Now it
13713 is initialised to -1 but X509_time_adj() now has to check the value
13714 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13715 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13716 *Steve Henson, reported by Kenneth R. Robinette
13717 <support@securenetterm.com>*
13719 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13720 result in a zero length in the ASN1_INTEGER structure which was
13721 not consistent with the structure when d2i_ASN1_INTEGER() was used
13722 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13723 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13724 where it did not print out a minus for negative ASN1_INTEGER.
13728 * Add summary printout to ocsp utility. The various functions which
13729 convert status values to strings have been renamed to:
13730 OCSP_response_status_str(), OCSP_cert_status_str() and
13731 OCSP_crl_reason_str() and are no longer static. New options
13732 to verify nonce values and to disable verification. OCSP response
13733 printout format cleaned up.
13737 * Add additional OCSP certificate checks. These are those specified
13738 in RFC2560. This consists of two separate checks: the CA of the
13739 certificate being checked must either be the OCSP signer certificate
13740 or the issuer of the OCSP signer certificate. In the latter case the
13741 OCSP signer certificate must contain the OCSP signing extended key
13742 usage. This check is performed by attempting to match the OCSP
13743 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13744 in the OCSP_CERTID structures of the response.
13748 * Initial OCSP certificate verification added to OCSP_basic_verify()
13749 and related routines. This uses the standard OpenSSL certificate
13750 verify routines to perform initial checks (just CA validity) and
13751 to obtain the certificate chain. Then additional checks will be
13752 performed on the chain. Currently the root CA is checked to see
13753 if it is explicitly trusted for OCSP signing. This is used to set
13754 a root CA as a global signing root: that is any certificate that
13755 chains to that CA is an acceptable OCSP signing certificate.
13759 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13760 extensions from a separate configuration file.
13761 As when reading extensions from the main configuration file,
13762 the '-extensions ...' option may be used for specifying the
13765 *Massimiliano Pala <madwolf@comune.modena.it>*
13767 * New OCSP utility. Allows OCSP requests to be generated or
13768 read. The request can be sent to a responder and the output
13769 parsed, outputted or printed in text form. Not complete yet:
13770 still needs to check the OCSP response validity.
13774 * New subcommands for 'openssl ca':
13775 `openssl ca -status <serial>` prints the status of the cert with
13776 the given serial number (according to the index file).
13777 `openssl ca -updatedb` updates the expiry status of certificates
13780 *Massimiliano Pala <madwolf@comune.modena.it>*
13782 * New '-newreq-nodes' command option to CA.pl. This is like
13783 '-newreq', but calls 'openssl req' with the '-nodes' option
13784 so that the resulting key is not encrypted.
13786 *Damien Miller <djm@mindrot.org>*
13788 * New configuration for the GNU Hurd.
13790 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13792 * Initial code to implement OCSP basic response verify. This
13793 is currently incomplete. Currently just finds the signer's
13794 certificate and verifies the signature on the response.
13798 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13799 value of OPENSSLDIR. This is available via the new '-d' option
13800 to 'openssl version', and is also included in 'openssl version -a'.
13804 * Allowing defining memory allocation callbacks that will be given
13805 file name and line number information in additional arguments
13806 (a `const char*` and an int). The basic functionality remains, as
13807 well as the original possibility to just replace malloc(),
13808 realloc() and free() by functions that do not know about these
13809 additional arguments. To register and find out the current
13810 settings for extended allocation functions, the following
13811 functions are provided:
13813 CRYPTO_set_mem_ex_functions
13814 CRYPTO_set_locked_mem_ex_functions
13815 CRYPTO_get_mem_ex_functions
13816 CRYPTO_get_locked_mem_ex_functions
13818 These work the same way as CRYPTO_set_mem_functions and friends.
13819 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13820 extended allocation function is enabled.
13821 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13822 a conventional allocation function is enabled.
13824 *Richard Levitte, Bodo Moeller*
13826 * Finish off removing the remaining LHASH function pointer casts.
13827 There should no longer be any prototype-casting required when using
13828 the LHASH abstraction, and any casts that remain are "bugs". See
13829 the callback types and macros at the head of lhash.h for details
13830 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13834 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13835 If /dev/[u]random devices are not available or do not return enough
13836 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13838 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13839 /etc/entropy will be queried once each in this sequence, querying stops
13840 when enough entropy was collected without querying more sockets.
13844 * Change the Unix RAND_poll() variant to be able to poll several
13845 random devices, as specified by DEVRANDOM, until a sufficient amount
13846 of data has been collected. We spend at most 10 ms on each file
13847 (select timeout) and read in non-blocking mode. DEVRANDOM now
13848 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13849 (previously it was just the string "/dev/urandom"), so on typical
13850 platforms the 10 ms delay will never occur.
13851 Also separate out the Unix variant to its own file, rand_unix.c.
13852 For VMS, there's a currently-empty rand_vms.c.
13856 * Move OCSP client related routines to ocsp_cl.c. These
13857 provide utility functions which an application needing
13858 to issue a request to an OCSP responder and analyse the
13859 response will typically need: as opposed to those which an
13860 OCSP responder itself would need which will be added later.
13862 OCSP_request_sign() signs an OCSP request with an API similar
13863 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13864 response. OCSP_response_get1_basic() extracts basic response
13865 from response. OCSP_resp_find_status(): finds and extracts status
13866 information from an OCSP_CERTID structure (which will be created
13867 when the request structure is built). These are built from lower
13868 level functions which work on OCSP_SINGLERESP structures but
13869 won't normally be used unless the application wishes to examine
13870 extensions in the OCSP response for example.
13872 Replace nonce routines with a pair of functions.
13873 OCSP_request_add1_nonce() adds a nonce value and optionally
13874 generates a random value. OCSP_check_nonce() checks the
13875 validity of the nonce in an OCSP response.
13879 * Change function OCSP_request_add() to OCSP_request_add0_id().
13880 This doesn't copy the supplied OCSP_CERTID and avoids the
13881 need to free up the newly created id. Change return type
13882 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13883 This can then be used to add extensions to the request.
13884 Deleted OCSP_request_new(), since most of its functionality
13885 is now in OCSP_REQUEST_new() (and the case insensitive name
13886 clash) apart from the ability to set the request name which
13887 will be added elsewhere.
13891 * Update OCSP API. Remove obsolete extensions argument from
13892 various functions. Extensions are now handled using the new
13893 OCSP extension code. New simple OCSP HTTP function which
13894 can be used to send requests and parse the response.
13898 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13899 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13900 uses the special reorder version of SET OF to sort the attributes
13901 and reorder them to match the encoded order. This resolves a long
13902 standing problem: a verify on a PKCS7 structure just after signing
13903 it used to fail because the attribute order did not match the
13904 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13905 it uses the received order. This is necessary to tolerate some broken
13906 software that does not order SET OF. This is handled by encoding
13907 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13908 to produce the required SET OF.
13912 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13913 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13914 files to get correct declarations of the ASN.1 item variables.
13918 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13919 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13920 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13921 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13922 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13923 ASN1_ITEM and no wrapper functions.
13927 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13928 replace the old function pointer based I/O routines. Change most of
13929 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13933 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13934 lines, recognize more "algorithms" that can be deselected, and make
13935 it complain about algorithm deselection that isn't recognised.
13939 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13940 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13941 to use new functions. Add NO_ASN1_OLD which can be set to remove
13942 some old style ASN1 functions: this can be used to determine if old
13943 code will still work when these eventually go away.
13947 * New extension functions for OCSP structures, these follow the
13948 same conventions as certificates and CRLs.
13952 * New function X509V3_add1_i2d(). This automatically encodes and
13953 adds an extension. Its behaviour can be customised with various
13954 flags to append, replace or delete. Various wrappers added for
13955 certificates and CRLs.
13959 * Fix to avoid calling the underlying ASN1 print routine when
13960 an extension cannot be parsed. Correct a typo in the
13961 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13965 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13966 entries for variables.
13970 * Add functionality to `apps/openssl.c` for detecting locking
13971 problems: As the program is single-threaded, all we have
13972 to do is register a locking callback using an array for
13973 storing which locks are currently held by the program.
13977 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13978 SSL_get_ex_data_X509_STORE_idx(), which is used in
13979 ssl_verify_cert_chain() and thus can be called at any time
13980 during TLS/SSL handshakes so that thread-safety is essential.
13981 Unfortunately, the ex_data design is not at all suited
13982 for multi-threaded use, so it probably should be abolished.
13986 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13988 *Broadcom, tweaked and integrated by Geoff Thorpe*
13990 * Move common extension printing code to new function
13991 X509V3_print_extensions(). Reorganise OCSP print routines and
13992 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13996 * New function X509_signature_print() to remove duplication in some
14001 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14002 set (this was treated exactly the same as SET OF previously). This
14003 is used to reorder the STACK representing the structure to match the
14004 encoding. This will be used to get round a problem where a PKCS7
14005 structure which was signed could not be verified because the STACK
14006 order did not reflect the encoded order.
14010 * Reimplement the OCSP ASN1 module using the new code.
14014 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14015 for its ASN1 operations. The old style function pointers still exist
14016 for now but they will eventually go away.
14020 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14021 completely replaces the old ASN1 functionality with a table driven
14022 encoder and decoder which interprets an ASN1_ITEM structure describing
14023 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14024 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14025 has also been converted to the new form.
14029 * Change BN_mod_exp_recp so that negative moduli are tolerated
14030 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14031 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14032 for negative moduli.
14036 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14037 of not touching the result's sign bit.
14041 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14046 * Changed the LHASH code to use prototypes for callbacks, and created
14047 macros to declare and implement thin (optionally static) functions
14048 that provide type-safety and avoid function pointer casting for the
14049 type-specific callbacks.
14053 * Added Kerberos Cipher Suites to be used with TLS, as written in
14055 *Veers Staats <staatsvr@asc.hpc.mil>,
14056 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14058 * Reformat the FAQ so the different questions and answers can be divided
14059 in sections depending on the subject.
14063 * Have the zlib compression code load ZLIB.DLL dynamically under
14068 * New function BN_mod_sqrt for computing square roots modulo a prime
14069 (using the probabilistic Tonelli-Shanks algorithm unless
14070 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14071 be handled deterministically).
14073 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14075 * Make BN_mod_inverse faster by explicitly handling small quotients
14076 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14077 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14081 * New function BN_kronecker.
14085 * Fix BN_gcd so that it works on negative inputs; the result is
14086 positive unless both parameters are zero.
14087 Previously something reasonably close to an infinite loop was
14088 possible because numbers could be growing instead of shrinking
14089 in the implementation of Euclid's algorithm.
14093 * Fix BN_is_word() and BN_is_one() macros to take into account the
14094 sign of the number in question.
14096 Fix BN_is_word(a,w) to work correctly for w == 0.
14098 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14099 because its test if the absolute value of 'a' equals 'w'.
14100 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14101 it exists mostly for use in the implementations of BN_is_zero(),
14102 BN_is_one(), and BN_is_word().
14106 * New function BN_swap.
14110 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14111 the exponentiation functions are more likely to produce reasonable
14112 results on negative inputs.
14116 * Change BN_mod_mul so that the result is always non-negative.
14117 Previously, it could be negative if one of the factors was negative;
14118 I don't think anyone really wanted that behaviour.
14122 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14123 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14124 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14125 and add new functions:
14134 BN_mod_lshift1_quick
14136 BN_mod_lshift_quick
14138 These functions always generate non-negative results.
14140 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14141 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14143 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14144 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14145 be reduced modulo `m`.
14147 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14150 The following entry accidentally appeared in the CHANGES file
14151 distributed with OpenSSL 0.9.7. The modifications described in
14152 it do *not* apply to OpenSSL 0.9.7.
14154 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14155 was actually never needed) and in BN_mul(). The removal in BN_mul()
14156 required a small change in bn_mul_part_recursive() and the addition
14157 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14158 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14159 bn_sub_words() and bn_add_words() except they take arrays with
14165 * In 'openssl passwd', verify passwords read from the terminal
14166 unless the '-salt' option is used (which usually means that
14167 verification would just waste user's time since the resulting
14168 hash is going to be compared with some given password hash)
14169 or the new '-noverify' option is used.
14171 This is an incompatible change, but it does not affect
14172 non-interactive use of 'openssl passwd' (passwords on the command
14173 line, '-stdin' option, '-in ...' option) and thus should not
14174 cause any problems.
14178 * Remove all references to RSAref, since there's no more need for it.
14182 * Make DSO load along a path given through an environment variable
14183 (SHLIB_PATH) with shl_load().
14187 * Constify the ENGINE code as a result of BIGNUM constification.
14188 Also constify the RSA code and most things related to it. In a
14189 few places, most notable in the depth of the ASN.1 code, ugly
14190 casts back to non-const were required (to be solved at a later
14195 * Make it so the openssl application has all engines loaded by default.
14199 * Constify the BIGNUM routines a little more.
14203 * Add the following functions:
14205 ENGINE_load_cswift()
14207 ENGINE_load_atalla()
14208 ENGINE_load_nuron()
14209 ENGINE_load_builtin_engines()
14211 That way, an application can itself choose if external engines that
14212 are built-in in OpenSSL shall ever be used or not. The benefit is
14213 that applications won't have to be linked with libdl or other dso
14214 libraries unless it's really needed.
14216 Changed 'openssl engine' to load all engines on demand.
14217 Changed the engine header files to avoid the duplication of some
14218 declarations (they differed!).
14222 * 'openssl engine' can now list capabilities.
14226 * Better error reporting in 'openssl engine'.
14230 * Never call load_dh_param(NULL) in s_server.
14234 * Add engine application. It can currently list engines by name and
14235 identity, and test if they are actually available.
14239 * Improve RPM specification file by forcing symbolic linking and making
14240 sure the installed documentation is also owned by root.root.
14242 *Damien Miller <djm@mindrot.org>*
14244 * Give the OpenSSL applications more possibilities to make use of
14245 keys (public as well as private) handled by engines.
14249 * Add OCSP code that comes from CertCo.
14253 * Add VMS support for the Rijndael code.
14257 * Added untested support for Nuron crypto accelerator.
14261 * Add support for external cryptographic devices. This code was
14262 previously distributed separately as the "engine" branch.
14264 *Geoff Thorpe, Richard Levitte*
14266 * Rework the filename-translation in the DSO code. It is now possible to
14267 have far greater control over how a "name" is turned into a filename
14268 depending on the operating environment and any oddities about the
14269 different shared library filenames on each system.
14273 * Support threads on FreeBSD-elf in Configure.
14277 * Fix for SHA1 assembly problem with MASM: it produces
14278 warnings about corrupt line number information when assembling
14279 with debugging information. This is caused by the overlapping
14282 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14285 NCONF_get_number() has no error checking at all. As a replacement,
14286 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14287 promoted strongly. The old NCONF_get_number is kept around for
14288 binary backward compatibility.
14289 Make it possible for methods to load from something other than a BIO,
14290 by providing a function pointer that is given a name instead of a BIO.
14291 For example, this could be used to load configuration data from an
14296 * Fix for non blocking accept BIOs. Added new I/O special reason
14297 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14298 with non blocking I/O was not possible because no retry code was
14299 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14304 * Added the beginnings of Rijndael support.
14308 * Fix for bug in DirectoryString mask setting. Add support for
14309 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14310 to allow certificate printing to more controllable, additional
14311 'certopt' option to 'x509' to allow new printing options to be
14316 * Clean old EAY MD5 hack from e_os.h.
14320 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14322 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14323 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14325 *Joe Orton, Steve Henson*
14327 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14329 * Fix additional bug revealed by the NISCC test suite:
14331 Stop bug triggering large recursion when presented with
14332 certain ASN.1 tags ([CVE-2003-0851])
14336 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14338 * Fix various bugs revealed by running the NISCC test suite:
14340 Stop out of bounds reads in the ASN1 code when presented with
14341 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14343 If verify callback ignores invalid public key errors don't try to check
14344 certificate signature with the NULL public key.
14348 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14349 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14354 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14355 extra data after the compression methods not only for TLS 1.0
14356 but also for SSL 3.0 (as required by the specification).
14358 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14360 * Change X509_certificate_type() to mark the key as exported/exportable
14361 when it's 512 *bits* long, not 512 bytes.
14365 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14367 * Countermeasure against the Klima-Pokorny-Rosa extension of
14368 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14369 a protocol version number mismatch like a decryption error
14370 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14374 * Turn on RSA blinding by default in the default implementation
14375 to avoid a timing attack. Applications that don't want it can call
14376 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14377 They would be ill-advised to do so in most cases.
14379 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14381 * Change RSA blinding code so that it works when the PRNG is not
14382 seeded (in this case, the secret RSA exponent is abused as
14383 an unpredictable seed -- if it is not unpredictable, there
14384 is no point in blinding anyway). Make RSA blinding thread-safe
14385 by remembering the creator's thread ID in rsa->blinding and
14386 having all other threads use local one-time blinding factors
14387 (this requires more computation than sharing rsa->blinding, but
14388 avoids excessive locking; and if an RSA object is not shared
14389 between threads, blinding will still be very fast).
14393 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14395 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14396 via timing by performing a MAC computation even if incorrect
14397 block cipher padding has been found. This is a countermeasure
14398 against active attacks where the attacker has to distinguish
14399 between bad padding and a MAC verification error. ([CVE-2003-0078])
14401 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14402 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14403 Martin Vuagnoux (EPFL, Ilion)*
14405 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14407 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14408 memory from its contents. This is done with a counter that will
14409 place alternating values in each byte. This can be used to solve
14410 two issues: 1) the removal of calls to memset() by highly optimizing
14411 compilers, and 2) cleansing with other values than 0, since those can
14412 be read through on certain media, for example a swap space on disk.
14416 * Bugfix: client side session caching did not work with external caching,
14417 because the session->cipher setting was not restored when reloading
14418 from the external cache. This problem was masked, when
14419 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14420 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14424 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14425 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14427 *Zeev Lieber <zeev-l@yahoo.com>*
14429 * Undo an undocumented change introduced in 0.9.6e which caused
14430 repeated calls to OpenSSL_add_all_ciphers() and
14431 OpenSSL_add_all_digests() to be ignored, even after calling
14436 * Change the default configuration reader to deal with last line not
14437 being properly terminated.
14441 * Change X509_NAME_cmp() so it applies the special rules on handling
14442 DN values that are of type PrintableString, as well as RDNs of type
14443 emailAddress where the value has the type ia5String.
14445 *stefank@valicert.com via Richard Levitte*
14447 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14448 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14449 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14450 the bitwise-OR of the two for use by the majority of applications
14451 wanting this behaviour, and update the docs. The documented
14452 behaviour and actual behaviour were inconsistent and had been
14453 changing anyway, so this is more a bug-fix than a behavioural
14456 *Geoff Thorpe, diagnosed by Nadav Har'El*
14458 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14459 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14463 * Fix initialization code race conditions in
14464 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14465 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14466 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14467 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14468 ssl2_get_cipher_by_char(),
14469 ssl3_get_cipher_by_char().
14471 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14473 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14474 the cached sessions are flushed, as the remove_cb() might use ex_data
14475 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14476 (see [openssl.org #212]).
14478 *Geoff Thorpe, Lutz Jaenicke*
14480 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14481 length, instead of the encoding length to d2i_ASN1_OBJECT.
14485 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14487 * [In 0.9.6g-engine release:]
14488 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14490 *Lynn Gazis <lgazis@rainbow.com>*
14492 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14494 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14495 and get fix the header length calculation.
14496 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14497 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14499 * Use proper error handling instead of 'assertions' in buffer
14500 overflow checks added in 0.9.6e. This prevents DoS (the
14501 assertions could call abort()).
14503 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14505 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14507 * Add various sanity checks to asn1_get_length() to reject
14508 the ASN1 length bytes if they exceed sizeof(long), will appear
14509 negative or the content length exceeds the length of the
14512 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14514 * Fix cipher selection routines: ciphers without encryption had no flags
14515 for the cipher strength set and where therefore not handled correctly
14516 by the selection routines (PR #130).
14520 * Fix EVP_dsa_sha macro.
14525 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14526 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14527 that was added in OpenSSL 0.9.6d.
14529 As the countermeasure turned out to be incompatible with some
14530 broken SSL implementations, the new option is part of SSL_OP_ALL.
14531 SSL_OP_ALL is usually employed when compatibility with weird SSL
14532 implementations is desired (e.g. '-bugs' option to 's_client' and
14533 's_server'), so the new option is automatically set in many
14538 * Changes in security patch:
14540 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14541 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14542 Air Force Materiel Command, USAF, under agreement number
14545 * Add various sanity checks to asn1_get_length() to reject
14546 the ASN1 length bytes if they exceed sizeof(long), will appear
14547 negative or the content length exceeds the length of the
14548 supplied buffer. ([CVE-2002-0659])
14550 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14552 * Assertions for various potential buffer overflows, not known to
14553 happen in practice.
14555 *Ben Laurie (CHATS)*
14557 * Various temporary buffers to hold ASCII versions of integers were
14558 too small for 64 bit platforms. ([CVE-2002-0655])
14559 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14561 * Remote buffer overflow in SSL3 protocol - an attacker could
14562 supply an oversized session ID to a client. ([CVE-2002-0656])
14564 *Ben Laurie (CHATS)*
14566 * Remote buffer overflow in SSL2 protocol - an attacker could
14567 supply an oversized client master key. ([CVE-2002-0656])
14569 *Ben Laurie (CHATS)*
14571 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14573 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14574 encoded as NULL) with id-dsa-with-sha1.
14576 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14578 * Check various `X509_...()` return values in `apps/req.c`.
14580 *Nils Larsch <nla@trustcenter.de>*
14582 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14583 an end-of-file condition would erroneously be flagged, when the CRLF
14584 was just at the end of a processed block. The bug was discovered when
14585 processing data through a buffering memory BIO handing the data to a
14586 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14587 <ptsekov@syntrex.com> and Nedelcho Stanev.
14591 * Implement a countermeasure against a vulnerability recently found
14592 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14593 before application data chunks to avoid the use of known IVs
14594 with data potentially chosen by the attacker.
14598 * Fix length checks in ssl3_get_client_hello().
14602 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14603 to prevent ssl3_read_internal() from incorrectly assuming that
14604 ssl3_read_bytes() found application data while handshake
14605 processing was enabled when in fact s->s3->in_read_app_data was
14606 merely automatically cleared during the initial handshake.
14608 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14610 * Fix object definitions for Private and Enterprise: they were not
14611 recognized in their shortname (=lowercase) representation. Extend
14612 obj_dat.pl to issue an error when using undefined keywords instead
14613 of silently ignoring the problem (Svenning Sorensen
14614 <sss@sss.dnsalias.net>).
14618 * Fix DH_generate_parameters() so that it works for 'non-standard'
14619 generators, i.e. generators other than 2 and 5. (Previously, the
14620 code did not properly initialise the 'add' and 'rem' values to
14621 BN_generate_prime().)
14623 In the new general case, we do not insist that 'generator' is
14624 actually a primitive root: This requirement is rather pointless;
14625 a generator of the order-q subgroup is just as good, if not
14630 * Map new X509 verification errors to alerts. Discovered and submitted by
14631 Tom Wu <tom@arcot.com>.
14635 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14636 returning non-zero before the data has been completely received
14637 when using non-blocking I/O.
14639 *Bodo Moeller; problem pointed out by John Hughes*
14641 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14643 *Ben Laurie, Lutz Jaenicke*
14645 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14646 Yoram Zahavi <YoramZ@gilian.com>).
14650 * Add information about CygWin 1.3 and on, and preserve proper
14651 configuration for the versions before that.
14653 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14655 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14656 check whether we deal with a copy of a session and do not delete from
14657 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14658 <izhar@checkpoint.com>.
14662 * Do not store session data into the internal session cache, if it
14663 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14664 flag is set). Proposed by Aslam <aslam@funk.com>.
14668 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14673 * [In 0.9.6d-engine release:]
14674 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14676 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14678 * Add the configuration target linux-s390x.
14680 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14682 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14683 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14684 variable as an indication that a ClientHello message has been
14685 received. As the flag value will be lost between multiple
14686 invocations of ssl3_accept when using non-blocking I/O, the
14687 function may not be aware that a handshake has actually taken
14688 place, thus preventing a new session from being added to the
14691 To avoid this problem, we now set s->new_session to 2 instead of
14692 using a local variable.
14694 *Lutz Jaenicke, Bodo Moeller*
14696 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14697 if the SSL_R_LENGTH_MISMATCH error is detected.
14699 *Geoff Thorpe, Bodo Moeller*
14701 * New 'shared_ldflag' column in Configure platform table.
14705 * Fix EVP_CIPHER_mode macro.
14707 *"Dan S. Camper" <dan@bti.net>*
14709 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14710 type, we must throw them away by setting rr->length to 0.
14712 *D P Chang <dpc@qualys.com>*
14714 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14716 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14717 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14718 worked incorrectly for those cases where range = `10..._2` and
14719 `3*range` is two bits longer than range.)
14723 * Only add signing time to PKCS7 structures if it is not already
14728 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14729 OBJ_ld_ce should be OBJ_id_ce.
14730 Also some ip-pda OIDs in crypto/objects/objects.txt were
14731 incorrect (cf. RFC 3039).
14733 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14735 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14736 returns early because it has nothing to do.
14738 *Andy Schneider <andy.schneider@bjss.co.uk>*
14740 * [In 0.9.6c-engine release:]
14741 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14743 *Andy Schneider <andy.schneider@bjss.co.uk>*
14745 * [In 0.9.6c-engine release:]
14746 Add support for Cryptographic Appliance's keyserver technology.
14747 (Use engine 'keyclient')
14749 *Cryptographic Appliances and Geoff Thorpe*
14751 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14752 is called via tools/c89.sh because arguments have to be
14753 rearranged (all '-L' options must appear before the first object
14756 *Richard Shapiro <rshapiro@abinitio.com>*
14758 * [In 0.9.6c-engine release:]
14759 Add support for Broadcom crypto accelerator cards, backported
14762 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14764 * [In 0.9.6c-engine release:]
14765 Add support for SureWare crypto accelerator cards from
14766 Baltimore Technologies. (Use engine 'sureware')
14768 *Baltimore Technologies and Mark Cox*
14770 * [In 0.9.6c-engine release:]
14771 Add support for crypto accelerator cards from Accelerated
14772 Encryption Processing, www.aep.ie. (Use engine 'aep')
14774 *AEP Inc. and Mark Cox*
14776 * Add a configuration entry for gcc on UnixWare.
14778 *Gary Benson <gbenson@redhat.com>*
14780 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14781 messages are stored in a single piece (fixed-length part and
14782 variable-length part combined) and fix various bugs found on the way.
14786 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14787 instead. BIO_gethostbyname() does not know what timeouts are
14788 appropriate, so entries would stay in cache even when they have
14790 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14792 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14793 faced with a pathologically small ClientHello fragment that does
14794 not contain client_version: Instead of aborting with an error,
14795 simply choose the highest available protocol version (i.e.,
14796 TLS 1.0 unless it is disabled). In practice, ClientHello
14797 messages are never sent like this, but this change gives us
14798 strictly correct behaviour at least for TLS.
14802 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14803 never resets s->method to s->ctx->method when called from within
14804 one of the SSL handshake functions.
14806 *Bodo Moeller; problem pointed out by Niko Baric*
14808 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14809 (sent using the client's version number) if client_version is
14810 smaller than the protocol version in use. Also change
14811 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14812 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14813 the client will at least see that alert.
14817 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14822 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14823 client receives HelloRequest while in a handshake.
14825 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14827 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14828 should end in 'break', not 'goto end' which circumvents various
14829 cleanups done in state SSL_ST_OK. But session related stuff
14830 must be disabled for SSL_ST_OK in the case that we just sent a
14833 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14834 before just sending a HelloRequest.
14836 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14838 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14839 reveal whether illegal block cipher padding was found or a MAC
14840 verification error occurred. (Neither SSLerr() codes nor alerts
14841 are directly visible to potential attackers, but the information
14842 may leak via logfiles.)
14844 Similar changes are not required for the SSL 2.0 implementation
14845 because the number of padding bytes is sent in clear for SSL 2.0,
14846 and the extra bytes are just ignored. However ssl/s2_pkt.c
14847 failed to verify that the purported number of padding bytes is in
14852 * Add OpenUNIX-8 support including shared libraries
14853 (Boyd Lynn Gerber <gerberb@zenez.com>).
14857 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14858 'wristwatch attack' using huge encoding parameters (cf.
14859 James H. Manger's CRYPTO 2001 paper). Note that the
14860 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14861 encoding parameters and hence was not vulnerable.
14865 * BN_sqr() bug fix.
14867 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14869 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14870 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14871 followed by modular reduction.
14873 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14875 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14876 equivalent based on BN_pseudo_rand() instead of BN_rand().
14880 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14881 This function was broken, as the check for a new client hello message
14882 to handle SGC did not allow these large messages.
14883 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14887 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14891 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14892 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14896 * Rework the configuration and shared library support for Tru64 Unix.
14897 The configuration part makes use of modern compiler features and
14898 still retains old compiler behavior for those that run older versions
14899 of the OS. The shared library support part includes a variant that
14900 uses the RPATH feature, and is available through the special
14901 configuration target "alpha-cc-rpath", which will never be selected
14904 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14906 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14907 with the same message size as in ssl3_get_certificate_request().
14908 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14909 messages might inadvertently be reject as too long.
14911 *Petr Lampa <lampa@fee.vutbr.cz>*
14913 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14917 * Modified SSL library such that the verify_callback that has been set
14918 specifically for an SSL object with SSL_set_verify() is actually being
14919 used. Before the change, a verify_callback set with this function was
14920 ignored and the verify_callback() set in the SSL_CTX at the time of
14921 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14922 to allow the necessary settings.
14926 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14927 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14928 done automatically (in contradiction to the requirements of the C
14929 standard). This made problems when used from OpenSSH.
14933 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14934 dh->length and always used
14936 BN_rand_range(priv_key, dh->p).
14938 BN_rand_range() is not necessary for Diffie-Hellman, and this
14939 specific range makes Diffie-Hellman unnecessarily inefficient if
14940 dh->length (recommended exponent length) is much smaller than the
14941 length of dh->p. We could use BN_rand_range() if the order of
14942 the subgroup was stored in the DH structure, but we only have
14947 BN_rand(priv_key, l, ...)
14949 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14956 RSA_eay_public_encrypt
14957 RSA_eay_private_decrypt
14958 RSA_eay_private_encrypt (signing)
14959 RSA_eay_public_decrypt (signature verification)
14961 (default implementations for RSA_public_encrypt,
14962 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14963 always reject numbers >= n.
14967 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14968 to synchronize access to 'locking_thread'. This is necessary on
14969 systems where access to 'locking_thread' (an 'unsigned long'
14970 variable) is not atomic.
14974 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14975 *before* setting the 'crypto_lock_rand' flag. The previous code had
14976 a race condition if 0 is a valid thread ID.
14978 *Travis Vitek <vitek@roguewave.com>*
14980 * Add support for shared libraries under Irix.
14982 *Albert Chin-A-Young <china@thewrittenword.com>*
14984 * Add configuration option to build on Linux on both big-endian and
14985 little-endian MIPS.
14987 *Ralf Baechle <ralf@uni-koblenz.de>*
14989 * Add the possibility to create shared libraries on HP-UX.
14993 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14995 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14996 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14997 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14998 PRNG state recovery was possible based on the output of
14999 one PRNG request appropriately sized to gain knowledge on
15000 'md' followed by enough consecutive 1-byte PRNG requests
15001 to traverse all of 'state'.
15003 1. When updating 'md_local' (the current thread's copy of 'md')
15004 during PRNG output generation, hash all of the previous
15005 'md_local' value, not just the half used for PRNG output.
15007 2. Make the number of bytes from 'state' included into the hash
15008 independent from the number of PRNG bytes requested.
15010 The first measure alone would be sufficient to avoid
15011 Markku-Juhani's attack. (Actually it had never occurred
15012 to me that the half of 'md_local' used for chaining was the
15013 half from which PRNG output bytes were taken -- I had always
15014 assumed that the secret half would be used.) The second
15015 measure makes sure that additional data from 'state' is never
15016 mixed into 'md_local' in small portions; this heuristically
15017 further strengthens the PRNG.
15021 * Fix crypto/bn/asm/mips3.s.
15025 * When only the key is given to "enc", the IV is undefined. Print out
15026 an error message in this case.
15030 * Handle special case when X509_NAME is empty in X509 printing routines.
15034 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15035 positive and less than q.
15039 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15040 used: it isn't thread safe and the add_lock_callback should handle
15043 *Paul Rose <Paul.Rose@bridge.com>*
15045 * Verify that incoming data obeys the block size in
15046 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15052 *Ulf Möller, Bodo Möller*
15054 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15055 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15056 when fixing the server behaviour for backwards-compatible 'client
15057 hello' messages. (Note that the attack is impractical against
15058 SSL 3.0 and TLS 1.0 anyway because length and version checking
15059 means that the probability of guessing a valid ciphertext is
15060 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15063 Before 0.9.5, the countermeasure (hide the error by generating a
15064 random 'decryption result') did not work properly because
15065 ERR_clear_error() was missing, meaning that SSL_get_error() would
15066 detect the supposedly ignored error.
15068 Both problems are now fixed.
15072 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15073 (previously it was 1024).
15077 * Fix for compatibility mode trust settings: ignore trust settings
15078 unless some valid trust or reject settings are present.
15082 * Fix for blowfish EVP: its a variable length cipher.
15086 * Fix various bugs related to DSA S/MIME verification. Handle missing
15087 parameters in DSA public key structures and return an error in the
15088 DSA routines if parameters are absent.
15092 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15093 in the current directory if neither $RANDFILE nor $HOME was set.
15094 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15095 caused some confusion to Windows users who haven't defined $HOME.
15096 Thus RAND_file_name() is changed again: e_os.h can define a
15097 DEFAULT_HOME, which will be used if $HOME is not set.
15098 For Windows, we use "C:"; on other platforms, we still require
15099 environment variables.
15101 * Move 'if (!initialized) RAND_poll()' into regions protected by
15102 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15103 having multiple threads call RAND_poll() concurrently.
15107 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15108 combination of a flag and a thread ID variable.
15109 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15110 flag), *other* threads can enter ssleay_add_bytes without obeying
15111 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15112 that they do not hold after the first thread unsets add_do_not_lock).
15116 * Change bctest again: '-x' expressions are not available in all
15117 versions of 'test'.
15121 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15123 * Fix a couple of memory leaks in PKCS7_dataDecode()
15125 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15127 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15128 the default extension for executables, if any. Also, make the perl
15129 scripts that use symlink() to test if it really exists and use "cp"
15130 if it doesn't. All this made OpenSSL compilable and installable in
15135 * Fix for asn1_GetSequence() for indefinite length constructed data.
15136 If SEQUENCE is length is indefinite just set c->slen to the total
15137 amount of data available.
15139 *Steve Henson, reported by shige@FreeBSD.org*
15141 *This change does not apply to 0.9.7.*
15143 * Change bctest to avoid here-documents inside command substitution
15144 (workaround for FreeBSD /bin/sh bug).
15145 For compatibility with Ultrix, avoid shell functions (introduced
15146 in the bctest version that searches along $PATH).
15150 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15151 with des_encrypt() defined on some operating systems, like Solaris
15156 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15157 On the Importance of Eliminating Errors in Cryptographic
15158 Computations, J. Cryptology 14 (2001) 2, 101-119,
15159 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15163 * MIPS assembler BIGNUM division bug fix.
15167 * Disabled incorrect Alpha assembler code.
15171 * Fix PKCS#7 decode routines so they correctly update the length
15172 after reading an EOC for the EXPLICIT tag.
15176 *This change does not apply to 0.9.7.*
15178 * Fix bug in PKCS#12 key generation routines. This was triggered
15179 if a 3DES key was generated with a 0 initial byte. Include
15180 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15181 (but broken) behaviour.
15185 * Enhance bctest to search for a working bc along $PATH and print
15188 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15190 * Fix memory leaks in err.c: free err_data string if necessary;
15191 don't write to the wrong index in ERR_set_error_data.
15195 * Implement ssl23_peek (analogous to ssl23_read), which previously
15200 * Replace rdtsc with `_emit` statements for VC++ version 5.
15202 *Jeremy Cooper <jeremy@baymoo.org>*
15204 * Make it possible to reuse SSLv2 sessions.
15208 * In copy_email() check for >= 0 as a return value for
15209 X509_NAME_get_index_by_NID() since 0 is a valid index.
15211 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15213 * Avoid coredump with unsupported or invalid public keys by checking if
15214 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15215 PKCS7_verify() fails with non detached data.
15219 * Don't use getenv in library functions when run as setuid/setgid.
15220 New function OPENSSL_issetugid().
15224 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15225 due to incorrect handling of multi-threading:
15227 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15229 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15231 3. Count how many times MemCheck_off() has been called so that
15232 nested use can be treated correctly. This also avoids
15233 inband-signalling in the previous code (which relied on the
15234 assumption that thread ID 0 is impossible).
15238 * Add "-rand" option also to s_client and s_server.
15242 * Fix CPU detection on Irix 6.x.
15243 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15244 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15246 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15251 *This change does not apply to 0.9.7.*
15253 * Use the cached encoding of an X509_NAME structure rather than
15254 copying it. This is apparently the reason for the libsafe "errors"
15255 but the code is actually correct.
15259 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15260 Bleichenbacher's DSA attack.
15261 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15262 to be set and top=0 forces the highest bit to be set; top=-1 is new
15263 and leaves the highest bit random.
15265 *Ulf Moeller, Bodo Moeller*
15267 * In the `NCONF_...`-based implementations for `CONF_...` queries
15268 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15269 a temporary CONF structure with the data component set to NULL
15270 (which gives segmentation faults in lh_retrieve).
15271 Instead, use NULL for the CONF pointer in CONF_get_string and
15272 CONF_get_number (which may use environment variables) and directly
15273 return NULL from CONF_get_section.
15277 * Fix potential buffer overrun for EBCDIC.
15281 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15282 keyUsage if basicConstraints absent for a CA.
15286 * Make SMIME_write_PKCS7() write mail header values with a format that
15287 is more generally accepted (no spaces before the semicolon), since
15288 some programs can't parse those values properly otherwise. Also make
15289 sure BIO's that break lines after each write do not create invalid
15294 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15295 macros previously used would not encode an empty SEQUENCE OF
15296 and break the signature.
15300 *This change does not apply to 0.9.7.*
15302 * Zero the premaster secret after deriving the master secret in
15307 * Add some EVP_add_digest_alias registrations (as found in
15308 OpenSSL_add_all_digests()) to SSL_library_init()
15309 aka OpenSSL_add_ssl_algorithms(). This provides improved
15310 compatibility with peers using X.509 certificates
15311 with unconventional AlgorithmIdentifier OIDs.
15315 * Fix for Irix with NO_ASM.
15317 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15319 * ./config script fixes.
15321 *Ulf Moeller, Richard Levitte*
15323 * Fix 'openssl passwd -1'.
15327 * Change PKCS12_key_gen_asc() so it can cope with non null
15328 terminated strings whose length is passed in the passlen
15329 parameter, for example from PEM callbacks. This was done
15330 by adding an extra length parameter to asc2uni().
15332 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15334 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15335 call failed, free the DSA structure.
15339 * Fix to uni2asc() to cope with zero length Unicode strings.
15340 These are present in some PKCS#12 files.
15344 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15345 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15346 when writing a 32767 byte record.
15348 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15350 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15351 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15353 (RSA objects have a reference count access to which is protected
15354 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15355 so they are meant to be shared between threads.)
15356 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15357 "Reddie, Steven" <Steven.Reddie@ca.com>*
15359 * Fix a deadlock in CRYPTO_mem_leaks().
15363 * Use better test patterns in bntest.
15367 * rand_win.c fix for Borland C.
15371 * BN_rshift bugfix for n == 0.
15375 * Add a 'bctest' script that checks for some known 'bc' bugs
15376 so that 'make test' does not abort just because 'bc' is broken.
15380 * Store verify_result within SSL_SESSION also for client side to
15381 avoid potential security hole. (Re-used sessions on the client side
15382 always resulted in verify_result==X509_V_OK, not using the original
15383 result of the server certificate verification.)
15387 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15388 SSL3_RT_APPLICATION_DATA, return 0.
15389 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15394 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15395 releases, have been re-implemented by renaming the previous
15396 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15397 and ssl3_read_internal, respectively, and adding 'peek' parameters
15398 to them. The new ssl[23]_{read,peek} functions are calls to
15399 ssl[23]_read_internal with the 'peek' flag set appropriately.
15400 A 'peek' parameter has also been added to ssl3_read_bytes, which
15401 does the actual work for ssl3_read_internal.
15405 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15406 the method-specific "init()" handler. Also clean up ex_data after
15407 calling the method-specific "finish()" handler. Previously, this was
15408 happening the other way round.
15412 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15413 The previous value, 12, was not always sufficient for BN_mod_exp().
15417 * Make sure that shared libraries get the internal name engine with
15418 the full version number and not just 0. This should mark the
15419 shared libraries as not backward compatible. Of course, this should
15420 be changed again when we can guarantee backward binary compatibility.
15424 * Fix typo in get_cert_by_subject() in by_dir.c
15426 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15428 * Rework the system to generate shared libraries:
15430 - Make note of the expected extension for the shared libraries and
15431 if there is a need for symbolic links from for example libcrypto.so.0
15432 to libcrypto.so.0.9.7. There is extended info in Configure for
15435 - Make as few rebuilds of the shared libraries as possible.
15437 - Still avoid linking the OpenSSL programs with the shared libraries.
15439 - When installing, install the shared libraries separately from the
15444 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15446 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15447 and not in SSL_clear because the latter is also used by the
15448 accept/connect functions; previously, the settings made by
15449 SSL_set_read_ahead would be lost during the handshake.
15451 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15453 * Correct util/mkdef.pl to be selective about disabled algorithms.
15454 Previously, it would create entries for disabled algorithms no
15459 * Added several new manual pages for SSL_* function.
15463 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15465 * In ssl23_get_client_hello, generate an error message when faced
15466 with an initial SSL 3.0/TLS record that is too small to contain the
15467 first two bytes of the ClientHello message, i.e. client_version.
15468 (Note that this is a pathologic case that probably has never happened
15469 in real life.) The previous approach was to use the version number
15470 from the record header as a substitute; but our protocol choice
15471 should not depend on that one because it is not authenticated
15472 by the Finished messages.
15476 * More robust randomness gathering functions for Windows.
15478 *Jeffrey Altman <jaltman@columbia.edu>*
15480 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15481 not set then we don't setup the error code for issuer check errors
15482 to avoid possibly overwriting other errors which the callback does
15483 handle. If an application does set the flag then we assume it knows
15484 what it is doing and can handle the new informational codes
15489 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15490 a general "ANY" type, as such it should be able to decode anything
15491 including tagged types. However it didn't check the class so it would
15492 wrongly interpret tagged types in the same way as their universal
15493 counterpart and unknown types were just rejected. Changed so that the
15494 tagged and unknown types are handled in the same way as a SEQUENCE:
15495 that is the encoding is stored intact. There is also a new type
15496 "V_ASN1_OTHER" which is used when the class is not universal, in this
15497 case we have no idea what the actual type is so we just lump them all
15502 * On VMS, stdout may very well lead to a file that is written to
15503 in a record-oriented fashion. That means that every write() will
15504 write a separate record, which will be read separately by the
15505 programs trying to read from it. This can be very confusing.
15507 The solution is to put a BIO filter in the way that will buffer
15508 text until a linefeed is reached, and then write everything a
15509 line at a time, so every record written will be an actual line,
15510 not chunks of lines and not (usually doesn't happen, but I've
15511 seen it once) several lines in one record. BIO_f_linebuffer() is
15514 Currently, it's a VMS-only method, because that's where it has
15515 been tested well enough.
15519 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15520 it can return incorrect results.
15521 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15522 but it was in 0.9.6-beta[12].)
15526 * Disable the check for content being present when verifying detached
15527 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15528 include zero length content when signing messages.
15532 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15533 BIO_ctrl (for BIO pairs).
15537 * Add DSO method for VMS.
15541 * Bug fix: Montgomery multiplication could produce results with the
15546 * Add RPM specification openssl.spec and modify it to build three
15547 packages. The default package contains applications, application
15548 documentation and run-time libraries. The devel package contains
15549 include files, static libraries and function documentation. The
15550 doc package contains the contents of the doc directory. The original
15551 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15555 * Add a large number of documentation files for many SSL routines.
15557 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15559 * Add a configuration entry for Sony News 4.
15561 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15563 * Don't set the two most significant bits to one when generating a
15564 random number < q in the DSA library.
15568 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15569 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15570 the underlying transport is blocking) if a handshake took place.
15571 (The default behaviour is needed by applications such as s_client
15572 and s_server that use select() to determine when to use SSL_read;
15573 but for applications that know in advance when to expect data, it
15574 just makes things more complicated.)
15578 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15583 * Add a few more EBCDIC conditionals that make `req` and `x509`
15584 work better on such systems.
15586 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15588 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15589 Update PKCS12_parse() so it copies the friendlyName and the
15590 keyid to the certificates aux info.
15594 * Fix bug in PKCS7_verify() which caused an infinite loop
15595 if there was more than one signature.
15597 *Sven Uszpelkat <su@celocom.de>*
15599 * Major change in util/mkdef.pl to include extra information
15600 about each symbol, as well as presenting variables as well
15601 as functions. This change means that there's n more need
15602 to rebuild the .num files when some algorithms are excluded.
15606 * Allow the verify time to be set by an application,
15607 rather than always using the current time.
15611 * Phase 2 verify code reorganisation. The certificate
15612 verify code now looks up an issuer certificate by a
15613 number of criteria: subject name, authority key id
15614 and key usage. It also verifies self signed certificates
15615 by the same criteria. The main comparison function is
15616 X509_check_issued() which performs these checks.
15618 Lot of changes were necessary in order to support this
15619 without completely rewriting the lookup code.
15621 Authority and subject key identifier are now cached.
15623 The LHASH 'certs' is X509_STORE has now been replaced
15624 by a STACK_OF(X509_OBJECT). This is mainly because an
15625 LHASH can't store or retrieve multiple objects with
15626 the same hash value.
15628 As a result various functions (which were all internal
15629 use only) have changed to handle the new X509_STORE
15630 structure. This will break anything that messed round
15631 with X509_STORE internally.
15633 The functions X509_STORE_add_cert() now checks for an
15634 exact match, rather than just subject name.
15636 The X509_STORE API doesn't directly support the retrieval
15637 of multiple certificates matching a given criteria, however
15638 this can be worked round by performing a lookup first
15639 (which will fill the cache with candidate certificates)
15640 and then examining the cache for matches. This is probably
15641 the best we can do without throwing out X509_LOOKUP
15642 entirely (maybe later...).
15644 The X509_VERIFY_CTX structure has been enhanced considerably.
15646 All certificate lookup operations now go via a get_issuer()
15647 callback. Although this currently uses an X509_STORE it
15648 can be replaced by custom lookups. This is a simple way
15649 to bypass the X509_STORE hackery necessary to make this
15650 work and makes it possible to use more efficient techniques
15651 in future. A very simple version which uses a simple
15652 STACK for its trusted certificate store is also provided
15653 using X509_STORE_CTX_trusted_stack().
15655 The verify_cb() and verify() callbacks now have equivalents
15656 in the X509_STORE_CTX structure.
15658 X509_STORE_CTX also has a 'flags' field which can be used
15659 to customise the verify behaviour.
15663 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15664 excludes S/MIME capabilities.
15668 * When a certificate request is read in keep a copy of the
15669 original encoding of the signed data and use it when outputting
15670 again. Signatures then use the original encoding rather than
15671 a decoded, encoded version which may cause problems if the
15672 request is improperly encoded.
15676 * For consistency with other BIO_puts implementations, call
15677 buffer_write(b, ...) directly in buffer_puts instead of calling
15680 In BIO_puts, increment b->num_write as in BIO_write.
15682 *Peter.Sylvester@EdelWeb.fr*
15684 * Fix BN_mul_word for the case where the word is 0. (We have to use
15685 BN_zero, we may not return a BIGNUM with an array consisting of
15686 words set to zero.)
15690 * Avoid calling abort() from within the library when problems are
15691 detected, except if preprocessor symbols have been defined
15692 (such as REF_CHECK, BN_DEBUG etc.).
15696 * New openssl application 'rsautl'. This utility can be
15697 used for low-level RSA operations. DER public key
15698 BIO/fp routines also added.
15702 * New Configure entry and patches for compiling on QNX 4.
15704 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15706 * A demo state-machine implementation was sponsored by
15707 Nuron (<http://www.nuron.com/>) and is now available in
15708 demos/state_machine.
15712 * New options added to the 'dgst' utility for signature
15713 generation and verification.
15717 * Unrecognized PKCS#7 content types are now handled via a
15718 catch all ASN1_TYPE structure. This allows unsupported
15719 types to be stored as a "blob" and an application can
15720 encode and decode it manually.
15724 * Fix various signed/unsigned issues to make a_strex.c
15725 compile under VC++.
15727 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15729 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15730 length if passed a buffer. ASN1_INTEGER_to_BN failed
15731 if passed a NULL BN and its argument was negative.
15733 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15735 * Modification to PKCS#7 encoding routines to output definite
15736 length encoding. Since currently the whole structures are in
15737 memory there's not real point in using indefinite length
15738 constructed encoding. However if OpenSSL is compiled with
15739 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15743 * Added BIO_vprintf() and BIO_vsnprintf().
15747 * Added more prefixes to parse for in the strings written
15748 through a logging bio, to cover all the levels that are available
15749 through syslog. The prefixes are now:
15751 PANIC, EMERG, EMR => LOG_EMERG
15752 ALERT, ALR => LOG_ALERT
15753 CRIT, CRI => LOG_CRIT
15754 ERROR, ERR => LOG_ERR
15755 WARNING, WARN, WAR => LOG_WARNING
15756 NOTICE, NOTE, NOT => LOG_NOTICE
15757 INFO, INF => LOG_INFO
15758 DEBUG, DBG => LOG_DEBUG
15760 and as before, if none of those prefixes are present at the
15761 beginning of the string, LOG_ERR is chosen.
15763 On Win32, the `LOG_*` levels are mapped according to this:
15765 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15766 LOG_WARNING => EVENTLOG_WARNING_TYPE
15767 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15771 * Made it possible to reconfigure with just the configuration
15772 argument "reconf" or "reconfigure". The command line arguments
15773 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15774 and are retrieved from there when reconfiguring.
15780 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15782 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15786 * The obj_dat.pl script was messing up the sorting of object
15787 names. The reason was that it compared the quoted version
15788 of strings as a result "OCSP" > "OCSP Signing" because
15789 " > SPACE. Changed script to store unquoted versions of
15790 names and add quotes on output. It was also omitting some
15791 names from the lookup table if they were given a default
15792 value (that is if SN is missing it is given the same
15793 value as LN and vice versa), these are now added on the
15794 grounds that if an object has a name we should be able to
15795 look it up. Finally added warning output when duplicate
15796 short or long names are found.
15800 * Changes needed for Tandem NSK.
15802 *Scott Uroff <scott@xypro.com>*
15804 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15805 RSA_padding_check_SSLv23(), special padding was never detected
15806 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15807 version rollback attacks was not effective.
15809 In s23_clnt.c, don't use special rollback-attack detection padding
15810 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15811 client; similarly, in s23_srvr.c, don't do the rollback check if
15812 SSL 2.0 is the only protocol enabled in the server.
15816 * Make it possible to get hexdumps of unprintable data with 'openssl
15817 asn1parse'. By implication, the functions ASN1_parse_dump() and
15818 BIO_dump_indent() are added.
15822 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15823 these print out strings and name structures based on various
15824 flags including RFC2253 support and proper handling of
15825 multibyte characters. Added options to the 'x509' utility
15826 to allow the various flags to be set.
15830 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15831 Also change the functions X509_cmp_current_time() and
15832 X509_gmtime_adj() work with an ASN1_TIME structure,
15833 this will enable certificates using GeneralizedTime in validity
15834 dates to be checked.
15838 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15839 negative public key encodings) on by default,
15840 NO_NEG_PUBKEY_BUG can be set to disable it.
15844 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15845 content octets. An i2c_ASN1_OBJECT is unnecessary because
15846 the encoding can be trivially obtained from the structure.
15850 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15851 not read locks (`CRYPTO_r_[un]lock`).
15855 * A first attempt at creating official support for shared
15856 libraries through configuration. I've kept it so the
15857 default is static libraries only, and the OpenSSL programs
15858 are always statically linked for now, but there are
15859 preparations for dynamic linking in place.
15860 This has been tested on Linux and Tru64.
15864 * Randomness polling function for Win9x, as described in:
15865 Peter Gutmann, Software Generation of Practically Strong
15870 * Fix so PRNG is seeded in req if using an already existing
15875 * New options to smime application. -inform and -outform
15876 allow alternative formats for the S/MIME message including
15877 PEM and DER. The -content option allows the content to be
15878 specified separately. This should allow things like Netscape
15879 form signing output easier to verify.
15883 * Fix the ASN1 encoding of tags using the 'long form'.
15887 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15888 STRING types. These convert content octets to and from the
15889 underlying type. The actual tag and length octets are
15890 already assumed to have been read in and checked. These
15891 are needed because all other string types have virtually
15892 identical handling apart from the tag. By having versions
15893 of the ASN1 functions that just operate on content octets
15894 IMPLICIT tagging can be handled properly. It also allows
15895 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15896 and ASN1_INTEGER are identical apart from the tag.
15900 * Change the handling of OID objects as follows:
15902 - New object identifiers are inserted in objects.txt, following
15903 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15904 - objects.pl is used to process obj_mac.num and create a new
15906 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15909 This is currently kind of a hack, and the perl code in objects.pl
15910 isn't very elegant, but it works as I intended. The simplest way
15911 to check that it worked correctly is to look in obj_dat.h and
15912 check the array nid_objs and make sure the objects haven't moved
15913 around (this is important!). Additions are OK, as well as
15914 consistent name changes.
15918 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15922 * Addition of the command line parameter '-rand file' to 'openssl req'.
15923 The given file adds to whatever has already been seeded into the
15924 random pool through the RANDFILE configuration file option or
15925 environment variable, or the default random state file.
15929 * mkstack.pl now sorts each macro group into lexical order.
15930 Previously the output order depended on the order the files
15931 appeared in the directory, resulting in needless rewriting
15936 * Patches to make OpenSSL compile under Win32 again. Mostly
15937 work arounds for the VC++ problem that it treats func() as
15938 func(void). Also stripped out the parts of mkdef.pl that
15939 added extra typesafe functions: these no longer exist.
15943 * Reorganisation of the stack code. The macros are now all
15944 collected in safestack.h . Each macro is defined in terms of
15945 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15946 DEBUG_SAFESTACK is now handled in terms of function casts,
15947 this has the advantage of retaining type safety without the
15948 use of additional functions. If DEBUG_SAFESTACK is not defined
15949 then the non typesafe macros are used instead. Also modified the
15950 mkstack.pl script to handle the new form. Needs testing to see
15951 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15952 the default if no major problems. Similar behaviour for ASN1_SET_OF
15953 and PKCS12_STACK_OF.
15957 * When some versions of IIS use the 'NET' form of private key the
15958 key derivation algorithm is different. Normally MD5(password) is
15959 used as a 128 bit RC4 key. In the modified case
15960 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15961 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15962 as the old Netscape_RSA functions except they have an additional
15963 'sgckey' parameter which uses the modified algorithm. Also added
15964 an -sgckey command line option to the rsa utility. Thanks to
15965 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15966 algorithm to openssl-dev.
15970 * The evp_local.h macros were using 'c.##kname' which resulted in
15971 invalid expansion on some systems (SCO 5.0.5 for example).
15972 Corrected to 'c.kname'.
15974 *Phillip Porch <root@theporch.com>*
15976 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15977 a STACK of email addresses from a certificate or request, these look
15978 in the subject name and the subject alternative name extensions and
15979 omit any duplicate addresses.
15983 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15984 This makes DSA verification about 2 % faster.
15988 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15989 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15990 plus overhead for 1024 bit moduli).
15991 This makes exponentiations about 0.5 % faster for 1024 bit
15992 exponents (as measured by "openssl speed rsa2048").
15996 * Rename memory handling macros to avoid conflicts with other
15998 Malloc => OPENSSL_malloc
15999 Malloc_locked => OPENSSL_malloc_locked
16000 Realloc => OPENSSL_realloc
16001 Free => OPENSSL_free
16005 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16006 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16010 * CygWin32 support.
16012 *John Jarvie <jjarvie@newsguy.com>*
16014 * The type-safe stack code has been rejigged. It is now only compiled
16015 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16016 by default all type-specific stack functions are "#define"d back to
16017 standard stack functions. This results in more streamlined output
16018 but retains the type-safety checking possibilities of the original
16023 * The STACK code has been cleaned up, and certain type declarations
16024 that didn't make a lot of sense have been brought in line. This has
16025 also involved a cleanup of sorts in safestack.h to more correctly
16026 map type-safe stack functions onto their plain stack counterparts.
16027 This work has also resulted in a variety of "const"ifications of
16028 lots of the code, especially `_cmp` operations which should normally
16029 be prototyped with "const" parameters anyway.
16033 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16034 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16035 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16036 where all of 'md' is used each time the PRNG is used, but 'state'
16037 is used only indexed by a cyclic counter. As entropy may not be
16038 well distributed from the beginning, 'md' is important as a
16039 chaining variable. However, the output function chains only half
16040 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16041 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16042 in all of 'state' being rewritten, with the new values depending
16043 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16047 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16048 the handshake is continued after ssl_verify_cert_chain();
16049 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16050 can lead to 'unexplainable' connection aborts later.
16052 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16054 * Major EVP API cipher revision.
16055 Add hooks for extra EVP features. This allows various cipher
16056 parameters to be set in the EVP interface. Support added for variable
16057 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16058 setting of RC2 and RC5 parameters.
16060 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16063 Remove lots of duplicated code from the EVP library. For example *every*
16064 cipher init() function handles the 'iv' in the same way according to the
16065 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16066 for CFB and OFB modes they zero ctx->num.
16068 New functionality allows removal of S/MIME code RC2 hack.
16070 Most of the routines have the same form and so can be declared in terms
16073 By shifting this to the top level EVP_CipherInit() it can be removed from
16074 all individual ciphers. If the cipher wants to handle IVs or keys
16075 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16078 Change lots of functions like EVP_EncryptUpdate() to now return a
16079 value: although software versions of the algorithms cannot fail
16080 any installed hardware versions can.
16084 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16085 this option is set, tolerate broken clients that send the negotiated
16086 protocol version number instead of the requested protocol version
16091 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16092 i.e. non-zero for export ciphersuites, zero otherwise.
16093 Previous versions had this flag inverted, inconsistent with
16094 rsa_tmp_cb (..._TMP_RSA_CB).
16096 *Bodo Moeller; problem reported by Amit Chopra*
16098 * Add missing DSA library text string. Work around for some IIS
16099 key files with invalid SEQUENCE encoding.
16103 * Add a document (doc/standards.txt) that list all kinds of standards
16104 and so on that are implemented in OpenSSL.
16108 * Enhance c_rehash script. Old version would mishandle certificates
16109 with the same subject name hash and wouldn't handle CRLs at all.
16110 Added -fingerprint option to crl utility, to support new c_rehash
16115 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16119 * Fix for SSL server purpose checking. Server checking was
16120 rejecting certificates which had extended key usage present
16121 but no ssl client purpose.
16123 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16125 * Make PKCS#12 code work with no password. The PKCS#12 spec
16126 is a little unclear about how a blank password is handled.
16127 Since the password in encoded as a BMPString with terminating
16128 double NULL a zero length password would end up as just the
16129 double NULL. However no password at all is different and is
16130 handled differently in the PKCS#12 key generation code. NS
16131 treats a blank password as zero length. MSIE treats it as no
16132 password on export: but it will try both on import. We now do
16133 the same: PKCS12_parse() tries zero length and no password if
16134 the password is set to "" or NULL (NULL is now a valid password:
16135 it wasn't before) as does the pkcs12 application.
16139 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16140 perror when PEM_read_bio_X509_REQ fails, the error message must
16141 be obtained from the error queue.
16145 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16146 it in ERR_remove_state if appropriate, and change ERR_get_state
16147 accordingly to avoid race conditions (this is necessary because
16148 thread_hash is no longer constant once set).
16152 * Bugfix for linux-elf makefile.one.
16156 * RSA_get_default_method() will now cause a default
16157 RSA_METHOD to be chosen if one doesn't exist already.
16158 Previously this was only set during a call to RSA_new()
16159 or RSA_new_method(NULL) meaning it was possible for
16160 RSA_get_default_method() to return NULL.
16164 * Added native name translation to the existing DSO code
16165 that will convert (if the flag to do so is set) filenames
16166 that are sufficiently small and have no path information
16167 into a canonical native form. Eg. "blah" converted to
16168 "libblah.so" or "blah.dll" etc.
16172 * New function ERR_error_string_n(e, buf, len) which is like
16173 ERR_error_string(e, buf), but writes at most 'len' bytes
16174 including the 0 terminator. For ERR_error_string_n, 'buf'
16177 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16179 * CONF library reworked to become more general. A new CONF
16180 configuration file reader "class" is implemented as well as a
16181 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16182 old `CONF_*` functions are still there, but are reimplemented to
16183 work in terms of the new functions. Also, a set of functions
16184 to handle the internal storage of the configuration data is
16185 provided to make it easier to write new configuration file
16186 reader "classes" (I can definitely see something reading a
16187 configuration file in XML format, for example), called `_CONF_*`,
16188 or "the configuration storage API"...
16190 The new configuration file reading functions are:
16192 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16193 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16195 NCONF_default, NCONF_WIN32
16197 NCONF_dump_fp, NCONF_dump_bio
16199 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16200 NCONF_new creates a new CONF object. This works in the same way
16201 as other interfaces in OpenSSL, like the BIO interface.
16202 `NCONF_dump_*` dump the internal storage of the configuration file,
16203 which is useful for debugging. All other functions take the same
16204 arguments as the old `CONF_*` functions with the exception of the
16205 first that must be a `CONF *` instead of a `LHASH *`.
16207 To make it easier to use the new classes with the old `CONF_*` functions,
16208 the function CONF_set_default_method is provided.
16212 * Add '-tls1' option to 'openssl ciphers', which was already
16213 mentioned in the documentation but had not been implemented.
16214 (This option is not yet really useful because even the additional
16215 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16219 * Initial DSO code added into libcrypto for letting OpenSSL (and
16220 OpenSSL-based applications) load shared libraries and bind to
16221 them in a portable way.
16223 *Geoff Thorpe, with contributions from Richard Levitte*
16225 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16227 * Make sure _lrotl and _lrotr are only used with MSVC.
16229 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16230 (the default implementation of RAND_status).
16232 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16233 to '-clrext' (= clear extensions), as intended and documented.
16234 *Bodo Moeller; inconsistency pointed out by Michael Attili
16235 <attili@amaxo.com>*
16237 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16238 was larger than the MD block size.
16240 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16242 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16243 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16244 using the passed key: if the passed key was a private key the result
16245 of X509_print(), for example, would be to print out all the private key
16250 * des_quad_cksum() byte order bug fix.
16251 *Ulf Möller, using the problem description in krb4-0.9.7, where
16252 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16254 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16257 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16259 * For easily testing in shell scripts whether some command
16260 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16261 returns with exit code 0 iff no command of the given name is available.
16262 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16263 the output goes to stdout and nothing is printed to stderr.
16264 Additional arguments are always ignored.
16266 Since for each cipher there is a command of the same name,
16267 the 'no-cipher' compilation switches can be tested this way.
16269 ('openssl no-XXX' is not able to detect pseudo-commands such
16270 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16274 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16278 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16279 is set; it will be thrown away anyway because each handshake creates
16281 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16282 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16283 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16284 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16288 * New s_client option -ign_eof: EOF at stdin is ignored, and
16289 'Q' and 'R' lose their special meanings (quit/renegotiate).
16290 This is part of what -quiet does; unlike -quiet, -ign_eof
16291 does not suppress any output.
16295 * Add compatibility options to the purpose and trust code. The
16296 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16297 accepts a certificate or CA, this was the previous behaviour,
16298 with all the associated security issues.
16300 X509_TRUST_COMPAT is the old trust behaviour: only and
16301 automatically trust self signed roots in certificate store. A
16302 new trust setting X509_TRUST_DEFAULT is used to specify that
16303 a purpose has no associated trust setting and it should instead
16304 use the value in the default purpose.
16308 * Fix the PKCS#8 DSA private key code so it decodes keys again
16309 and fix a memory leak.
16313 * In util/mkerr.pl (which implements 'make errors'), preserve
16314 reason strings from the previous version of the .c file, as
16315 the default to have only downcase letters (and digits) in
16316 automatically generated reasons codes is not always appropriate.
16320 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16321 using strerror. Previously, ERR_reason_error_string() returned
16322 library names as reason strings for SYSerr; but SYSerr is a special
16323 case where small numbers are errno values, not library numbers.
16327 * Add '-dsaparam' option to 'openssl dhparam' application. This
16328 converts DSA parameters into DH parameters. (When creating parameters,
16329 DSA_generate_parameters is used.)
16333 * Include 'length' (recommended exponent length) in C code generated
16334 by 'openssl dhparam -C'.
16338 * The second argument to set_label in perlasm was already being used
16339 so couldn't be used as a "file scope" flag. Moved to third argument
16344 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16345 instead of RAND_bytes for encryption IVs and salts.
16349 * Include RAND_status() into RAND_METHOD instead of implementing
16350 it only for md_rand.c Otherwise replacing the PRNG by calling
16351 RAND_set_rand_method would be impossible.
16355 * Don't let DSA_generate_key() enter an infinite loop if the random
16356 number generation fails.
16360 * New 'rand' application for creating pseudo-random output.
16364 * Added configuration support for Linux/IA64
16366 *Rolf Haberrecker <rolf@suse.de>*
16368 * Assembler module support for Mingw32.
16372 * Shared library support for HPUX (in shlib/).
16374 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16376 * Shared library support for Solaris gcc.
16378 *Lutz Behnke <behnke@trustcenter.de>*
16380 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16382 * PKCS7_encrypt() was adding text MIME headers twice because they
16383 were added manually and by SMIME_crlf_copy().
16387 * In bntest.c don't call BN_rand with zero bits argument.
16389 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16391 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16392 case was implemented. This caused BN_div_recp() to fail occasionally.
16396 * Add an optional second argument to the set_label() in the perl
16397 assembly language builder. If this argument exists and is set
16398 to 1 it signals that the assembler should use a symbol whose
16399 scope is the entire file, not just the current function. This
16400 is needed with MASM which uses the format label:: for this scope.
16402 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16404 * Change the ASN1 types so they are typedefs by default. Before
16405 almost all types were #define'd to ASN1_STRING which was causing
16406 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16411 * Change names of new functions to the new get1/get0 naming
16412 convention: After 'get1', the caller owns a reference count
16413 and has to call `..._free`; 'get0' returns a pointer to some
16414 data structure without incrementing reference counters.
16415 (Some of the existing 'get' functions increment a reference
16416 counter, some don't.)
16417 Similarly, 'set1' and 'add1' functions increase reference
16418 counters or duplicate objects.
16422 * Allow for the possibility of temp RSA key generation failure:
16423 the code used to assume it always worked and crashed on failure.
16427 * Fix potential buffer overrun problem in BIO_printf().
16428 *Ulf Möller, using public domain code by Patrick Powell; problem
16429 pointed out by David Sacerdote <das33@cornell.edu>*
16431 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16432 RAND_egd() and RAND_status(). In the command line application,
16433 the EGD socket can be specified like a seed file using RANDFILE
16438 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16439 Some CAs (e.g. Verisign) distribute certificates in this form.
16443 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16444 list to exclude them. This means that no special compilation option
16445 is needed to use anonymous DH: it just needs to be included in the
16450 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16451 EVP_MD_type. The old functionality is available in a new macro called
16452 EVP_MD_md(). Change code that uses it and update docs.
16456 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16457 where the `void *` argument is replaced by a function pointer argument.
16458 Previously `void *` was abused to point to functions, which works on
16459 many platforms, but is not correct. As these functions are usually
16460 called by macros defined in OpenSSL header files, most source code
16461 should work without changes.
16465 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16466 sections with information on -D... compiler switches used for
16467 compiling the library so that applications can see them. To enable
16468 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16469 must be defined. E.g.,
16470 #define OPENSSL_ALGORITHM_DEFINES
16471 #include <openssl/opensslconf.h>
16472 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16474 *Richard Levitte, Ulf and Bodo Möller*
16476 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16481 * Change the 'other' type in certificate aux info to a STACK_OF
16482 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16483 the required ASN1 format: arbitrary types determined by an OID.
16487 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16488 argument to 'req'. This is not because the function is newer or
16489 better than others it just uses the work 'NEW' in the certificate
16490 request header lines. Some software needs this.
16494 * Reorganise password command line arguments: now passwords can be
16495 obtained from various sources. Delete the PEM_cb function and make
16496 it the default behaviour: i.e. if the callback is NULL and the
16497 usrdata argument is not NULL interpret it as a null terminated pass
16498 phrase. If usrdata and the callback are NULL then the pass phrase
16499 is prompted for as usual.
16503 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16504 the support is automatically enabled. The resulting binaries will
16505 autodetect the card and use it if present.
16507 *Ben Laurie and Compaq Inc.*
16509 * Work around for Netscape hang bug. This sends certificate request
16510 and server done in one record. Since this is perfectly legal in the
16511 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16512 the bugs/SSLv3 entry for more info.
16516 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16520 * Add -rand argument to smime and pkcs12 applications and read/write
16525 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16529 * Add command line password options to the remaining applications.
16533 * Bug fix for BN_div_recp() for numerators with an even number of
16538 * More tests in bntest.c, and changed test_bn output.
16542 * ./config recognizes MacOS X now.
16546 * Bug fix for BN_div() when the first words of num and divisor are
16547 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16551 * Add support for various broken PKCS#8 formats, and command line
16552 options to produce them.
16556 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16557 get temporary BIGNUMs from a BN_CTX.
16561 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16566 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16567 include a #define from the old name to the new. The original intent
16568 was that statically linked binaries could for example just call
16569 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16570 link with digests. This never worked because SSLeay_add_all_digests()
16571 and SSLeay_add_all_ciphers() were in the same source file so calling
16572 one would link with the other. They are now in separate source files.
16576 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16580 * Use a less unusual form of the Miller-Rabin primality test (it used
16581 a binary algorithm for exponentiation integrated into the Miller-Rabin
16582 loop, our standard modexp algorithms are faster).
16586 * Support for the EBCDIC character set completed.
16588 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16590 * Source code cleanups: use const where appropriate, eliminate casts,
16591 use `void *` instead of `char *` in lhash.
16595 * Bugfix: ssl3_send_server_key_exchange was not restartable
16596 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16597 this the server could overwrite ephemeral keys that the client
16602 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16603 using 50 iterations of the Rabin-Miller test.
16605 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16606 iterations of the Rabin-Miller test as required by the appendix
16607 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16608 As BN_is_prime_fasttest includes trial division, DSA parameter
16609 generation becomes much faster.
16611 This implies a change for the callback functions in DSA_is_prime
16612 and DSA_generate_parameters: The callback function is called once
16613 for each positive witness in the Rabin-Miller test, not just
16614 occasionally in the inner loop; and the parameters to the
16615 callback function now provide an iteration count for the outer
16616 loop rather than for the current invocation of the inner loop.
16617 DSA_generate_parameters additionally can call the callback
16618 function with an 'iteration count' of -1, meaning that a
16619 candidate has passed the trial division test (when q is generated
16620 from an application-provided seed, trial division is skipped).
16624 * New function BN_is_prime_fasttest that optionally does trial
16625 division before starting the Rabin-Miller test and has
16626 an additional BN_CTX * argument (whereas BN_is_prime always
16627 has to allocate at least one BN_CTX).
16628 'callback(1, -1, cb_arg)' is called when a number has passed the
16629 trial division stage.
16633 * Fix for bug in CRL encoding. The validity dates weren't being handled
16638 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16642 * New function BN_pseudo_rand().
16646 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16647 bignum version of BN_from_montgomery() with the working code from
16648 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16653 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16654 made it impossible to use the same SSL_SESSION data structure in
16655 SSL2 clients in multiple threads.
16659 * The return value of RAND_load_file() no longer counts bytes obtained
16660 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16661 to seed the PRNG (previously an explicit byte count was required).
16663 *Ulf Möller, Bodo Möller*
16665 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16666 used `char *` instead of `void *` and had casts all over the place.
16670 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16674 * Retain source code compatibility for BN_prime_checks macro:
16675 BN_is_prime(..., BN_prime_checks, ...) now uses
16676 BN_prime_checks_for_size to determine the appropriate number of
16677 Rabin-Miller iterations.
16681 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16682 DH_CHECK_P_NOT_SAFE_PRIME.
16683 (Check if this is true? OpenPGP calls them "strong".)
16687 * Merge the functionality of "dh" and "gendh" programs into a new program
16688 "dhparam". The old programs are retained for now but will handle DH keys
16689 (instead of parameters) in future.
16693 * Make the ciphers, s_server and s_client programs check the return values
16694 when a new cipher list is set.
16698 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16699 ciphers. Before when the 56bit ciphers were enabled the sorting was
16702 The syntax for the cipher sorting has been extended to support sorting by
16703 cipher-strength (using the strength_bits hard coded in the tables).
16704 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16706 Fix a bug in the cipher-command parser: when supplying a cipher command
16707 string with an "undefined" symbol (neither command nor alphanumeric
16708 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16709 an error is flagged.
16711 Due to the strength-sorting extension, the code of the
16712 ssl_create_cipher_list() function was completely rearranged. I hope that
16713 the readability was also increased :-)
16715 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16717 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16718 for the first serial number and places 2 in the serial number file. This
16719 avoids problems when the root CA is created with serial number zero and
16720 the first user certificate has the same issuer name and serial number
16725 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16726 the new code. Add documentation for this stuff.
16730 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16731 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16732 structures and behave in an analogous way to the X509v3 functions:
16733 they shouldn't be called directly but wrapper functions should be used
16736 So we also now have some wrapper functions that call the X509at functions
16737 when passed certificate requests. (TO DO: similar things can be done with
16738 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16739 things. Some of these need some d2i or i2d and print functionality
16740 because they handle more complex structures.)
16744 * Add missing #ifndefs that caused missing symbols when building libssl
16745 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16746 NO_RSA in `ssl/s2*.c`.
16748 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16750 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16751 has a return value which indicates the quality of the random data
16752 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16753 error queue. New function RAND_pseudo_bytes() generates output that is
16754 guaranteed to be unique but not unpredictable. RAND_add is like
16755 RAND_seed, but takes an extra argument for an entropy estimate
16756 (RAND_seed always assumes full entropy).
16760 * Do more iterations of Rabin-Miller probable prime test (specifically,
16761 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16762 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16763 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16764 false-positive rate of at most 2^-80 for random input.
16768 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16772 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16773 in the 0.9.5 release), this returns the chain
16774 from an X509_CTX structure with a dup of the stack and all
16775 the X509 reference counts upped: so the stack will exist
16776 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16779 Also make SSL_SESSION_print() print out the verify return
16784 * Add manpage for the pkcs12 command. Also change the default
16785 behaviour so MAC iteration counts are used unless the new
16786 -nomaciter option is used. This improves file security and
16787 only older versions of MSIE (4.0 for example) need it.
16791 * Honor the no-xxx Configure options when creating .DEF files.
16795 * Add PKCS#10 attributes to field table: challengePassword,
16796 unstructuredName and unstructuredAddress. These are taken from
16797 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16798 international characters are used.
16800 More changes to X509_ATTRIBUTE code: allow the setting of types
16801 based on strings. Remove the 'loc' parameter when adding
16802 attributes because these will be a SET OF encoding which is sorted
16807 * Initial changes to the 'req' utility to allow request generation
16808 automation. This will allow an application to just generate a template
16809 file containing all the field values and have req construct the
16812 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16813 used all over the place including certificate requests and PKCS#7
16814 structures. They are currently handled manually where necessary with
16815 some primitive wrappers for PKCS#7. The new functions behave in a
16816 manner analogous to the X509 extension functions: they allow
16817 attributes to be looked up by NID and added.
16819 Later something similar to the X509V3 code would be desirable to
16820 automatically handle the encoding, decoding and printing of the
16821 more complex types. The string types like challengePassword can
16822 be handled by the string table functions.
16824 Also modified the multi byte string table handling. Now there is
16825 a 'global mask' which masks out certain types. The table itself
16826 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16827 is useful when for example there is only one permissible type
16828 (as in countryName) and using the mask might result in no valid
16833 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16834 SSL_get_peer_finished to allow applications to obtain the latest
16835 Finished messages sent to the peer or expected from the peer,
16836 respectively. (SSL_get_peer_finished is usually the Finished message
16837 actually received from the peer, otherwise the protocol will be aborted.)
16839 As the Finished message are message digests of the complete handshake
16840 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16841 be used for external authentication procedures when the authentication
16842 provided by SSL/TLS is not desired or is not enough.
16846 * Enhanced support for Alpha Linux is added. Now ./config checks if
16847 the host supports BWX extension and if Compaq C is present on the
16848 $PATH. Just exploiting of the BWX extension results in 20-30%
16849 performance kick for some algorithms, e.g. DES and RC4 to mention
16850 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16855 * Add support for MS "fast SGC". This is arguably a violation of the
16856 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16857 weak crypto and after checking the certificate is SGC a second one
16858 with strong crypto. MS SGC stops the first handshake after receiving
16859 the server certificate message and sends a second client hello. Since
16860 a server will typically do all the time consuming operations before
16861 expecting any further messages from the client (server key exchange
16862 is the most expensive) there is little difference between the two.
16864 To get OpenSSL to support MS SGC we have to permit a second client
16865 hello message after we have sent server done. In addition we have to
16866 reset the MAC if we do get this second client hello.
16870 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16871 if a DER encoded private key is RSA or DSA traditional format. Changed
16872 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16873 format DER encoded private key. Newer code should use PKCS#8 format which
16874 has the key type encoded in the ASN1 structure. Added DER private key
16875 support to pkcs8 application.
16879 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16880 ciphersuites has been selected (as required by the SSL 3/TLS 1
16881 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16882 is set, we interpret this as a request to violate the specification
16883 (the worst that can happen is a handshake failure, and 'correct'
16884 behaviour would result in a handshake failure anyway).
16888 * In SSL_CTX_add_session, take into account that there might be multiple
16889 SSL_SESSION structures with the same session ID (e.g. when two threads
16890 concurrently obtain them from an external cache).
16891 The internal cache can handle only one SSL_SESSION with a given ID,
16892 so if there's a conflict, we now throw out the old one to achieve
16897 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16898 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16899 some routines that use cipher OIDs: some ciphers do not have OIDs
16900 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16905 * Simplify the trust setting structure and code. Now we just have
16906 two sequences of OIDs for trusted and rejected settings. These will
16907 typically have values the same as the extended key usage extension
16908 and any application specific purposes.
16910 The trust checking code now has a default behaviour: it will just
16911 check for an object with the same NID as the passed id. Functions can
16912 be provided to override either the default behaviour or the behaviour
16913 for a given id. SSL client, server and email already have functions
16914 in place for compatibility: they check the NID and also return "trusted"
16915 if the certificate is self signed.
16919 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16920 traditional format into an EVP_PKEY structure.
16924 * Add a password callback function PEM_cb() which either prompts for
16925 a password if usr_data is NULL or otherwise assumes it is a null
16926 terminated password. Allow passwords to be passed on command line
16927 environment or config files in a few more utilities.
16931 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16932 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16933 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16934 Update documentation.
16938 * Support for ASN1 "NULL" type. This could be handled before by using
16939 ASN1_TYPE but there wasn't any function that would try to read a NULL
16940 and produce an error if it couldn't. For compatibility we also have
16941 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16942 don't allocate anything because they don't need to.
16946 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16949 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16951 * Rebuild of the memory allocation routines used by OpenSSL code and
16952 possibly others as well. The purpose is to make an interface that
16953 provide hooks so anyone can build a separate set of allocation and
16954 deallocation routines to be used by OpenSSL, for example memory
16955 pool implementations, or something else, which was previously hard
16956 since Malloc(), Realloc() and Free() were defined as macros having
16957 the values malloc, realloc and free, respectively (except for Win32
16958 compilations). The same is provided for memory debugging code.
16959 OpenSSL already comes with functionality to find memory leaks, but
16960 this gives people a chance to debug other memory problems.
16962 With these changes, a new set of functions and macros have appeared:
16964 CRYPTO_set_mem_debug_functions() [F]
16965 CRYPTO_get_mem_debug_functions() [F]
16966 CRYPTO_dbg_set_options() [F]
16967 CRYPTO_dbg_get_options() [F]
16968 CRYPTO_malloc_debug_init() [M]
16970 The memory debug functions are NULL by default, unless the library
16971 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16972 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16973 gives the standard debugging functions that come with OpenSSL) or
16974 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16975 provided by the library user) must be used. When the standard
16976 debugging functions are used, CRYPTO_dbg_set_options can be used to
16977 request additional information:
16978 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16979 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16981 Also, things like CRYPTO_set_mem_functions will always give the
16982 expected result (the new set of functions is used for allocation
16983 and deallocation) at all times, regardless of platform and compiler
16986 To finish it up, some functions that were never use in any other
16987 way than through macros have a new API and new semantic:
16989 CRYPTO_dbg_malloc()
16990 CRYPTO_dbg_realloc()
16993 All macros of value have retained their old syntax.
16995 *Richard Levitte and Bodo Moeller*
16997 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16998 ordering of SMIMECapabilities wasn't in "strength order" and there
16999 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17004 * Some ASN1 types with illegal zero length encoding (INTEGER,
17005 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17007 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17009 * Merge in my S/MIME library for OpenSSL. This provides a simple
17010 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17011 functionality to handle multipart/signed properly) and a utility
17012 called 'smime' to call all this stuff. This is based on code I
17013 originally wrote for Celo who have kindly allowed it to be
17014 included in OpenSSL.
17018 * Add variants des_set_key_checked and des_set_key_unchecked of
17019 des_set_key (aka des_key_sched). Global variable des_check_key
17020 decides which of these is called by des_set_key; this way
17021 des_check_key behaves as it always did, but applications and
17022 the library itself, which was buggy for des_check_key == 1,
17023 have a cleaner way to pick the version they need.
17027 * New function PKCS12_newpass() which changes the password of a
17032 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17033 dynamic mix. In both cases the ids can be used as an index into the
17034 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17035 functions so they accept a list of the field values and the
17036 application doesn't need to directly manipulate the X509_TRUST
17041 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17046 * Modify the way the V3 extension code looks up extensions. This now
17047 works in a similar way to the object code: we have some "standard"
17048 extensions in a static table which is searched with OBJ_bsearch()
17049 and the application can add dynamic ones if needed. The file
17050 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17051 updated whenever a new extension is added to the core code and kept
17052 in ext_nid order. There is a simple program 'tabtest.c' which checks
17053 this. New extensions are not added too often so this file can readily
17054 be maintained manually.
17056 There are two big advantages in doing things this way. The extensions
17057 can be looked up immediately and no longer need to be "added" using
17058 X509V3_add_standard_extensions(): this function now does nothing.
17059 Side note: I get *lots* of email saying the extension code doesn't
17060 work because people forget to call this function.
17061 Also no dynamic allocation is done unless new extensions are added:
17062 so if we don't add custom extensions there is no need to call
17063 X509V3_EXT_cleanup().
17067 * Modify enc utility's salting as follows: make salting the default. Add a
17068 magic header, so unsalted files fail gracefully instead of just decrypting
17069 to garbage. This is because not salting is a big security hole, so people
17070 should be discouraged from doing it.
17074 * Fixes and enhancements to the 'x509' utility. It allowed a message
17075 digest to be passed on the command line but it only used this
17076 parameter when signing a certificate. Modified so all relevant
17077 operations are affected by the digest parameter including the
17078 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17079 DSA key was used because it didn't fix the digest.
17083 * Initial certificate chain verify code. Currently tests the untrusted
17084 certificates for consistency with the verify purpose (which is set
17085 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17087 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17088 this is because it will reject chains with invalid extensions whereas
17089 every previous version of OpenSSL and SSLeay made no checks at all.
17091 Trust code: checks the root CA for the relevant trust settings. Trust
17092 settings have an initial value consistent with the verify purpose: e.g.
17093 if the verify purpose is for SSL client use it expects the CA to be
17094 trusted for SSL client use. However the default value can be changed to
17095 permit custom trust settings: one example of this would be to only trust
17096 certificates from a specific "secure" set of CAs.
17098 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17099 which should be used for version portability: especially since the
17100 verify structure is likely to change more often now.
17102 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17103 to set them. If not set then assume SSL clients will verify SSL servers
17106 Two new options to the verify program: -untrusted allows a set of
17107 untrusted certificates to be passed in and -purpose which sets the
17108 intended purpose of the certificate. If a purpose is set then the
17109 new chain verify code is used to check extension consistency.
17113 * Support for the authority information access extension.
17117 * Modify RSA and DSA PEM read routines to transparently handle
17118 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17119 public keys in a format compatible with certificate
17120 SubjectPublicKeyInfo structures. Unfortunately there were already
17121 functions called *_PublicKey_* which used various odd formats so
17122 these are retained for compatibility: however the DSA variants were
17123 never in a public release so they have been deleted. Changed dsa/rsa
17124 utilities to handle the new format: note no releases ever handled public
17125 keys so we should be OK.
17127 The primary motivation for this change is to avoid the same fiasco
17128 that dogs private keys: there are several incompatible private key
17129 formats some of which are standard and some OpenSSL specific and
17130 require various evil hacks to allow partial transparent handling and
17131 even then it doesn't work with DER formats. Given the option anything
17132 other than PKCS#8 should be dumped: but the other formats have to
17133 stay in the name of compatibility.
17135 With public keys and the benefit of hindsight one standard format
17136 is used which works with EVP_PKEY, RSA or DSA structures: though
17137 it clearly returns an error if you try to read the wrong kind of key.
17139 Added a -pubkey option to the 'x509' utility to output the public key.
17140 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17141 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17142 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17143 that do the same as the `EVP_PKEY_assign_*()` except they up the
17144 reference count of the added key (they don't "swallow" the
17149 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17150 CRLs would fail if the file contained no certificates or no CRLs:
17151 added a new function to read in both types and return the number
17152 read: this means that if none are read it will be an error. The
17153 DER versions of the certificate and CRL reader would always fail
17154 because it isn't possible to mix certificates and CRLs in DER format
17155 without choking one or the other routine. Changed this to just read
17156 a certificate: this is the best we can do. Also modified the code
17157 in `apps/verify.c` to take notice of return codes: it was previously
17158 attempting to read in certificates from NULL pointers and ignoring
17159 any errors: this is one reason why the cert and CRL reader seemed
17160 to work. It doesn't check return codes from the default certificate
17161 routines: these may well fail if the certificates aren't installed.
17165 * Code to support otherName option in GeneralName.
17169 * First update to verify code. Change the verify utility
17170 so it warns if it is passed a self signed certificate:
17171 for consistency with the normal behaviour. X509_verify
17172 has been modified to it will now verify a self signed
17173 certificate if *exactly* the same certificate appears
17174 in the store: it was previously impossible to trust a
17175 single self signed certificate. This means that:
17176 openssl verify ss.pem
17177 now gives a warning about a self signed certificate but
17178 openssl verify -CAfile ss.pem ss.pem
17183 * For servers, store verify_result in SSL_SESSION data structure
17184 (and add it to external session representation).
17185 This is needed when client certificate verifications fails,
17186 but an application-provided verification callback (set by
17187 SSL_CTX_set_cert_verify_callback) allows accepting the session
17188 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17189 but returns 1): When the session is reused, we have to set
17190 ssl->verify_result to the appropriate error code to avoid
17193 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17195 * Fix a bug in the new PKCS#7 code: it didn't consider the
17196 case in PKCS7_dataInit() where the signed PKCS7 structure
17197 didn't contain any existing data because it was being created.
17199 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17201 * Add a salt to the key derivation routines in enc.c. This
17202 forms the first 8 bytes of the encrypted file. Also add a
17203 -S option to allow a salt to be input on the command line.
17207 * New function X509_cmp(). Oddly enough there wasn't a function
17208 to compare two certificates. We do this by working out the SHA1
17209 hash and comparing that. X509_cmp() will be needed by the trust
17214 * SSL_get1_session() is like SSL_get_session(), but increments
17215 the reference count in the SSL_SESSION returned.
17217 *Geoff Thorpe <geoff@eu.c2.net>*
17219 * Fix for 'req': it was adding a null to request attributes.
17220 Also change the X509_LOOKUP and X509_INFO code to handle
17221 certificate auxiliary information.
17225 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17230 * Add the possibility to add extra information to the memory leak
17231 detecting output, to form tracebacks, showing from where each
17232 allocation was originated: CRYPTO_push_info("constant string") adds
17233 the string plus current file name and line number to a per-thread
17234 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17235 is like calling CYRPTO_pop_info() until the stack is empty.
17236 Also updated memory leak detection code to be multi-thread-safe.
17240 * Add options -text and -noout to pkcs7 utility and delete the
17241 encryption options which never did anything. Update docs.
17245 * Add options to some of the utilities to allow the pass phrase
17246 to be included on either the command line (not recommended on
17247 OSes like Unix) or read from the environment. Update the
17248 manpages and fix a few bugs.
17252 * Add a few manpages for some of the openssl commands.
17256 * Fix the -revoke option in ca. It was freeing up memory twice,
17257 leaking and not finding already revoked certificates.
17261 * Extensive changes to support certificate auxiliary information.
17262 This involves the use of X509_CERT_AUX structure and X509_AUX
17263 functions. An X509_AUX function such as PEM_read_X509_AUX()
17264 can still read in a certificate file in the usual way but it
17265 will also read in any additional "auxiliary information". By
17266 doing things this way a fair degree of compatibility can be
17267 retained: existing certificates can have this information added
17268 using the new 'x509' options.
17270 Current auxiliary information includes an "alias" and some trust
17271 settings. The trust settings will ultimately be used in enhanced
17272 certificate chain verification routines: currently a certificate
17273 can only be trusted if it is self signed and then it is trusted
17278 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17279 The problem was that one of the replacement routines had not been working
17280 since SSLeay releases. For now the offending routine has been replaced
17281 with non-optimised assembler. Even so, this now gives around 95%
17282 performance improvement for 1024 bit RSA signs.
17286 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17287 handling. Most clients have the effective key size in bits equal to
17288 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17289 A few however don't do this and instead use the size of the decrypted key
17290 to determine the RC2 key length and the AlgorithmIdentifier to determine
17291 the effective key length. In this case the effective key length can still
17292 be 40 bits but the key length can be 168 bits for example. This is fixed
17293 by manually forcing an RC2 key into the EVP_PKEY structure because the
17294 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17295 the key length and effective key length are equal.
17299 * Add a bunch of functions that should simplify the creation of
17300 X509_NAME structures. Now you should be able to do:
17301 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17302 and have it automatically work out the correct field type and fill in
17303 the structures. The more adventurous can try:
17304 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17305 and it will (hopefully) work out the correct multibyte encoding.
17309 * Change the 'req' utility to use the new field handling and multibyte
17310 copy routines. Before the DN field creation was handled in an ad hoc
17311 way in req, ca, and x509 which was rather broken and didn't support
17312 BMPStrings or UTF8Strings. Since some software doesn't implement
17313 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17314 using the dirstring_type option. See the new comment in the default
17315 openssl.cnf for more info.
17319 * Make crypto/rand/md_rand.c more robust:
17320 - Assure unique random numbers after fork().
17321 - Make sure that concurrent threads access the global counter and
17322 md serializably so that we never lose entropy in them
17323 or use exactly the same state in multiple threads.
17324 Access to the large state is not always serializable because
17325 the additional locking could be a performance killer, and
17326 md should be large enough anyway.
17330 * New file `apps/app_rand.c` with commonly needed functionality
17331 for handling the random seed file.
17333 Use the random seed file in some applications that previously did not:
17335 dsaparam -genkey (which also ignored its '-rand' option),
17338 x509 (when signing).
17339 Except on systems with /dev/urandom, it is crucial to have a random
17340 seed file at least for key creation, DSA signing, and for DH exchanges;
17341 for RSA signatures we could do without one.
17343 gendh and gendsa (unlike genrsa) used to read only the first byte
17344 of each file listed in the '-rand' option. The function as previously
17345 found in genrsa is now in app_rand.c and is used by all programs
17346 that support '-rand'.
17350 * In RAND_write_file, use mode 0600 for creating files;
17351 don't just chmod when it may be too late.
17355 * Report an error from X509_STORE_load_locations
17356 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17360 * New function ASN1_mbstring_copy() this copies a string in either
17361 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17362 into an ASN1_STRING type. A mask of permissible types is passed
17363 and it chooses the "minimal" type to use or an error if not type
17368 * Add function equivalents to the various macros in asn1.h. The old
17369 macros are retained with an `M_` prefix. Code inside the library can
17370 use the `M_` macros. External code (including the openssl utility)
17371 should *NOT* in order to be "shared library friendly".
17375 * Add various functions that can check a certificate's extensions
17376 to see if it usable for various purposes such as SSL client,
17377 server or S/MIME and CAs of these types. This is currently
17378 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17379 verification. Also added a -purpose flag to x509 utility to
17380 print out all the purposes.
17384 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17389 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17390 for, obtain and decode and extension and obtain its critical flag.
17391 This allows all the necessary extension code to be handled in a
17392 single function call.
17396 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17397 platforms. See crypto/rc4/rc4_enc.c for further details.
17401 * New -noout option to asn1parse. This causes no output to be produced
17402 its main use is when combined with -strparse and -out to extract data
17403 from a file (which may not be in ASN.1 format).
17407 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17408 when producing the local key id.
17410 *Richard Levitte <levitte@stacken.kth.se>*
17412 * New option -dhparam in s_server. This allows a DH parameter file to be
17413 stated explicitly. If it is not stated then it tries the first server
17414 certificate file. The previous behaviour hard coded the filename
17419 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17420 a public key to be input or output. For example:
17421 openssl rsa -in key.pem -pubout -out pubkey.pem
17422 Also added necessary DSA public key functions to handle this.
17426 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17427 in the message. This was handled by allowing
17428 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17430 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17432 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17433 to the end of the strings whereas this didn't. This would cause problems
17434 if strings read with d2i_ASN1_bytes() were later modified.
17436 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17438 * Fix for base64 decode bug. When a base64 bio reads only one line of
17439 data and it contains EOF it will end up returning an error. This is
17440 caused by input 46 bytes long. The cause is due to the way base64
17441 BIOs find the start of base64 encoded data. They do this by trying a
17442 trial decode on each line until they find one that works. When they
17443 do a flag is set and it starts again knowing it can pass all the
17444 data directly through the decoder. Unfortunately it doesn't reset
17445 the context it uses. This means that if EOF is reached an attempt
17446 is made to pass two EOFs through the context and this causes the
17447 resulting error. This can also cause other problems as well. As is
17448 usual with these problems it takes *ages* to find and the fix is
17449 trivial: move one line.
17451 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17453 * Ugly workaround to get s_client and s_server working under Windows. The
17454 old code wouldn't work because it needed to select() on sockets and the
17455 tty (for keypresses and to see if data could be written). Win32 only
17456 supports select() on sockets so we select() with a 1s timeout on the
17457 sockets and then see if any characters are waiting to be read, if none
17458 are present then we retry, we also assume we can always write data to
17459 the tty. This isn't nice because the code then blocks until we've
17460 received a complete line of data and it is effectively polling the
17461 keyboard at 1s intervals: however it's quite a bit better than not
17462 working at all :-) A dedicated Windows application might handle this
17463 with an event loop for example.
17467 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17468 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17469 will be called when RSA_sign() and RSA_verify() are used. This is useful
17470 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17471 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17472 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17473 This necessitated the support of an extra signature type NID_md5_sha1
17474 for SSL signatures and modifications to the SSL library to use it instead
17475 of calling RSA_public_decrypt() and RSA_private_encrypt().
17479 * Add new -verify -CAfile and -CApath options to the crl program, these
17480 will lookup a CRL issuers certificate and verify the signature in a
17481 similar way to the verify program. Tidy up the crl program so it
17482 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17483 less strict. It will now permit CRL extensions even if it is not
17484 a V2 CRL: this will allow it to tolerate some broken CRLs.
17488 * Initialize all non-automatic variables each time one of the openssl
17489 sub-programs is started (this is necessary as they may be started
17490 multiple times from the "OpenSSL>" prompt).
17492 *Lennart Bang, Bodo Moeller*
17494 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17495 removing all other RSA functionality (this is what NO_RSA does). This
17496 is so (for example) those in the US can disable those operations covered
17497 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17502 * Non-copying interface to BIO pairs.
17503 (still largely untested)
17507 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17508 ASCII string. This was handled independently in various places before.
17512 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17513 UTF8 strings a character at a time.
17517 * Use client_version from client hello to select the protocol
17518 (s23_srvr.c) and for RSA client key exchange verification
17519 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17523 * Add various utility functions to handle SPKACs, these were previously
17524 handled by poking round in the structure internals. Added new function
17525 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17526 print, verify and generate SPKACs. Based on an original idea from
17527 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17531 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17535 * Allow the config file extension section to be overwritten on the
17536 command line. Based on an original idea from Massimiliano Pala
17537 <madwolf@comune.modena.it>. The new option is called -extensions
17538 and can be applied to ca, req and x509. Also -reqexts to override
17539 the request extensions in req and -crlexts to override the crl extensions
17544 * Add new feature to the SPKAC handling in ca. Now you can include
17545 the same field multiple times by preceding it by "XXXX." for example:
17548 this is the same syntax as used in the req config file.
17552 * Allow certificate extensions to be added to certificate requests. These
17553 are specified in a 'req_extensions' option of the req section of the
17554 config file. They can be printed out with the -text option to req but
17555 are otherwise ignored at present.
17559 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17560 data read consists of only the final block it would not decrypted because
17561 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17562 A misplaced 'break' also meant the decrypted final block might not be
17563 copied until the next read.
17567 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17568 a few extra parameters to the DH structure: these will be useful if
17569 for example we want the value of 'q' or implement X9.42 DH.
17573 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17574 provides hooks that allow the default DSA functions or functions on a
17575 "per key" basis to be replaced. This allows hardware acceleration and
17576 hardware key storage to be handled without major modification to the
17577 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17578 associated functions.
17582 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17583 as "read only": it can't be written to and the buffer it points to will
17584 not be freed. Reading from a read only BIO is much more efficient than
17585 a normal memory BIO. This was added because there are several times when
17586 an area of memory needs to be read from a BIO. The previous method was
17587 to create a memory BIO and write the data to it, this results in two
17588 copies of the data and an O(n^2) reading algorithm. There is a new
17589 function BIO_new_mem_buf() which creates a read only memory BIO from
17590 an area of memory. Also modified the PKCS#7 routines to use read only
17595 * Bugfix: ssl23_get_client_hello did not work properly when called in
17596 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17597 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17598 but a retry condition occurred while trying to read the rest.
17602 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17603 NID_pkcs7_encrypted by default: this was wrong since this should almost
17604 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17605 the encrypted data type: this is a more sensible place to put it and it
17606 allows the PKCS#12 code to be tidied up that duplicated this
17611 * Changed obj_dat.pl script so it takes its input and output files on
17612 the command line. This should avoid shell escape redirection problems
17617 * Initial support for certificate extension requests, these are included
17618 in things like Xenroll certificate requests. Included functions to allow
17619 extensions to be obtained and added.
17623 * -crlf option to s_client and s_server for sending newlines as
17624 CRLF (as required by many protocols).
17628 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17630 * Install libRSAglue.a when OpenSSL is built with RSAref.
17632 *Ralf S. Engelschall*
17634 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17636 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17638 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17643 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17644 DH parameters/keys (q is lost during that conversion, but the resulting
17645 DH parameters contain its length).
17647 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17648 much faster than DH_generate_parameters (which creates parameters
17649 where `p = 2*q + 1`), and also the smaller q makes DH computations
17650 much more efficient (160-bit exponentiation instead of 1024-bit
17651 exponentiation); so this provides a convenient way to support DHE
17652 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17653 utter importance to use
17654 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17656 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17657 when such DH parameters are used, because otherwise small subgroup
17658 attacks may become possible!
17662 * Avoid memory leak in i2d_DHparams.
17666 * Allow the -k option to be used more than once in the enc program:
17667 this allows the same encrypted message to be read by multiple recipients.
17671 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17672 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17673 it will always use the numerical form of the OID, even if it has a short
17678 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17679 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17680 otherwise bn_mod_exp was called. In the case of hardware keys for example
17681 no private key components need be present and it might store extra data
17682 in the RSA structure, which cannot be accessed from bn_mod_exp.
17683 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17684 private key operations.
17688 * Added support for SPARC Linux.
17692 * pem_password_cb function type incompatibly changed from
17693 typedef int pem_password_cb(char *buf, int size, int rwflag);
17695 ....(char *buf, int size, int rwflag, void *userdata);
17696 so that applications can pass data to their callbacks:
17697 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17698 additional void * argument, which is just handed through whenever
17699 the password callback is called.
17701 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17703 New function SSL_CTX_set_default_passwd_cb_userdata.
17705 Compatibility note: As many C implementations push function arguments
17706 onto the stack in reverse order, the new library version is likely to
17707 interoperate with programs that have been compiled with the old
17708 pem_password_cb definition (PEM_whatever takes some data that
17709 happens to be on the stack as its last argument, and the callback
17710 just ignores this garbage); but there is no guarantee whatsoever that
17713 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17714 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17715 problems not only on Windows, but also on some Unix platforms.
17716 To avoid problematic command lines, these definitions are now in an
17717 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17718 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17722 * MIPS III/IV assembler module is reimplemented.
17726 * More DES library cleanups: remove references to srand/rand and
17727 delete an unused file.
17731 * Add support for the free Netwide assembler (NASM) under Win32,
17732 since not many people have MASM (ml) and it can be hard to obtain.
17733 This is currently experimental but it seems to work OK and pass all
17734 the tests. Check out INSTALL.W32 for info.
17738 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17739 without temporary keys kept an extra copy of the server key,
17740 and connections with temporary keys did not free everything in case
17745 * New function RSA_check_key and new openssl rsa option -check
17746 for verifying the consistency of RSA keys.
17748 *Ulf Moeller, Bodo Moeller*
17750 * Various changes to make Win32 compile work:
17751 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17752 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17753 comparison" warnings.
17754 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17758 * Add a debugging option to PKCS#5 v2 key generation function: when
17759 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17760 derived keys are printed to stderr.
17764 * Copy the flags in ASN1_STRING_dup().
17766 *Roman E. Pavlov <pre@mo.msk.ru>*
17768 * The x509 application mishandled signing requests containing DSA
17769 keys when the signing key was also DSA and the parameters didn't match.
17771 It was supposed to omit the parameters when they matched the signing key:
17772 the verifying software was then supposed to automatically use the CA's
17773 parameters if they were absent from the end user certificate.
17775 Omitting parameters is no longer recommended. The test was also
17776 the wrong way round! This was probably due to unusual behaviour in
17777 EVP_cmp_parameters() which returns 1 if the parameters match.
17778 This meant that parameters were omitted when they *didn't* match and
17779 the certificate was useless. Certificates signed with 'ca' didn't have
17782 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17784 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17785 The interface is as follows:
17786 Applications can use
17787 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17788 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17789 "off" is now the default.
17790 The library internally uses
17791 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17792 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17793 to disable memory-checking temporarily.
17795 Some inconsistent states that previously were possible (and were
17796 even the default) are now avoided.
17798 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17799 with each memory chunk allocated; this is occasionally more helpful
17800 than just having a counter.
17802 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17804 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17809 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17810 which largely parallels "options", but is for changing API behaviour,
17811 whereas "options" are about protocol behaviour.
17812 Initial "mode" flags are:
17814 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17815 a single record has been written.
17816 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17817 retries use the same buffer location.
17818 (But all of the contents must be
17823 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17826 * Fix problems with no-hmac etc.
17828 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17830 * New functions RSA_get_default_method(), RSA_set_method() and
17831 RSA_get_method(). These allows replacement of RSA_METHODs without having
17832 to mess around with the internals of an RSA structure.
17836 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17837 Also really enable memory leak checks in openssl.c and in some
17840 *Chad C. Mulligan, Bodo Moeller*
17842 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17843 up the length of negative integers. This has now been simplified to just
17844 store the length when it is first determined and use it later, rather
17845 than trying to keep track of where data is copied and updating it to
17847 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17849 * Add a new function PKCS7_signatureVerify. This allows the verification
17850 of a PKCS#7 signature but with the signing certificate passed to the
17851 function itself. This contrasts with PKCS7_dataVerify which assumes the
17852 certificate is present in the PKCS#7 structure. This isn't always the
17853 case: certificates can be omitted from a PKCS#7 structure and be
17854 distributed by "out of band" means (such as a certificate database).
17858 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17859 function prototypes in pem.h, also change util/mkdef.pl to add the
17860 necessary function names.
17864 * mk1mf.pl (used by Windows builds) did not properly read the
17865 options set by Configure in the top level Makefile, and Configure
17866 was not even able to write more than one option correctly.
17867 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17871 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17872 file to be loaded from a BIO or FILE pointer. The BIO version will
17873 for example allow memory BIOs to contain config info.
17877 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17878 Whoever hopes to achieve shared-library compatibility across versions
17879 must use this, not the compile-time macro.
17880 (Exercise 0.9.4: Which is the minimum library version required by
17882 Note: All this applies only to multi-threaded programs, others don't
17887 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17888 through a BIO pair triggered the default case, i.e.
17889 SSLerr(...,SSL_R_UNKNOWN_STATE).
17893 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17894 can use the SSL library even if none of the specific BIOs is
17899 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17900 for the encoded length.
17902 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17904 * Add initial documentation of the X509V3 functions.
17908 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17909 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17910 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17911 secure PKCS#8 private key format with a high iteration count.
17915 * Fix determination of Perl interpreter: A perl or perl5
17916 *directory* in $PATH was also accepted as the interpreter.
17918 *Ralf S. Engelschall*
17920 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17921 wrong with it but it was very old and did things like calling
17922 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17923 unusual formatting.
17927 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17928 to use the new extension code.
17932 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17933 with macros. This should make it easier to change their form, add extra
17934 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17939 * Add to configuration table a new entry that can specify an alternative
17940 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17941 according to Mark Crispin <MRC@Panda.COM>.
17945 * DES CBC did not update the IV. Weird.
17949 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17950 Changing the behaviour of the former might break existing programs --
17951 where IV updating is needed, des_ncbc_encrypt can be used.
17954 * When bntest is run from "make test" it drives bc to check its
17955 calculations, as well as internally checking them. If an internal check
17956 fails, it needs to cause bc to give a non-zero result or make test carries
17957 on without noticing the failure. Fixed.
17961 * DES library cleanups.
17965 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17966 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17967 ciphers. NOTE: although the key derivation function has been verified
17968 against some published test vectors it has not been extensively tested
17969 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17974 * Instead of "mkdir -p", which is not fully portable, use new
17975 Perl script "util/mkdir-p.pl".
17979 * Rewrite the way password based encryption (PBE) is handled. It used to
17980 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17981 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17982 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17983 the 'parameter' field of the AlgorithmIdentifier is passed to the
17984 underlying key generation function so it must do its own ASN1 parsing.
17985 This has also changed the EVP_PBE_CipherInit() function which now has a
17986 'parameter' argument instead of literal salt and iteration count values
17987 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17991 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17992 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17993 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17994 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17995 value was just used as a "magic string" and not used directly its
17996 value doesn't matter.
18000 * Introduce some semblance of const correctness to BN. Shame C doesn't
18005 * "linux-sparc64" configuration (ultrapenguin).
18007 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18008 "linux-sparc" configuration.
18010 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18012 * config now generates no-xxx options for missing ciphers.
18016 * Support the EBCDIC character set (work in progress).
18017 File ebcdic.c not yet included because it has a different license.
18019 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18021 * Support BS2000/OSD-POSIX.
18023 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18025 * Make callbacks for key generation use `void *` instead of `char *`.
18029 * Make S/MIME samples compile (not yet tested).
18033 * Additional typesafe stacks.
18037 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18041 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18043 * New configuration variant "sco5-gcc".
18045 * Updated some demos.
18047 *Sean O Riordain, Wade Scholine*
18049 * Add missing BIO_free at exit of pkcs12 application.
18053 * Fix memory leak in conf.c.
18057 * Updates for Win32 to assembler version of MD5.
18061 * Set #! path to perl in `apps/der_chop` to where we found it
18062 instead of using a fixed path.
18066 * SHA library changes for irix64-mips4-cc.
18070 * Improvements for VMS support.
18074 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18076 * Bignum library bug fix. IRIX 6 passes "make test" now!
18077 This also avoids the problems with SC4.2 and unpatched SC5.
18079 *Andy Polyakov <appro@fy.chalmers.se>*
18081 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18082 These are required because of the typesafe stack would otherwise break
18083 existing code. If old code used a structure member which used to be STACK
18084 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18085 sk_num or sk_value it would produce an error because the num, data members
18086 are not present in STACK_OF. Now it just produces a warning. sk_set
18087 replaces the old method of assigning a value to sk_value
18088 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18089 that does this will no longer work (and should use sk_set instead) but
18090 this could be regarded as a "questionable" behaviour anyway.
18094 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18095 correctly handle encrypted S/MIME data.
18099 * Change type of various DES function arguments from des_cblock
18100 (which means, in function argument declarations, pointer to char)
18101 to des_cblock * (meaning pointer to array with 8 char elements),
18102 which allows the compiler to do more typechecking; it was like
18103 that back in SSLeay, but with lots of ugly casts.
18105 Introduce new type const_des_cblock.
18109 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18110 problems: find RecipientInfo structure that matches recipient certificate
18111 and initialise the ASN1 structures properly based on passed cipher.
18115 * Belatedly make the BN tests actually check the results.
18119 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18120 to and from BNs: it was completely broken. New compilation option
18121 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18122 key elements as negative integers.
18126 * Reorganize and speed up MD5.
18128 *Andy Polyakov <appro@fy.chalmers.se>*
18132 *Richard Levitte <richard@levitte.org>*
18134 * New option -out to asn1parse to allow the parsed structure to be
18135 output to a file. This is most useful when combined with the -strparse
18136 option to examine the output of things like OCTET STRINGS.
18140 * Make SSL library a little more fool-proof by not requiring any longer
18141 that `SSL_set_{accept,connect}_state` be called before
18142 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18143 in many applications because usually everything *appeared* to work as
18144 intended anyway -- now it really works as intended).
18148 * Move openssl.cnf out of lib/.
18152 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18153 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18154 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18156 *Ralf S. Engelschall*
18158 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18159 handle PKCS#7 enveloped data properly.
18161 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18163 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18164 copying pointers. The cert_st handling is changed by this in
18165 various ways (and thus what used to be known as ctx->default_cert
18166 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18167 any longer when s->cert does not give us what we need).
18168 ssl_cert_instantiate becomes obsolete by this change.
18169 As soon as we've got the new code right (possibly it already is?),
18170 we have solved a couple of bugs of the earlier code where s->cert
18171 was used as if it could not have been shared with other SSL structures.
18173 Note that using the SSL API in certain dirty ways now will result
18174 in different behaviour than observed with earlier library versions:
18175 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18176 does not influence s as it used to.
18178 In order to clean up things more thoroughly, inside SSL_SESSION
18179 we don't use CERT any longer, but a new structure SESS_CERT
18180 that holds per-session data (if available); currently, this is
18181 the peer's certificate chain and, for clients, the server's certificate
18182 and temporary key. CERT holds only those values that can have
18183 meaningful defaults in an SSL_CTX.
18187 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18188 from the internal representation. Various PKCS#7 fixes: remove some
18189 evil casts and set the enc_dig_alg field properly based on the signing
18194 * Allow PKCS#12 password to be set from the command line or the
18195 environment. Let 'ca' get its config file name from the environment
18196 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18201 * Allow certificate policies extension to use an IA5STRING for the
18202 organization field. This is contrary to the PKIX definition but
18203 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18208 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18209 without disallowing inline assembler and the like for non-pedantic builds.
18213 * Support Borland C++ builder.
18215 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18221 * SHA-1 cleanups and performance enhancements.
18223 *Andy Polyakov <appro@fy.chalmers.se>*
18225 * Sparc v8plus assembler for the bignum library.
18227 *Andy Polyakov <appro@fy.chalmers.se>*
18229 * Accept any -xxx and +xxx compiler options in Configure.
18233 * Update HPUX configuration.
18237 * Add missing `sk_<type>_unshift()` function to safestack.h
18239 *Ralf S. Engelschall*
18241 * New function SSL_CTX_use_certificate_chain_file that sets the
18242 "extra_cert"s in addition to the certificate. (This makes sense
18243 only for "PEM" format files, as chains as a whole are not
18248 * Support verify_depth from the SSL API.
18249 x509_vfy.c had what can be considered an off-by-one-error:
18250 Its depth (which was not part of the external interface)
18251 was actually counting the number of certificates in a chain;
18252 now it really counts the depth.
18256 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18257 instead of X509err, which often resulted in confusing error
18258 messages since the error codes are not globally unique
18259 (e.g. an alleged error in ssl3_accept when a certificate
18260 didn't match the private key).
18262 * New function SSL_CTX_set_session_id_context that allows to set a default
18263 value (so that you don't need SSL_set_session_id_context for each
18264 connection using the SSL_CTX).
18268 * OAEP decoding bug fix.
18272 * Support INSTALL_PREFIX for package builders, as proposed by
18277 * New Configure options "threads" and "no-threads". For systems
18278 where the proper compiler options are known (currently Solaris
18279 and Linux), "threads" is the default.
18283 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18287 * Install various scripts to $(OPENSSLDIR)/misc, not to
18288 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18289 such as /usr/local/bin.
18293 * "make linux-shared" to build shared libraries.
18295 *Niels Poppe <niels@netbox.org>*
18297 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18301 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18302 extension adding in x509 utility.
18306 * Remove NOPROTO sections and error code comments.
18310 * Partial rewrite of the DEF file generator to now parse the ANSI
18315 * New Configure options --prefix=DIR and --openssldir=DIR.
18319 * Complete rewrite of the error code script(s). It is all now handled
18320 by one script at the top level which handles error code gathering,
18321 header rewriting and C source file generation. It should be much better
18322 than the old method: it now uses a modified version of Ulf's parser to
18323 read the ANSI prototypes in all header files (thus the old K&R definitions
18324 aren't needed for error creation any more) and do a better job of
18325 translating function codes into names. The old 'ASN1 error code embedded
18326 in a comment' is no longer necessary and it doesn't use .err files which
18327 have now been deleted. Also the error code call doesn't have to appear all
18328 on one line (which resulted in some large lines...).
18332 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18336 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18337 0 (which usually indicates a closed connection), but continue reading.
18341 * Fix some race conditions.
18345 * Add support for CRL distribution points extension. Add Certificate
18346 Policies and CRL distribution points documentation.
18350 * Move the autogenerated header file parts to crypto/opensslconf.h.
18354 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18355 8 of keying material. Merlin has also confirmed interop with this fix
18356 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18358 *Merlin Hughes <merlin@baltimore.ie>*
18360 * Fix lots of warnings.
18362 *Richard Levitte <levitte@stacken.kth.se>*
18364 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18365 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18367 *Richard Levitte <levitte@stacken.kth.se>*
18369 * Fix problems with sizeof(long) == 8.
18371 *Andy Polyakov <appro@fy.chalmers.se>*
18373 * Change functions to ANSI C.
18377 * Fix typos in error codes.
18379 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18381 * Remove defunct assembler files from Configure.
18385 * SPARC v8 assembler BIGNUM implementation.
18387 *Andy Polyakov <appro@fy.chalmers.se>*
18389 * Support for Certificate Policies extension: both print and set.
18390 Various additions to support the r2i method this uses.
18394 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18395 return a const string when you are expecting an allocated buffer.
18399 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18400 types DirectoryString and DisplayText.
18404 * Add code to allow r2i extensions to access the configuration database,
18405 add an LHASH database driver and add several ctx helper functions.
18409 * Fix an evil bug in bn_expand2() which caused various BN functions to
18410 fail when they extended the size of a BIGNUM.
18414 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18415 support typesafe stack.
18419 * Fix typo in SSL_[gs]et_options().
18421 *Nils Frostberg <nils@medcom.se>*
18423 * Delete various functions and files that belonged to the (now obsolete)
18424 old X509V3 handling code.
18428 * New Configure option "rsaref".
18432 * Don't auto-generate pem.h.
18436 * Introduce type-safe ASN.1 SETs.
18440 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18442 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18444 * Introduce type-safe STACKs. This will almost certainly break lots of code
18445 that links with OpenSSL (well at least cause lots of warnings), but fear
18446 not: the conversion is trivial, and it eliminates loads of evil casts. A
18447 few STACKed things have been converted already. Feel free to convert more.
18448 In the fullness of time, I'll do away with the STACK type altogether.
18452 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18453 specified in `<certfile>` by updating the entry in the index.txt file.
18454 This way one no longer has to edit the index.txt file manually for
18455 revoking a certificate. The -revoke option does the gory details now.
18457 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18459 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18460 `-text` option at all and this way the `-noout -text` combination was
18461 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18463 *Ralf S. Engelschall*
18465 * Make sure a corresponding plain text error message exists for the
18466 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18467 verify callback function determined that a certificate was revoked.
18469 *Ralf S. Engelschall*
18471 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18472 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18473 all available ciphers including rc5, which was forgotten until now.
18474 In order to let the testing shell script know which algorithms
18475 are available, a new (up to now undocumented) command
18476 `openssl list-cipher-commands` is used.
18480 * Bugfix: s_client occasionally would sleep in select() when
18481 it should have checked SSL_pending() first.
18485 * New functions DSA_do_sign and DSA_do_verify to provide access to
18486 the raw DSA values prior to ASN.1 encoding.
18490 * Tweaks to Configure
18492 *Niels Poppe <niels@netbox.org>*
18494 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18499 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18503 * New config option to avoid instructions that are illegal on the 80386.
18504 The default code is faster, but requires at least a 486.
18508 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18509 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18510 same as SSL2_VERSION anyway.
18514 * New "-showcerts" option for s_client.
18518 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18519 application. Various cleanups and fixes.
18523 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18524 modify error routines to work internally. Add error codes and PBE init
18525 to library startup routines.
18529 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18530 packing functions to asn1 and evp. Changed function names and error
18531 codes along the way.
18535 * PKCS12 integration: and so it begins... First of several patches to
18536 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18537 objects to objects.h
18541 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18542 and display support for Thawte strong extranet extension.
18546 * Add LinuxPPC support.
18548 *Jeff Dubrule <igor@pobox.org>*
18550 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18551 bn_div_words in alpha.s.
18553 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18555 * Make sure the RSA OAEP test is skipped under -DRSAref because
18556 OAEP isn't supported when OpenSSL is built with RSAref.
18558 *Ulf Moeller <ulf@fitug.de>*
18560 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18561 so they no longer are missing under -DNOPROTO.
18563 *Soren S. Jorvang <soren@t.dk>*
18565 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18567 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18568 doesn't work when the session is reused. Coming soon!
18572 * Fix a security hole, that allows sessions to be reused in the wrong
18573 context thus bypassing client cert protection! All software that uses
18574 client certs and session caches in multiple contexts NEEDS PATCHING to
18575 allow session reuse! A fuller solution is in the works.
18577 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18579 * Some more source tree cleanups (removed obsolete files
18580 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18581 permission on "config" script to be executable) and a fix for the INSTALL
18584 *Ulf Moeller <ulf@fitug.de>*
18586 * Remove some legacy and erroneous uses of malloc, free instead of
18589 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18591 * Make rsa_oaep_test return non-zero on error.
18593 *Ulf Moeller <ulf@fitug.de>*
18595 * Add support for native Solaris shared libraries. Configure
18596 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18597 if someone would make that last step automatic.
18599 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18601 * ctx_size was not built with the right compiler during "make links". Fixed.
18605 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18606 except NULL ciphers". This means the default cipher list will no longer
18607 enable NULL ciphers. They need to be specifically enabled e.g. with
18608 the string "DEFAULT:eNULL".
18612 * Fix to RSA private encryption routines: if p < q then it would
18613 occasionally produce an invalid result. This will only happen with
18614 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18618 * Be less restrictive and allow also `perl util/perlpath.pl
18619 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18620 because this way one can also use an interpreter named `perl5` (which is
18621 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18622 installed as `perl`).
18624 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18626 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18628 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18630 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18631 advapi32.lib to Win32 build and change the pem test comparison
18632 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18633 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18634 and crypto/des/ede_cbcm_enc.c.
18638 * DES quad checksum was broken on big-endian architectures. Fixed.
18642 * Comment out two functions in bio.h that aren't implemented. Fix up the
18643 Win32 test batch file so it (might) work again. The Win32 test batch file
18644 is horrible: I feel ill....
18648 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18649 in e_os.h. Audit of header files to check ANSI and non ANSI
18650 sections: 10 functions were absent from non ANSI section and not exported
18651 from Windows DLLs. Fixed up libeay.num for new functions.
18655 * Make `openssl version` output lines consistent.
18657 *Ralf S. Engelschall*
18659 * Fix Win32 symbol export lists for BIO functions: Added
18660 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18661 to ms/libeay{16,32}.def.
18663 *Ralf S. Engelschall*
18665 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18666 fine under Unix and passes some trivial tests I've now added. But the
18667 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18668 added to make sure no one expects that this stuff really works in the
18669 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18670 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18673 *Ralf S. Engelschall*
18675 * Fix the generation of two part addresses in perl.
18677 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18679 * Add config entry for Linux on MIPS.
18681 *John Tobey <jtobey@channel1.com>*
18683 * Make links whenever Configure is run, unless we are on Windoze.
18687 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18688 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18693 * Add a useful kludge to allow package maintainers to specify compiler and
18694 other platforms details on the command line without having to patch the
18695 Configure script every time: One now can use
18696 `perl Configure <id>:<details>`,
18697 i.e. platform ids are allowed to have details appended
18698 to them (separated by colons). This is treated as there would be a static
18699 pre-configured entry in Configure's %table under key `<id>` with value
18700 `<details>` and `perl Configure <id>` is called. So, when you want to
18701 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18702 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18703 now, which overrides the FreeBSD-elf entry on-the-fly.
18705 *Ralf S. Engelschall*
18707 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18711 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18712 on the `perl Configure ...` command line. This way one can compile
18713 OpenSSL libraries with Position Independent Code (PIC) which is needed
18714 for linking it into DSOs.
18716 *Ralf S. Engelschall*
18718 * Remarkably, export ciphers were totally broken and no-one had noticed!
18723 * Cleaned up the LICENSE document: The official contact for any license
18724 questions now is the OpenSSL core team under openssl-core@openssl.org.
18725 And add a paragraph about the dual-license situation to make sure people
18726 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18727 to the OpenSSL toolkit.
18729 *Ralf S. Engelschall*
18731 * General source tree makefile cleanups: Made `making xxx in yyy...`
18732 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18733 Additionally cleaned up the `make links` target: Remove unnecessary
18734 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18735 to speed processing and no longer clutter the display with confusing
18736 stuff. Instead only the actually done links are displayed.
18738 *Ralf S. Engelschall*
18740 * Permit null encryption ciphersuites, used for authentication only. It used
18741 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18742 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18747 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18748 signed attributes when verifying signatures (this would break them),
18749 the detached data encoding was wrong and public keys obtained using
18750 X509_get_pubkey() weren't freed.
18754 * Add text documentation for the BUFFER functions. Also added a work around
18755 to a Win95 console bug. This was triggered by the password read stuff: the
18756 last character typed gets carried over to the next fread(). If you were
18757 generating a new cert request using 'req' for example then the last
18758 character of the passphrase would be CR which would then enter the first
18763 * Added the new 'Includes OpenSSL Cryptography Software' button as
18764 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18765 button and can be used by applications based on OpenSSL to show the
18766 relationship to the OpenSSL project.
18768 *Ralf S. Engelschall*
18770 * Remove confusing variables in function signatures in files
18771 ssl/ssl_lib.c and ssl/ssl.h.
18773 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18775 * Don't install bss_file.c under PREFIX/include/
18777 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18779 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18780 functions that return function pointers and has support for NT specific
18781 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18782 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18783 unsigned to signed types: this was killing the Win32 compile.
18787 * Add new certificate file to stack functions,
18788 SSL_add_dir_cert_subjects_to_stack() and
18789 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18790 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18791 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18792 This means that Apache-SSL and similar packages don't have to mess around
18793 to add as many CAs as they want to the preferred list.
18797 * Experiment with doxygen documentation. Currently only partially applied to
18799 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18800 openssl.doxy as the configuration file.
18804 * Get rid of remaining C++-style comments which strict C compilers hate.
18806 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18808 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18809 compiled in by default: it has problems with large keys.
18813 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18814 DH private keys and/or callback functions which directly correspond to
18815 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18816 is needed for applications which have to configure certificates on a
18817 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18819 For the RSA certificate situation is makes no difference, but
18820 for the DSA certificate situation this fixes the "no shared cipher"
18821 problem where the OpenSSL cipher selection procedure failed because the
18822 temporary keys were not overtaken from the context and the API provided
18823 no way to reconfigure them.
18824 The new functions now let applications reconfigure the stuff and they
18825 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18826 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18827 non-public-API function ssl_cert_instantiate() is used as a helper
18828 function and also to reduce code redundancy inside ssl_rsa.c.
18830 *Ralf S. Engelschall*
18832 * Move s_server -dcert and -dkey options out of the undocumented feature
18833 area because they are useful for the DSA situation and should be
18834 recognized by the users.
18836 *Ralf S. Engelschall*
18838 * Fix the cipher decision scheme for export ciphers: the export bits are
18839 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18840 SSL_EXP_MASK. So, the original variable has to be used instead of the
18841 already masked variable.
18843 *Richard Levitte <levitte@stacken.kth.se>*
18845 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18847 *Richard Levitte <levitte@stacken.kth.se>*
18849 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18850 from `int` to `unsigned int` because it is a length and initialized by
18851 EVP_DigestFinal() which expects an `unsigned int *`.
18853 *Richard Levitte <levitte@stacken.kth.se>*
18855 * Don't hard-code path to Perl interpreter on shebang line of Configure
18856 script. Instead use the usual Shell->Perl transition trick.
18858 *Ralf S. Engelschall*
18860 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18861 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18862 -noout -modulus` as it's already the case for `openssl rsa -noout
18863 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18864 currently the public key is printed (a decision which was already done by
18865 `openssl dsa -modulus` in the past) which serves a similar purpose.
18866 Additionally the NO_RSA no longer completely removes the whole -modulus
18867 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18870 *Ralf S. Engelschall*
18872 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18873 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18875 *Arne Ansper <arne@ats.cyber.ee>*
18877 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18878 to be added. Now both 'req' and 'ca' can use new objects defined in the
18883 * Add cool BIO that does syslog (or event log on NT).
18885 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18887 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18888 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18889 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18890 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18894 * Add preliminary config info for new extension code.
18898 * Make RSA_NO_PADDING really use no padding.
18900 *Ulf Moeller <ulf@fitug.de>*
18902 * Generate errors when private/public key check is done.
18906 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18907 for some CRL extensions and new objects added.
18911 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18912 key usage extension and fuller support for authority key id.
18916 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18917 padding method for RSA, which is recommended for new applications in PKCS
18918 #1 v2.0 (RFC 2437, October 1998).
18919 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18920 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18921 against Bleichbacher's attack on RSA.
18922 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18925 * Updates to the new SSL compression code
18927 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18929 * Fix so that the version number in the master secret, when passed
18930 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18931 (because the server will not accept higher), that the version number
18932 is 0x03,0x01, not 0x03,0x00
18934 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18936 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18937 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18938 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18942 * Support for RAW extensions where an arbitrary extension can be
18943 created by including its DER encoding. See `apps/openssl.cnf` for
18948 * Make sure latest Perl versions don't interpret some generated C array
18949 code as Perl array code in the crypto/err/err_genc.pl script.
18951 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18953 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18954 not many people have the assembler. Various Win32 compilation fixes and
18955 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18956 build instructions.
18960 * Modify configure script 'Configure' to automatically create crypto/date.h
18961 file under Win32 and also build pem.h from pem.org. New script
18962 util/mkfiles.pl to create the MINFO file on environments that can't do a
18963 'make files': perl util/mkfiles.pl >MINFO should work.
18967 * Major rework of DES function declarations, in the pursuit of correctness
18968 and purity. As a result, many evil casts evaporated, and some weirdness,
18969 too. You may find this causes warnings in your code. Zapping your evil
18970 casts will probably fix them. Mostly.
18974 * Fix for a typo in asn1.h. Bug fix to object creation script
18975 obj_dat.pl. It considered a zero in an object definition to mean
18976 "end of object": none of the objects in objects.h have any zeros
18977 so it wasn't spotted.
18979 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18981 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18982 Masking (CBCM). In the absence of test vectors, the best I have been able
18983 to do is check that the decrypt undoes the encrypt, so far. Send me test
18984 vectors if you have them.
18988 * Correct calculation of key length for export ciphers (too much space was
18989 allocated for null ciphers). This has not been tested!
18993 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18994 message is now correct (it understands "crypto" and "ssl" on its
18995 command line). There is also now an "update" option. This will update
18996 the util/ssleay.num and util/libeay.num files with any new functions.
18998 perl util/mkdef.pl crypto ssl update
18999 it will update them.
19003 * Overhauled the Perl interface:
19004 - ported BN stuff to OpenSSL's different BN library
19005 - made the perl/ source tree CVS-aware
19006 - renamed the package from SSLeay to OpenSSL (the files still contain
19007 their history because I've copied them in the repository)
19008 - removed obsolete files (the test scripts will be replaced
19009 by better Test::Harness variants in the future)
19011 *Ralf S. Engelschall*
19013 * First cut for a very conservative source tree cleanup:
19014 1. merge various obsolete readme texts into doc/ssleay.txt
19015 where we collect the old documents and readme texts.
19016 2. remove the first part of files where I'm already sure that we no
19017 longer need them because of three reasons: either they are just temporary
19018 files which were left by Eric or they are preserved original files where
19019 I've verified that the diff is also available in the CVS via "cvs diff
19020 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19021 the crypto/md/ stuff).
19023 *Ralf S. Engelschall*
19025 * More extension code. Incomplete support for subject and issuer alt
19026 name, issuer and authority key id. Change the i2v function parameters
19027 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19028 what that's for :-) Fix to ASN1 macro which messed up
19029 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19033 * Preliminary support for ENUMERATED type. This is largely copied from the
19038 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19040 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19042 * Make sure `make rehash` target really finds the `openssl` program.
19044 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19046 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19047 like to hear about it if this slows down other processors.
19051 * Add CygWin32 platform information to Configure script.
19053 *Alan Batie <batie@aahz.jf.intel.com>*
19055 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19057 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19059 * New program nseq to manipulate netscape certificate sequences
19063 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19068 * Fixes to BN code. Previously the default was to define BN_RECURSION
19069 but the BN code had some problems that would cause failures when
19070 doing certificate verification and some other functions.
19072 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19074 * Add ASN1 and PEM code to support netscape certificate sequences.
19078 * Add ASN1 and PEM code to support netscape certificate sequences.
19082 * Add several PKIX and private extended key usage OIDs.
19086 * Modify the 'ca' program to handle the new extension code. Modify
19087 openssl.cnf for new extension format, add comments.
19091 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19092 and add a sample to openssl.cnf so req -x509 now adds appropriate
19097 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19098 error code, add initial support to X509_print() and x509 application.
19102 * Takes a deep breath and start adding X509 V3 extension support code. Add
19103 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19104 stuff is currently isolated and isn't even compiled yet.
19108 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19109 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19110 Removed the versions check from X509 routines when loading extensions:
19111 this allows certain broken certificates that don't set the version
19112 properly to be processed.
19116 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19117 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19118 can still be regenerated with "make depend".
19122 * Spelling mistake in C version of CAST-128.
19124 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19126 * Changes to the error generation code. The perl script err-code.pl
19127 now reads in the old error codes and retains the old numbers, only
19128 adding new ones if necessary. It also only changes the .err files if new
19129 codes are added. The makefiles have been modified to only insert errors
19130 when needed (to avoid needlessly modifying header files). This is done
19131 by only inserting errors if the .err file is newer than the auto generated
19132 C file. To rebuild all the error codes from scratch (the old behaviour)
19133 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19134 or delete all the .err files.
19138 * CAST-128 was incorrectly implemented for short keys. The C version has
19139 been fixed, but is untested. The assembler versions are also fixed, but
19140 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19141 to regenerate it if needed.
19142 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19143 Hagino <itojun@kame.net>*
19145 * File was opened incorrectly in randfile.c.
19147 *Ulf Möller <ulf@fitug.de>*
19149 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19150 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19151 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19152 al: it's just almost always a UTCTime. Note this patch adds new error
19153 codes so do a "make errors" if there are problems.
19157 * Correct Linux 1 recognition in config.
19159 *Ulf Möller <ulf@fitug.de>*
19161 * Remove pointless MD5 hash when using DSA keys in ca.
19163 *Anonymous <nobody@replay.com>*
19165 * Generate an error if given an empty string as a cert directory. Also
19166 generate an error if handed NULL (previously returned 0 to indicate an
19167 error, but didn't set one).
19169 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19171 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19175 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19176 parameters. This was causing a warning which killed off the Win32 compile.
19180 * Remove C++ style comments from crypto/bn/bn_local.h.
19182 *Neil Costigan <neil.costigan@celocom.com>*
19184 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19185 based on a text string, looking up short and long names and finally
19186 "dot" format. The "dot" format stuff didn't work. Added new function
19187 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19188 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19189 OID is not part of the table.
19193 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19194 X509_LOOKUP_by_alias().
19198 * Sort openssl functions by name.
19202 * Get the `gendsa` command working and add it to the `list` command. Remove
19203 encryption from sample DSA keys (in case anyone is interested the password
19208 * Make *all* `*_free` functions accept a NULL pointer.
19210 *Frans Heymans <fheymans@isaserver.be>*
19212 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19215 *Anonymous <nobody@replay.com>*
19217 * s_server should send the CAfile as acceptable CAs, not its own cert.
19219 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19221 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19223 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19225 * Temp key "for export" tests were wrong in s3_srvr.c.
19227 *Anonymous <nobody@replay.com>*
19229 * Add prototype for temp key callback functions
19230 SSL_CTX_set_tmp_{rsa,dh}_callback().
19234 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19235 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19239 * X509_name_add_entry() freed the wrong thing after an error.
19241 *Arne Ansper <arne@ats.cyber.ee>*
19243 * rsa_eay.c would attempt to free a NULL context.
19245 *Arne Ansper <arne@ats.cyber.ee>*
19247 * BIO_s_socket() had a broken should_retry() on Windoze.
19249 *Arne Ansper <arne@ats.cyber.ee>*
19251 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19253 *Arne Ansper <arne@ats.cyber.ee>*
19255 * Make sure the already existing X509_STORE->depth variable is initialized
19256 in X509_STORE_new(), but document the fact that this variable is still
19257 unused in the certificate verification process.
19259 *Ralf S. Engelschall*
19261 * Fix the various library and `apps/` files to free up pkeys obtained from
19262 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19266 * Fix reference counting in X509_PUBKEY_get(). This makes
19267 demos/maurice/example2.c work, amongst others, probably.
19269 *Steve Henson and Ben Laurie*
19271 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19272 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19273 are no longer created. This way we have a single and consistent command
19274 line interface `openssl <command>`, similar to `cvs <command>`.
19276 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19278 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19279 BIT STRING wrapper always have zero unused bits.
19283 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19287 * Make the top-level INSTALL documentation easier to understand.
19291 * Makefiles updated to exit if an error occurs in a sub-directory
19292 make (including if user presses ^C) [Paul Sutton]
19294 * Make Montgomery context stuff explicit in RSA data structure.
19298 * Fix build order of pem and err to allow for generated pem.h.
19302 * Fix renumbering bug in X509_NAME_delete_entry().
19306 * Enhanced the err-ins.pl script so it makes the error library number
19307 global and can add a library name. This is needed for external ASN1 and
19308 other error libraries.
19312 * Fixed sk_insert which never worked properly.
19316 * Fix ASN1 macros so they can handle indefinite length constructed
19317 EXPLICIT tags. Some non standard certificates use these: they can now
19322 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19323 into a single doc/ssleay.txt bundle. This way the information is still
19324 preserved but no longer messes up this directory. Now it's new room for
19325 the new set of documentation files.
19327 *Ralf S. Engelschall*
19329 * SETs were incorrectly DER encoded. This was a major pain, because they
19330 shared code with SEQUENCEs, which aren't coded the same. This means that
19331 almost everything to do with SETs or SEQUENCEs has either changed name or
19332 number of arguments.
19334 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19336 * Fix test data to work with the above.
19340 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19341 was already fixed by Eric for 0.9.1 it seems.
19343 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19345 * Autodetect FreeBSD3.
19349 * Fix various bugs in Configure. This affects the following platforms:
19353 unixware-2.0-pentium
19358 * Eliminate generated files from CVS. Reorder tests to regenerate files
19359 before they are needed.
19363 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19367 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19369 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19370 changed SSLeay to OpenSSL in version strings.
19372 *Ralf S. Engelschall*
19374 * Some fixups to the top-level documents.
19378 * Fixed the nasty bug where rsaref.h was not found under compile-time
19379 because the symlink to include/ was missing.
19381 *Ralf S. Engelschall*
19383 * Incorporated the popular no-RSA/DSA-only patches
19384 which allow to compile a RSA-free SSLeay.
19386 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19388 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19389 when "ssleay" is still not found.
19391 *Ralf S. Engelschall*
19393 * Added more platforms to Configure: Cray T3E, HPUX 11,
19395 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19397 * Updated the README file.
19399 *Ralf S. Engelschall*
19401 * Added various .cvsignore files in the CVS repository subdirs
19402 to make a "cvs update" really silent.
19404 *Ralf S. Engelschall*
19406 * Recompiled the error-definition header files and added
19407 missing symbols to the Win32 linker tables.
19409 *Ralf S. Engelschall*
19411 * Cleaned up the top-level documents;
19412 o new files: CHANGES and LICENSE
19413 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19414 o merged COPYRIGHT into LICENSE
19415 o removed obsolete TODO file
19416 o renamed MICROSOFT to INSTALL.W32
19418 *Ralf S. Engelschall*
19420 * Removed dummy files from the 0.9.1b source tree:
19421 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19422 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19423 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19424 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19425 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19427 *Ralf S. Engelschall*
19429 * Added various platform portability fixes.
19433 * The Genesis of the OpenSSL rpject:
19434 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19435 Young and Tim J. Hudson created while they were working for C2Net until
19438 *The OpenSSL Project*
19440 ### Changes between 0.9.0b and 0.9.1b [not released]
19442 * Updated a few CA certificates under certs/
19446 * Changed some BIGNUM api stuff.
19450 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19451 DGUX x86, Linux Alpha, etc.
19455 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19456 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19461 * Add -strparse option to asn1pars program which parses nested
19464 *Dr Stephen Henson <shenson@bigfoot.com>*
19466 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19470 * DSA fix for "ca" program.
19474 * Added "-genkey" option to "dsaparam" program.
19478 * Added RIPE MD160 (rmd160) message digest.
19482 * Added -a (all) option to "ssleay version" command.
19486 * Added PLATFORM define which is the id given to Configure.
19490 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19494 * Extended the ASN.1 parser routines.
19498 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19502 * Added a BN_CTX to the BN library.
19506 * Fixed the weak key values in DES library
19510 * Changed API in EVP library for cipher aliases.
19514 * Added support for RC2/64bit cipher.
19518 * Converted the lhash library to the crypto/mem.c functions.
19522 * Added more recognized ASN.1 object ids.
19526 * Added more RSA padding checks for SSL/TLS.
19530 * Added BIO proxy/filter functionality.
19534 * Added extra_certs to SSL_CTX which can be used
19535 send extra CA certificates to the client in the CA cert chain sending
19536 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19540 * Now Fortezza is denied in the authentication phase because
19541 this is key exchange mechanism is not supported by SSLeay at all.
19545 * Additional PKCS1 checks.
19549 * Support the string "TLSv1" for all TLS v1 ciphers.
19553 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19554 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19558 * Fixed a few memory leaks.
19562 * Fixed various code and comment typos.
19566 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19567 bytes sent in the client random.
19569 *Edward Bishop <ebishop@spyglass.com>*
19573 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19574 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19575 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19576 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19577 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19578 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19579 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19580 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19581 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19582 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19583 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19584 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19585 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19586 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19587 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19588 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19589 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19590 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19591 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19592 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19593 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19594 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19595 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19596 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19597 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19598 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19599 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19600 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19601 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19602 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19603 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19604 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19605 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19606 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19607 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19608 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19609 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19610 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19611 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19612 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19613 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19614 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19615 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19616 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19617 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19618 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19619 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19620 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19621 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19622 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19623 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19624 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19625 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19626 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19627 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19628 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19629 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19630 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19631 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19632 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19633 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19634 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19635 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19636 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19637 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19638 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19639 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19640 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19641 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19642 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19643 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19644 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19645 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19646 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19647 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19648 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19649 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19650 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19651 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19652 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19653 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19654 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19655 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19656 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19657 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19658 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19659 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19660 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19661 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19662 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19663 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19664 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19665 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19666 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19667 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19668 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19669 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19670 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19671 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19672 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19673 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19674 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19675 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19676 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19677 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19678 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19679 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19680 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19681 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19682 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19683 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19684 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19685 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19686 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19687 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19688 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19689 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19690 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19691 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19692 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19693 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19694 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19695 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19696 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19697 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19698 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19699 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19700 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19701 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19702 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19703 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19704 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19705 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19706 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19707 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19708 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19709 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19710 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19711 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19712 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19713 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19714 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19715 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19716 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19717 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19718 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19719 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19720 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19721 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19722 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19723 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19724 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19725 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19726 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19727 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19728 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19729 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19730 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19731 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19732 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19733 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19734 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655