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 * Added RIPEMD160 to the default provider.
31 * Add support for certificate compression (RFC8879), including
32 library support for Brotli and Zstandard compression.
36 * Add the ability to add custom attributes to PKCS12 files. Add a new API
37 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
38 for a user specified callback and optional argument.
39 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
40 added to the existing STACK_OF attrs.
44 * Major refactor of the libssl record layer
48 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
52 * Add a mac salt length option for the pkcs12 command.
56 * Add more SRTP protection profiles from RFC8723 and RFC8269.
60 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
62 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
64 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
65 supported and enabled.
69 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
70 to the list of ciphersuites providing Perfect Forward Secrecy as
71 required by SECLEVEL >= 3.
73 *Dmitry Belyavskiy, Nicola Tuveri*
75 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
76 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
77 SSL_get0_iana_groups() function-like macro, retrieves the list of
78 supported groups sent by the peer.
79 The function SSL_client_hello_get_extension_order() populates
80 a caller-supplied array with the list of extension types present in the
81 ClientHello, in order of appearance.
85 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
86 to make it possible to use empty passphrase strings.
90 * RNDR and RNDRRS support in provider functions to provide
91 random number generation for Arm CPUs (aarch64).
95 * s_client and s_server apps now explicitly say when the TLS version
96 does not include the renegotiation mechanism. This avoids confusion
97 between that scenario versus when the TLS version includes secure
98 renegotiation but the peer lacks support for it.
102 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
104 *Tomasz Kantecki, Andrey Matyukov*
106 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
107 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
108 of 160 bits and above and less than 224 bits were previously accepted by
109 default but are now no longer allowed. By default TLS compression was
110 already disabled in previous OpenSSL versions. At security level 2 it cannot
115 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
120 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
121 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
122 will need to load the legacy crypto provider.
126 * The various OBJ_* functions have been made thread safe.
130 * CCM8 cipher suites in TLS have been downgraded to security level zero
131 because they use a short authentication tag which lowers their strength.
135 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
140 * Parallel dual-prime 1536/2048-bit modular exponentiation for
141 AVX512_IFMA capable processors.
143 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
145 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
146 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
147 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
148 marked deprecated from OpenSSL 3.2 onwards and can be disabled by defining
149 `OPENSSL_NO_DEPRECATED_3_2`.
151 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
152 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
153 definitions for these functions regardless of whether
154 `OPENSSL_NO_DEPRECATED_3_2` is defined.
156 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
157 functions regardless of whether they are using them. It is recommended that
158 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
162 * When generating safe-prime DH parameters set the recommended private key
163 length equivalent to minimum key lengths as in RFC 7919.
167 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
168 such as adding a trace facility for debugging certificate chain building.
172 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
173 in particular supporting requests for central key generation, generalized
174 polling, and various types of genm/genp exchanges defined in CMP Updates.
178 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
179 like correcting the TLS and proxy support and adding tracing for debugging.
183 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
187 * Fixed and extended `util/check-format.pl` for checking adherence to the
188 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
189 The checks are meanwhile more complete and yield fewer false positives.
193 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
194 sending and receiving multiple messages in a single call. An implementation
195 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
199 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
200 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
201 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
202 paths which are searched for root certificates.
204 The existing `SSL_CERT_DIR` environment variable is deprecated.
205 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
206 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
207 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
208 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
209 for the purposes of determining root certificate stores.
213 * Support for loading root certificates from the Windows certificate store
214 has been added. The support is in the form of a store which recognises the
215 URI string of `org.openssl.winstore://`. This store is enabled by default and
216 can be disabled using the new compile-time option `no-winstore`.
223 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
224 listed here are only a brief description.
225 The migration guide contains more detailed information related to new features,
226 breaking changes, and mappings for the large list of deprecated functions.
228 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
230 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
232 * OpenSSL supports creating a custom cipher via the legacy
233 EVP_CIPHER_meth_new() function and associated function calls. This function
234 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
235 to use the new provider mechanism in order to implement custom ciphers.
237 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
238 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
239 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
240 and decryption initialisation functions). Instead of using the custom cipher
241 directly it incorrectly tries to fetch an equivalent cipher from the
242 available providers. An equivalent cipher is found based on the NID passed to
243 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
244 given cipher. However it is possible for an application to incorrectly pass
245 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
246 is used in this way the OpenSSL encryption/decryption initialisation function
247 will match the NULL cipher as being equivalent and will fetch this from the
248 available providers. This will succeed if the default provider has been
249 loaded (or if a third party provider has been loaded that offers this
250 cipher). Using the NULL cipher means that the plaintext is emitted as the
253 Applications are only affected by this issue if they call
254 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
255 encryption/decryption initialisation function. Applications that only use
256 SSL/TLS are not impacted by this issue.
261 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
266 * Fixed the linux-mips64 Configure target which was missing the
267 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
272 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
277 * Correctly handle a retransmitted ClientHello in DTLS
281 * Fixed detection of ktls support in cross-compile environment on Linux
285 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
290 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
291 report correct results in some cases
295 * Fix UWP builds by defining VirtualLock
299 * For known safe primes use the minimum key length according to RFC 7919.
300 Longer private key sizes unnecessarily raise the cycles needed to compute the
301 shared secret without any increase of the real security. This fixes a
302 regression from 1.1.1 where these shorter keys were generated for the known
307 * Added the loongarch64 target
311 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
312 only passed to the FIPS provider and not to the default or legacy provider.
316 * Fixed reported performance degradation on aarch64. Restored the
317 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
318 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
319 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
320 The new algorithm is still used for 32 bit targets.
324 * Added a missing header for memcmp that caused compilation failure on some
329 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
331 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
332 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
333 This issue makes the RSA implementation with 2048 bit private keys
334 incorrect on such machines and memory corruption will happen during
335 the computation. As a consequence of the memory corruption an attacker
336 may be able to trigger a remote code execution on the machine performing
339 SSL/TLS servers or other servers using 2048 bit RSA private keys running
340 on machines supporting AVX512IFMA instructions of the X86_64 architecture
341 are affected by this issue.
346 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
347 implementation would not encrypt the entirety of the data under some
348 circumstances. This could reveal sixteen bytes of data that was
349 preexisting in the memory that wasn't written. In the special case of
350 "in place" encryption, sixteen bytes of the plaintext would be revealed.
352 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
353 they are both unaffected.
356 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
358 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
360 * In addition to the c_rehash shell command injection identified in
361 CVE-2022-1292, further bugs where the c_rehash script does not
362 properly sanitise shell metacharacters to prevent command injection have been
365 When the CVE-2022-1292 was fixed it was not discovered that there
366 are other places in the script where the file names of certificates
367 being hashed were possibly passed to a command executed through the shell.
369 This script is distributed by some operating systems in a manner where
370 it is automatically executed. On such operating systems, an attacker
371 could execute arbitrary commands with the privileges of the script.
373 Use of the c_rehash script is considered obsolete and should be replaced
374 by the OpenSSL rehash command line tool.
377 *Daniel Fiala, Tomáš Mráz*
379 * Case insensitive string comparison no longer uses locales. It has instead
380 been directly implemented.
384 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
386 * Case insensitive string comparison is reimplemented via new locale-agnostic
387 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
388 comparison. The previous implementation had problems when the Turkish locale
393 * Fixed a bug in the c_rehash script which was not properly sanitising shell
394 metacharacters to prevent command injection. This script is distributed by
395 some operating systems in a manner where it is automatically executed. On
396 such operating systems, an attacker could execute arbitrary commands with the
397 privileges of the script.
399 Use of the c_rehash script is considered obsolete and should be replaced
400 by the OpenSSL rehash command line tool.
405 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
406 certificate on an OCSP response. The bug caused the function in the case
407 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
408 response (meaning a successful verification) even in the case where the
409 response signing certificate fails to verify.
411 It is anticipated that most users of `OCSP_basic_verify` will not use the
412 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
413 a negative value (indicating a fatal error) in the case of a certificate
414 verification failure. The normal expected return value in this case would be
417 This issue also impacts the command line OpenSSL "ocsp" application. When
418 verifying an ocsp response with the "-no_cert_checks" option the command line
419 application will report that the verification is successful even though it
420 has in fact failed. In this case the incorrect successful response will also
421 be accompanied by error messages showing the failure and contradicting the
422 apparently successful result.
427 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
428 AAD data as the MAC key. This made the MAC key trivially predictable.
430 An attacker could exploit this issue by performing a man-in-the-middle attack
431 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
432 that the modified data would still pass the MAC integrity check.
434 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
435 endpoint will always be rejected by the recipient and the connection will
436 fail at that point. Many application protocols require data to be sent from
437 the client to the server first. Therefore, in such a case, only an OpenSSL
438 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
440 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
441 sent in both directions. In this case both clients and servers could be
442 affected, regardless of the application protocol.
444 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
445 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
446 the handshake when using this ciphersuite.
448 The confidentiality of data is not impacted by this issue, i.e. an attacker
449 cannot decrypt data that has been encrypted using this ciphersuite - they can
452 In order for this attack to work both endpoints must legitimately negotiate
453 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
454 OpenSSL 3.0, and is not available within the default provider or the default
455 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
456 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
457 following must have occurred:
459 1) OpenSSL must have been compiled with the (non-default) compile time option
460 enable-weak-ssl-ciphers
462 2) OpenSSL must have had the legacy provider explicitly loaded (either
463 through application code or via configuration)
465 3) The ciphersuite must have been explicitly added to the ciphersuite list
467 4) The libssl security level must have been set to 0 (default is 1)
469 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
471 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
472 others that both endpoints have in common
477 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
478 occupied by the removed hash table entries.
480 This function is used when decoding certificates or keys. If a long lived
481 process periodically decodes certificates or keys its memory usage will
482 expand without bounds and the process might be terminated by the operating
483 system causing a denial of service. Also traversing the empty hash table
484 entries will take increasingly more time.
486 Typically such long lived processes might be TLS clients or TLS servers
487 configured to accept client certificate authentication.
490 *Hugo Landau, Aliaksei Levin*
492 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
493 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
494 statistics are no longer supported. For compatibility, these statistics are
495 still listed in the output but are now always reported as zero.
499 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
501 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
502 for non-prime moduli.
504 Internally this function is used when parsing certificates that contain
505 elliptic curve public keys in compressed form or explicit elliptic curve
506 parameters with a base point encoded in compressed form.
508 It is possible to trigger the infinite loop by crafting a certificate that
509 has invalid explicit curve parameters.
511 Since certificate parsing happens prior to verification of the certificate
512 signature, any process that parses an externally supplied certificate may thus
513 be subject to a denial of service attack. The infinite loop can also be
514 reached when parsing crafted private keys as they can contain explicit
515 elliptic curve parameters.
517 Thus vulnerable situations include:
519 - TLS clients consuming server certificates
520 - TLS servers consuming client certificates
521 - Hosting providers taking certificates or private keys from customers
522 - Certificate authorities parsing certification requests from subscribers
523 - Anything else which parses ASN.1 elliptic curve parameters
525 Also any other applications that use the BN_mod_sqrt() where the attacker
526 can control the parameter values are vulnerable to this DoS issue.
531 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
532 to the list of ciphersuites providing Perfect Forward Secrecy as
533 required by SECLEVEL >= 3.
535 *Dmitry Belyavskiy, Nicola Tuveri*
537 * Made the AES constant time code for no-asm configurations
538 optional due to the resulting 95% performance degradation.
539 The AES constant time code can be enabled, for no assembly
540 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
544 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
549 * The negative return value handling of the certificate verification callback
550 was reverted. The replacement is to set the verification retry state with
551 the SSL_set_retry_verify() function.
555 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
557 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
558 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
559 verify a certificate supplied by a server. That function may return a
560 negative return value to indicate an internal error (for example out of
561 memory). Such a negative return value is mishandled by OpenSSL and will cause
562 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
563 success and a subsequent call to SSL_get_error() to return the value
564 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
565 returned by OpenSSL if the application has previously called
566 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
567 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
568 totally unexpected and applications may not behave correctly as a result. The
569 exact behaviour will depend on the application but it could result in
570 crashes, infinite loops or other similar incorrect responses.
572 This issue is made more serious in combination with a separate bug in OpenSSL
573 3.0 that will cause X509_verify_cert() to indicate an internal error when
574 processing a certificate chain. This will occur where a certificate does not
575 include the Subject Alternative Name extension but where a Certificate
576 Authority has enforced name constraints. This issue can occur even with valid
582 * Corrected a few file name and file reference bugs in the build,
583 installation and setup scripts, which lead to installation verification
584 failures. Slightly enhanced the installation verification script.
588 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
593 * Fixed PVK encoder to properly query for the passphrase.
597 * Multiple fixes in the OSSL_HTTP API functions.
601 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
602 OSSL_PARAM_INTEGER data type and return error on negative numbers
603 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
604 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
608 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
612 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
616 * Multiple threading fixes.
620 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
624 * Allow fetching an operation from the provider that owns an unexportable key
625 as a fallback if that is still allowed by the property query.
629 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
631 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
636 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
637 S390X capability vector to zero. This simplifies testing of different code
638 paths on S390X architecture.
642 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
643 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
644 SP 800-38D". The communication will fail at this point.
648 * The EC_GROUP_clear_free() function is deprecated as there is nothing
649 confidential in EC_GROUP data.
653 * The byte order mark (BOM) character is ignored if encountered at the
654 beginning of a PEM-formatted file.
658 * Added CMS support for the Russian GOST algorithms.
662 * Due to move of the implementation of cryptographic operations
663 to the providers, validation of various operation parameters can
664 be postponed until the actual operation is executed where previously
665 it happened immediately when an operation parameter was set.
667 For example when setting an unsupported curve with
668 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
669 fail but later keygen operations with the EVP_PKEY_CTX will fail.
671 *OpenSSL team members and many third party contributors*
673 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
674 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
675 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
676 instead to retrieve these algorithms from a provider.
680 * On build targets where the multilib postfix is set in the build
681 configuration the libdir directory was changing based on whether
682 the lib directory with the multilib postfix exists on the system
683 or not. This unpredictable behavior was removed and eventual
684 multilib postfix is now always added to the default libdir. Use
685 `--libdir=lib` to override the libdir if adding the postfix is
690 * The triple DES key wrap functionality now conforms to RFC 3217 but is
691 no longer interoperable with OpenSSL 1.1.1.
695 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
696 function codes, this function can only cause problems for calling
701 * Add a configurable flag to output date formats as ISO 8601. Does not
702 change the default date format.
706 * Version of MSVC earlier than 1300 could get link warnings, which could
707 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
708 Support for this flag has been removed.
712 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
713 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
714 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
715 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
716 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
720 * The signatures of the functions to get and set options on SSL and
721 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
722 Some source code changes may be required.
726 * The public definitions of conf_method_st and conf_st have been
727 deprecated. They will be made opaque in a future release.
729 *Rich Salz and Tomáš Mráz*
731 * Client-initiated renegotiation is disabled by default. To allow it, use
732 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
733 flag, or the "ClientRenegotiation" config parameter as appropriate.
737 * Add "abspath" and "includedir" pragma's to config files, to prevent,
738 or modify relative pathname inclusion.
742 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
743 validated. Please consult the README-FIPS and
744 README-PROVIDERS files, as well as the migration guide.
746 *OpenSSL team members and many third party contributors*
748 * For the key types DH and DHX the allowed settable parameters are now different.
752 * The openssl commands that read keys, certificates, and CRLs now
753 automatically detect the PEM or DER format of the input files.
755 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
757 * Added enhanced PKCS#12 APIs which accept a library context.
761 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
765 * Added support for Kernel TLS (KTLS).
767 *Boris Pismenny, John Baldwin and Andrew Gallatin*
769 * Support for RFC 5746 secure renegotiation is now required by default for
770 SSL or TLS connections to succeed.
774 * The signature of the `copy` functional parameter of the
775 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
776 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
777 the signature of the `pub_decode` functional parameter of the
778 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
779 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
783 * The error return values from some control calls (ctrl) have changed.
787 * A public key check is now performed during EVP_PKEY_derive_set_peer().
791 * Many functions in the EVP_ namespace that are getters of values from
792 implementations or contexts were renamed to include get or get0 in their
793 names. Old names are provided as macro aliases for compatibility and
798 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
799 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
800 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
805 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
808 * The output from the command line applications may have minor
813 * The output from numerous "printing" may have minor changes.
817 * Windows thread synchronization uses read/write primitives (SRWLock) when
818 supported by the OS, otherwise CriticalSection continues to be used.
822 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
823 work on read only BIO source/sinks that do not support these functions.
824 This allows piping or redirection of a file BIO using stdin to be buffered
825 into memory. This is used internally in OSSL_DECODER_from_bio().
829 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
830 this function would return one of the values OSSL_STORE_INFO_NAME,
831 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
832 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
833 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
834 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
835 using this function should be amended to handle the changed return value.
839 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
840 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
841 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
842 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
843 contains more than one certificate identifier: This means that all
844 certificates referenced there MUST be part of the validation chain.
848 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
849 RC5, DESX and DES have been moved to the legacy provider.
853 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
854 RIPEMD-160 have been moved to the legacy provider.
858 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
863 * The deprecated functions EVP_PKEY_get0_RSA(),
864 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
865 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
866 well as the similarly named "get1" functions behave differently in
871 * A number of functions handling low-level keys or engines were deprecated
872 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
873 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
874 EVP_PKEY_get0_siphash().
878 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
879 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
880 will need to load the legacy crypto provider. This includes these PBE
881 algorithms which use this KDF:
882 - NID_pbeWithMD2AndDES_CBC
883 - NID_pbeWithMD5AndDES_CBC
884 - NID_pbeWithSHA1AndRC2_CBC
885 - NID_pbeWithMD2AndRC2_CBC
886 - NID_pbeWithMD5AndRC2_CBC
887 - NID_pbeWithSHA1AndDES_CBC
891 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
892 BIO_debug_callback() functions.
896 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
897 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
901 * The RAND_METHOD APIs have been deprecated.
905 * The SRP APIs have been deprecated.
909 * Add a compile time option to prevent the caching of provider fetched
910 algorithms. This is enabled by including the no-cached-fetch option
911 at configuration time.
915 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
916 count of PKCS12_DEFAULT_ITER.
918 *Tomáš Mráz and Sahana Prasad*
920 * The openssl speed command does not use low-level API calls anymore.
924 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
927 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
929 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
933 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
934 providers may supply their own group implementations (using either the "key
935 exchange" or the "key encapsulation" methods) which will automatically be
936 detected and used by libssl.
938 *Matt Caswell, Nicola Tuveri*
940 * The undocumented function X509_certificate_type() has been deprecated;
944 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
948 * Removed RSA padding mode for SSLv23 (which was only used for
949 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
950 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
955 * Deprecated the obsolete X9.31 RSA key generation related functions.
957 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
958 is not allowed to return a value > 1, this is no more taken as failure.
960 *Viktor Dukhovni and David von Oheimb*
962 * Deprecated the obsolete X9.31 RSA key generation related functions
963 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
964 BN_X931_generate_prime_ex().
968 * The default key generation method for the regular 2-prime RSA keys was
969 changed to the FIPS 186-4 B.3.6 method.
973 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
977 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
981 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
982 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
984 *Rich Salz, Richard Levitte, and David von Oheimb*
986 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
990 * Deprecated `OCSP_parse_url()`.
994 * Validation of SM2 keys has been separated from the validation of regular EC
999 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1000 switches: a validation failure triggers an early exit, returning a failure
1001 exit status to the parent process.
1005 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1006 to ignore unknown ciphers.
1010 * The `-cipher-commands` and `-digest-commands` options
1011 of the command line utility `list` have been deprecated.
1012 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1016 * Added convenience functions for generating asymmetric key pairs:
1017 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1018 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1022 * All of the low-level EC_KEY functions have been deprecated.
1024 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1026 * Deprecated all the libcrypto and libssl error string loading
1031 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1032 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1037 * The `-crypt` option to the `passwd` command line tool has been removed.
1041 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1046 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1050 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1051 EVP_PKEY_get1_tls_encodedpoint().
1055 * The security callback, which can be customised by application code, supports
1056 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1057 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1061 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1062 interface. Their functionality remains unchanged.
1066 * Added new option for 'openssl list', '-providers', which will display the
1067 list of loaded providers, their names, version and status. It optionally
1068 displays their gettable parameters.
1072 * Removed EVP_PKEY_set_alias_type().
1076 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1077 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1081 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1082 parameters are still checked for type safety at compile time via helper
1087 * Remove the RAND_DRBG API
1089 *Paul Dale and Matthias St. Pierre*
1091 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1092 as well as actual hostnames.
1096 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1097 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1098 conversely, silently ignore DTLS protocol version bounds when configuring
1099 TLS-based contexts. The commands can be repeated to set bounds of both
1100 types. The same applies with the corresponding "min_protocol" and
1101 "max_protocol" command-line switches, in case some application uses both TLS
1104 SSL_CTX instances that are created for a fixed protocol version (e.g.
1105 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1106 attempts to apply bounds to these protocol versions would result in an
1107 error. Now only the "version-flexible" SSL_CTX instances are subject to
1108 limits in configuration files in command-line options.
1112 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1117 * Reworked the recorded ERR codes to make better space for system errors.
1118 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1119 given code is a system error (true) or an OpenSSL error (false).
1123 * Reworked the test perl framework to better allow parallel testing.
1125 *Nicola Tuveri and David von Oheimb*
1127 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1128 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1132 * 'Configure' has been changed to figure out the configuration target if
1133 none is given on the command line. Consequently, the 'config' script is
1134 now only a mere wrapper. All documentation is changed to only mention
1137 *Rich Salz and Richard Levitte*
1139 * Added a library context `OSSL_LIB_CTX` that applications as well as
1140 other libraries can use to form a separate context within which
1141 libcrypto operations are performed.
1145 * Added various `_ex` functions to the OpenSSL API that support using
1146 a non-default `OSSL_LIB_CTX`.
1150 * Handshake now fails if Extended Master Secret extension is dropped
1155 * Dropped interactive mode from the `openssl` program.
1159 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1161 *David von Oheimb and Shane Lontis*
1163 * Deprecated `EC_METHOD_get_field_type()`.
1167 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1168 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1169 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1173 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1177 * Add CAdES-BES signature verification support, mostly derived
1178 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1180 *Filipe Raimundo da Silva*
1182 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1186 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1187 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1191 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1195 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1196 EC_KEY_precompute_mult().
1200 * Deprecated EC_POINTs_mul().
1204 * Removed FIPS_mode() and FIPS_mode_set().
1208 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1212 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1213 EC_POINT_get_Jprojective_coordinates_GFp().
1217 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1218 arrays to be more easily constructed via a series of utility functions.
1219 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1220 the various push functions and finally convert to a passable OSSL_PARAM
1221 array using OSSL_PARAM_BLD_to_param().
1225 * The security strength of SHA1 and MD5 based signatures in TLS has been
1230 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1231 contain a provider side internal key.
1235 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1239 * Project text documents not yet having a proper file name extension
1240 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1241 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1245 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1246 have been converted to Markdown with the goal to produce documents
1247 which not only look pretty when viewed online in the browser, but
1248 remain well readable inside a plain text editor.
1250 To achieve this goal, a 'minimalistic' Markdown style has been applied
1251 which avoids formatting elements that interfere too much with the
1252 reading flow in the text file. For example, it
1254 * avoids [ATX headings][] and uses [setext headings][] instead
1255 (which works for `<h1>` and `<h2>` headings only).
1256 * avoids [inline links][] and uses [reference links][] instead.
1257 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1259 [ATX headings]: https://github.github.com/gfm/#atx-headings
1260 [setext headings]: https://github.github.com/gfm/#setext-headings
1261 [inline links]: https://github.github.com/gfm/#inline-link
1262 [reference links]: https://github.github.com/gfm/#reference-link
1263 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1264 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1266 *Matthias St. Pierre*
1268 * The test suite is changed to preserve results of each test recipe.
1269 A new directory test-runs/ with subdirectories named like the
1270 test recipes are created in the build tree for this purpose.
1274 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1275 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1276 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1278 *David von Oheimb, Martin Peylo*
1280 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1281 It supports arbitrary request and response content types, GET redirection,
1282 TLS, connections via HTTP(S) proxies, connections and exchange via
1283 user-defined BIOs (allowing implicit connections), persistent connections,
1284 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1285 The legacy OCSP-focused (and only partly documented) API
1286 is retained for backward compatibility, while most of it is deprecated.
1290 * Added `util/check-format.pl`, a tool for checking adherence to the
1291 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1292 The checks performed are incomplete and yield some false positives.
1293 Still the tool should be useful for detecting most typical glitches.
1297 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1298 If domain name resolution yields multiple IP addresses all of them are tried
1299 after `connect()` failures.
1303 * All of the low-level RSA functions have been deprecated.
1307 * X509 certificates signed using SHA1 are no longer allowed at security
1312 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1313 modified to use PKEY APIs. These commands are now in maintenance mode
1314 and no new features will be added to them.
1318 * The command line utility rsautl has been deprecated.
1322 * The command line utilities genrsa and rsa have been modified to use PKEY
1323 APIs. They now write PKCS#8 keys by default. These commands are now in
1324 maintenance mode and no new features will be added to them.
1328 * All of the low-level DH functions have been deprecated.
1330 *Paul Dale and Matt Caswell*
1332 * All of the low-level DSA functions have been deprecated.
1336 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1337 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1341 * Deprecated low-level ECDH and ECDSA functions.
1345 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1349 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1350 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1351 a new formulation to include all the things it can be used for,
1352 as well as words of caution.
1356 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1360 * All of the low-level HMAC functions have been deprecated.
1362 *Paul Dale and David von Oheimb*
1364 * Over two thousand fixes were made to the documentation, including:
1365 - Common options (such as -rand/-writerand, TLS version control, etc)
1366 were refactored and point to newly-enhanced descriptions in openssl.pod.
1367 - Added style conformance for all options (with help from Richard Levitte),
1368 documented all reported missing options, added a CI build to check
1369 that all options are documented and that no unimplemented options
1371 - Documented some internals, such as all use of environment variables.
1372 - Addressed all internal broken L<> references.
1376 * All of the low-level CMAC functions have been deprecated.
1380 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1381 functions have been deprecated.
1383 *Paul Dale and David von Oheimb*
1385 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1386 set of functions. The documentation mentioned negative values for some
1387 errors, but this was never the case, so the mention of negative values
1390 Code that followed the documentation and thereby check with something
1391 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1395 * All of the low-level cipher functions have been deprecated.
1397 *Matt Caswell and Paul Dale*
1399 * Removed include/openssl/opensslconf.h.in and replaced it with
1400 include/openssl/configuration.h.in, which differs in not including
1401 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1402 was added to include both.
1404 This allows internal hacks where one might need to modify the set
1405 of configured macros, for example this if deprecated symbols are
1406 still supposed to be available internally:
1408 #include <openssl/configuration.h>
1410 #undef OPENSSL_NO_DEPRECATED
1411 #define OPENSSL_SUPPRESS_DEPRECATED
1413 #include <openssl/macros.h>
1415 This should not be used by applications that use the exported
1416 symbols, as that will lead to linking errors.
1420 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1421 used in exponentiation with 512-bit moduli. No EC algorithms are
1422 affected. Analysis suggests that attacks against 2-prime RSA1024,
1423 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1424 difficult to perform and are not believed likely. Attacks against DH512
1425 are considered just feasible. However, for an attack the target would
1426 have to re-use the DH512 private key, which is not recommended anyway.
1427 Also applications directly using the low-level API BN_mod_exp may be
1428 affected if they use BN_FLG_CONSTTIME.
1433 * Most memory-debug features have been deprecated, and the functionality
1434 replaced with no-ops.
1438 * Added documentation for the STACK API.
1442 * Introduced a new method type and API, OSSL_ENCODER, to represent
1443 generic encoders. These do the same sort of job that PEM writers
1444 and d2i functions do, but with support for methods supplied by
1445 providers, and the possibility for providers to support other
1450 * Introduced a new method type and API, OSSL_DECODER, to represent
1451 generic decoders. These do the same sort of job that PEM readers
1452 and i2d functions do, but with support for methods supplied by
1453 providers, and the possibility for providers to support other
1458 * Added a .pragma directive to the syntax of configuration files, to
1459 allow varying behavior in a supported and predictable manner.
1460 Currently added pragma:
1464 This allows dollar signs to be a keyword character unless it's
1465 followed by a opening brace or parenthesis. This is useful for
1466 platforms where dollar signs are commonly used in names, such as
1467 volume names and system directory names on VMS.
1471 * Added functionality to create an EVP_PKEY from user data.
1475 * Change the interpretation of the '--api' configuration option to
1476 mean that this is a desired API compatibility level with no
1477 further meaning. The previous interpretation, that this would
1478 also mean to remove all deprecated symbols up to and including
1479 the given version, no requires that 'no-deprecated' is also used
1480 in the configuration.
1482 When building applications, the desired API compatibility level
1483 can be set with the OPENSSL_API_COMPAT macro like before. For
1484 API compatibility version below 3.0, the old style numerical
1485 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1486 For version 3.0 and on, the value is expected to be the decimal
1487 value calculated from the major and minor version like this:
1489 MAJOR * 10000 + MINOR * 100
1493 -DOPENSSL_API_COMPAT=30000 For 3.0
1494 -DOPENSSL_API_COMPAT=30200 For 3.2
1496 To hide declarations that are deprecated up to and including the
1497 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1498 given when building the application as well.
1502 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1503 access to certificate and CRL stores via URIs and OSSL_STORE
1506 This adds the following functions:
1508 - X509_LOOKUP_store()
1509 - X509_STORE_load_file()
1510 - X509_STORE_load_path()
1511 - X509_STORE_load_store()
1512 - SSL_add_store_cert_subjects_to_stack()
1513 - SSL_CTX_set_default_verify_store()
1514 - SSL_CTX_load_verify_file()
1515 - SSL_CTX_load_verify_dir()
1516 - SSL_CTX_load_verify_store()
1520 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1521 The presence of this system service is determined at run-time.
1525 * Added functionality to create an EVP_PKEY context based on data
1526 for methods from providers. This takes an algorithm name and a
1527 property query string and simply stores them, with the intent
1528 that any operation that uses this context will use those strings
1529 to fetch the needed methods implicitly, thereby making the port
1530 of application written for pre-3.0 OpenSSL easier.
1534 * The undocumented function NCONF_WIN32() has been deprecated; for
1535 conversion details see the HISTORY section of doc/man5/config.pod
1539 * Introduced the new functions EVP_DigestSignInit_ex() and
1540 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1541 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1542 pages for further details.
1546 * Over two thousand fixes were made to the documentation, including:
1547 adding missing command flags, better style conformance, documentation
1550 *Rich Salz, Richard Levitte*
1552 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1553 X25519, X448, Ed25519 and Ed448.
1557 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1562 * Deprecated the public definition of `ERR_STATE` as well as the function
1563 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1568 * Added ERR functionality to give callers access to the stored function
1569 names that have replaced the older function code based functions.
1571 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1572 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1573 ERR_peek_error_all() and ERR_peek_last_error_all().
1575 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1576 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1577 ERR_func_error_string().
1581 * Extended testing to be verbose for failing tests only. The make variables
1582 VERBOSE_FAILURE or VF can be used to enable this:
1584 $ make VF=1 test # Unix
1585 $ mms /macro=(VF=1) test ! OpenVMS
1586 $ nmake VF=1 test # Windows
1590 * Added the `-copy_extensions` option to the `x509` command for use with
1591 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1592 all extensions in the request are copied to the certificate or vice versa.
1594 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1596 * Added the `-copy_extensions` option to the `req` command for use with
1597 `-x509`. When given with the `copy` or `copyall` argument,
1598 all extensions in the certification request are copied to the certificate.
1602 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1603 they generate are by default RFC 5280 compliant in the following sense:
1604 There is a subjectKeyIdentifier extension with a hash value of the public key
1605 and for not self-signed certs there is an authorityKeyIdentifier extension
1606 with a keyIdentifier field or issuer information identifying the signing key.
1607 This is done unless some configuration overrides the new default behavior,
1608 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1612 * Added several checks to `X509_verify_cert()` according to requirements in
1613 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1614 (which may be done by using the CLI option `-x509_strict`):
1615 * The basicConstraints of CA certificates must be marked critical.
1616 * CA certificates must explicitly include the keyUsage extension.
1617 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1618 * The issuer name of any certificate must not be empty.
1619 * The subject name of CA certs, certs with keyUsage crlSign,
1620 and certs without subjectAlternativeName must not be empty.
1621 * If a subjectAlternativeName extension is given it must not be empty.
1622 * The signatureAlgorithm field and the cert signature must be consistent.
1623 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1624 must not be marked critical.
1625 * The authorityKeyIdentifier must be given for X.509v3 certs
1626 unless they are self-signed.
1627 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1631 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1632 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1636 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1637 used even when parsing explicit parameters, when loading a encoded key
1638 or calling `EC_GROUP_new_from_ecpkparameters()`/
1639 `EC_GROUP_new_from_ecparameters()`.
1640 This prevents bypass of security hardening and performance gains,
1641 especially for curves with specialized EC_METHODs.
1642 By default, if a key encoded with explicit parameters is loaded and later
1643 encoded, the output is still encoded with explicit parameters, even if
1644 internally a "named" EC_GROUP is used for computation.
1648 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1649 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1650 NULL. After this change, only the cofactor parameter can be NULL. It also
1651 does some minimal sanity checks on the passed order.
1656 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1657 An attack is simple, if the first CMS_recipientInfo is valid but the
1658 second CMS_recipientInfo is chosen ciphertext. If the second
1659 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1660 encryption key will be replaced by garbage, and the message cannot be
1661 decoded, but if the RSA decryption fails, the correct encryption key is
1662 used and the recipient will not notice the attack.
1663 As a work around for this potential attack the length of the decrypted
1664 key must be equal to the cipher default key length, in case the
1665 certificate is not given and all recipientInfo are tried out.
1666 The old behaviour can be re-enabled in the CMS code by setting the
1667 CMS_DEBUG_DECRYPT flag.
1671 * Early start up entropy quality from the DEVRANDOM seed source has been
1672 improved for older Linux systems. The RAND subsystem will wait for
1673 /dev/random to be producing output before seeding from /dev/urandom.
1674 The seeded state is stored for future library initialisations using
1675 a system global shared memory segment. The shared memory identifier
1676 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1677 the desired value. The default identifier is 114.
1681 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1682 when primes for RSA keys are computed.
1683 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1684 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1685 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1686 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1687 This avoids possible fingerprinting of newly generated RSA modules.
1691 * Correct the extended master secret constant on EBCDIC systems. Without this
1692 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1693 negotiate EMS will fail. Unfortunately this also means that TLS connections
1694 between EBCDIC systems with this fix, and EBCDIC systems without this
1695 fix will fail if they negotiate EMS.
1699 * Changed the library initialisation so that the config file is now loaded
1700 by default. This was already the case for libssl. It now occurs for both
1701 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1702 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1706 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1707 where the former acts as a replacement for `ERR_put_error()`, and the
1708 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1709 `ERR_raise_data()` adds more flexibility by taking a format string and
1710 an arbitrary number of arguments following it, to be processed with
1715 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1716 to check if a named provider is loaded and available. When called, it
1717 will also activate all fallback providers if such are still present.
1721 * Enforce a minimum DH modulus size of 512 bits.
1725 * Changed DH parameters to generate the order q subgroup instead of 2q.
1726 Previously generated DH parameters are still accepted by DH_check
1727 but DH_generate_key works around that by clearing bit 0 of the
1728 private key for those. This avoids leaking bit 0 of the private key.
1732 * Significantly reduce secure memory usage by the randomness pools.
1736 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1741 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1742 algorithms. An implementation of a key exchange algorithm can be obtained
1743 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1744 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1745 the older EVP_PKEY_derive_init() function. See the man pages for the new
1746 functions for further details.
1750 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1754 * Removed the function names from error messages and deprecated the
1759 * Removed NextStep support and the macro OPENSSL_UNISTD
1763 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1764 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1765 Also removed "export var as function" capability; we do not export
1766 variables, only functions.
1770 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1771 an error and 1 indicating success. In previous versions of OpenSSL this
1772 was a void type. If a key was set longer than the maximum possible this
1777 * Support SM2 signing and verification schemes with X509 certificate.
1781 * Use SHA256 as the default digest for TS query in the `ts` app.
1785 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1789 * Default cipher lists/suites are now available via a function, the
1790 #defines are deprecated.
1794 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1795 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1796 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1800 * Join the directories crypto/x509 and crypto/x509v3
1804 * Added command 'openssl kdf' that uses the EVP_KDF API.
1808 * Added command 'openssl mac' that uses the EVP_MAC API.
1812 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1813 as default directories. Also added the command 'openssl info'
1814 for scripting purposes.
1818 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1823 * Add prediction resistance to the DRBG reseeding process.
1827 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1828 mandated by IEEE Std 1619-2018.
1832 * Added newline escaping functionality to a filename when using openssl dgst.
1833 This output format is to replicate the output format found in the `*sum`
1834 checksum programs. This aims to preserve backward compatibility.
1836 *Matt Eaton, Richard Levitte, and Paul Dale*
1838 * Removed the heartbeat message in DTLS feature, as it has very
1839 little usage and doesn't seem to fulfill a valuable purpose.
1840 The configuration option is now deprecated.
1844 * Changed the output of 'openssl {digestname} < file' to display the
1845 digest name in its output.
1849 * Added a new generic trace API which provides support for enabling
1850 instrumentation through trace output.
1852 *Richard Levitte & Matthias St. Pierre*
1854 * Added build tests for C++. These are generated files that only do one
1855 thing, to include one public OpenSSL head file each. This tests that
1856 the public header files can be usefully included in a C++ application.
1858 This test isn't enabled by default. It can be enabled with the option
1859 'enable-buildtest-c++'.
1863 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1867 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1871 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1875 * Added KMAC to EVP_MAC.
1879 * Added property based algorithm implementation selection framework to
1884 * Added SCA hardening for modular field inversion in EC_GROUP through
1885 a new dedicated field_inv() pointer in EC_METHOD.
1886 This also addresses a leakage affecting conversions from projective
1887 to affine coordinates.
1889 *Billy Bob Brumley, Nicola Tuveri*
1891 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1892 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1893 those algorithms that were already supported through the EVP_PKEY API
1894 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1895 and scrypt are now wrappers that call EVP_KDF.
1899 * Build devcrypto engine as a dynamic engine.
1901 *Eneas U de Queiroz*
1903 * Add keyed BLAKE2 to EVP_MAC.
1907 * Fix a bug in the computation of the endpoint-pair shared secret used
1908 by DTLS over SCTP. This breaks interoperability with older versions
1909 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1910 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1911 interoperability with such broken implementations. However, enabling
1912 this switch breaks interoperability with correct implementations.
1914 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1915 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1919 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1923 * Changed the license to the Apache License v2.0.
1927 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1929 - Major releases (indicated by incrementing the MAJOR release number)
1930 may introduce incompatible API/ABI changes.
1931 - Minor releases (indicated by incrementing the MINOR release number)
1932 may introduce new features but retain API/ABI compatibility.
1933 - Patch releases (indicated by incrementing the PATCH number)
1934 are intended for bug fixes and other improvements of existing
1935 features only (like improving performance or adding documentation)
1936 and retain API/ABI compatibility.
1940 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1944 * Remove the 'dist' target and add a tarball building script. The
1945 'dist' target has fallen out of use, and it shouldn't be
1946 necessary to configure just to create a source distribution.
1950 * Recreate the OS390-Unix config target. It no longer relies on a
1951 special script like it did for OpenSSL pre-1.1.0.
1955 * Instead of having the source directories listed in Configure, add
1956 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1961 * Add GMAC to EVP_MAC.
1965 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1969 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1970 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1971 to facilitate the continued use of MACs through raw private keys in
1972 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1976 * Deprecate ECDH_KDF_X9_62().
1980 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1981 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1982 are retained for backwards compatibility.
1986 * AES-XTS mode now enforces that its two keys are different to mitigate
1987 the attacked described in "Efficient Instantiations of Tweakable
1988 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1989 Details of this attack can be obtained from:
1990 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1994 * Rename the object files, i.e. give them other names than in previous
1995 versions. Their names now include the name of the final product, as
1996 well as its type mnemonic (bin, lib, shlib).
2000 * Added new option for 'openssl list', '-objects', which will display the
2001 list of built in objects, i.e. OIDs with names.
2005 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2006 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2011 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2012 improves application performance by removing data copies and providing
2013 applications with zero-copy system calls such as sendfile and splice.
2017 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2021 * `PKCS12_parse` now maintains the order of the parsed certificates
2022 when outputting them via `*ca` (rather than reversing it).
2026 * Deprecated pthread fork support methods.
2030 * Added support for FFDHE key exchange in TLS 1.3.
2034 * Added a new concept for OpenSSL plugability: providers. This
2035 functionality is designed to replace the ENGINE API and ENGINE
2036 implementations, and to be much more dynamic, allowing provider
2037 authors to introduce new algorithms among other things, as long as
2038 there's an API that supports the algorithm type.
2040 With this concept comes a new core API for interaction between
2041 libcrypto and provider implementations. Public libcrypto functions
2042 that want to use providers do so through this core API.
2044 The main documentation for this core API is found in
2045 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2046 refer to other manuals describing the API specific for supported
2047 algorithm types (also called operations).
2054 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2056 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2058 * Avoid loading of a dynamic engine twice.
2062 * Prioritise DANE TLSA issuer certs over peer certs
2066 * Fixed random API for MacOS prior to 10.12
2068 These MacOS versions don't support the CommonCrypto APIs
2072 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2074 * Fixed an SM2 Decryption Buffer Overflow.
2076 In order to decrypt SM2 encrypted data an application is expected to
2077 call the API function EVP_PKEY_decrypt(). Typically an application will
2078 call this function twice. The first time, on entry, the "out" parameter
2079 can be NULL and, on exit, the "outlen" parameter is populated with the
2080 buffer size required to hold the decrypted plaintext. The application
2081 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2082 again, but this time passing a non-NULL value for the "out" parameter.
2084 A bug in the implementation of the SM2 decryption code means that the
2085 calculation of the buffer size required to hold the plaintext returned
2086 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2087 size required by the second call. This can lead to a buffer overflow
2088 when EVP_PKEY_decrypt() is called by the application a second time with
2089 a buffer that is too small.
2091 A malicious attacker who is able present SM2 content for decryption to
2092 an application could cause attacker chosen data to overflow the buffer
2093 by up to a maximum of 62 bytes altering the contents of other data held
2094 after the buffer, possibly changing application behaviour or causing
2095 the application to crash. The location of the buffer is application
2096 dependent but is typically heap allocated.
2101 * Fixed various read buffer overruns processing ASN.1 strings
2103 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2104 structure which contains a buffer holding the string data and a field
2105 holding the buffer length. This contrasts with normal C strings which
2106 are represented as a buffer for the string data which is terminated
2107 with a NUL (0) byte.
2109 Although not a strict requirement, ASN.1 strings that are parsed using
2110 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2111 well as any string whose value has been set with the ASN1_STRING_set()
2112 function will additionally NUL terminate the byte array in the
2113 ASN1_STRING structure.
2115 However, it is possible for applications to directly construct valid
2116 ASN1_STRING structures which do not NUL terminate the byte array by
2117 directly setting the "data" and "length" fields in the ASN1_STRING
2118 array. This can also happen by using the ASN1_STRING_set0() function.
2120 Numerous OpenSSL functions that print ASN.1 data have been found to
2121 assume that the ASN1_STRING byte array will be NUL terminated, even
2122 though this is not guaranteed for strings that have been directly
2123 constructed. Where an application requests an ASN.1 structure to be
2124 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2125 been directly constructed by the application without NUL terminating
2126 the "data" field, then a read buffer overrun can occur.
2128 The same thing can also occur during name constraints processing
2129 of certificates (for example if a certificate has been directly
2130 constructed by the application instead of loading it via the OpenSSL
2131 parsing functions, and the certificate contains non NUL terminated
2132 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2133 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2135 If a malicious actor can cause an application to directly construct an
2136 ASN1_STRING and then process it through one of the affected OpenSSL
2137 functions then this issue could be hit. This might result in a crash
2138 (causing a Denial of Service attack). It could also result in the
2139 disclosure of private memory contents (such as private keys, or
2140 sensitive plaintext).
2145 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2147 * Fixed a problem with verifying a certificate chain when using the
2148 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2149 the certificates present in a certificate chain. It is not set by default.
2151 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2152 the chain that have explicitly encoded elliptic curve parameters was added
2153 as an additional strict check.
2155 An error in the implementation of this check meant that the result of a
2156 previous check to confirm that certificates in the chain are valid CA
2157 certificates was overwritten. This effectively bypasses the check
2158 that non-CA certificates must not be able to issue other certificates.
2160 If a "purpose" has been configured then there is a subsequent opportunity
2161 for checks that the certificate is a valid CA. All of the named "purpose"
2162 values implemented in libcrypto perform this check. Therefore, where
2163 a purpose is set the certificate chain will still be rejected even when the
2164 strict flag has been used. A purpose is set by default in libssl client and
2165 server certificate verification routines, but it can be overridden or
2166 removed by an application.
2168 In order to be affected, an application must explicitly set the
2169 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2170 for the certificate verification or, in the case of TLS client or server
2171 applications, override the default purpose.
2176 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2177 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2178 renegotiation ClientHello omits the signature_algorithms extension (where it
2179 was present in the initial ClientHello), but includes a
2180 signature_algorithms_cert extension then a NULL pointer dereference will
2181 result, leading to a crash and a denial of service attack.
2183 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2184 (which is the default configuration). OpenSSL TLS clients are not impacted by
2188 *Peter Kästle and Samuel Sapalski*
2190 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2192 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2193 create a unique hash value based on the issuer and serial number data
2194 contained within an X509 certificate. However, it was failing to correctly
2195 handle any errors that may occur while parsing the issuer field (which might
2196 occur if the issuer field is maliciously constructed). This may subsequently
2197 result in a NULL pointer deref and a crash leading to a potential denial of
2203 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2204 padding mode to correctly check for rollback attacks. This is considered a
2205 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2210 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2211 functions. Previously they could overflow the output length argument in some
2212 cases where the input length is close to the maximum permissible length for
2213 an integer on the platform. In such cases the return value from the function
2214 call would be 1 (indicating success), but the output length value would be
2215 negative. This could cause applications to behave incorrectly or crash.
2220 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2221 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2222 could be exploited in a side channel attack to recover the password. Since
2223 the attack is local host only this is outside of the current OpenSSL
2224 threat model and therefore no CVE is assigned.
2226 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2231 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2233 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2234 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2235 If an attacker can control both items being compared then this could lead
2236 to a possible denial of service attack. OpenSSL itself uses the
2237 GENERAL_NAME_cmp function for two purposes:
2238 1) Comparing CRL distribution point names between an available CRL and a
2239 CRL distribution point embedded in an X509 certificate
2240 2) When verifying that a timestamp response token signer matches the
2241 timestamp authority name (exposed via the API functions
2242 TS_RESP_verify_response and TS_RESP_verify_token)
2247 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2249 * Certificates with explicit curve parameters are now disallowed in
2250 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2254 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2255 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2256 conversely, silently ignore DTLS protocol version bounds when configuring
2257 TLS-based contexts. The commands can be repeated to set bounds of both
2258 types. The same applies with the corresponding "min_protocol" and
2259 "max_protocol" command-line switches, in case some application uses both TLS
2262 SSL_CTX instances that are created for a fixed protocol version (e.g.
2263 TLSv1_server_method()) also silently ignore version bounds. Previously
2264 attempts to apply bounds to these protocol versions would result in an
2265 error. Now only the "version-flexible" SSL_CTX instances are subject to
2266 limits in configuration files in command-line options.
2270 * Handshake now fails if Extended Master Secret extension is dropped
2275 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2277 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2279 * Fixed segmentation fault in SSL_check_chain()
2280 Server or client applications that call the SSL_check_chain() function
2281 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2282 dereference as a result of incorrect handling of the
2283 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2284 or unrecognised signature algorithm is received from the peer. This could
2285 be exploited by a malicious peer in a Denial of Service attack.
2290 * Added AES consttime code for no-asm configurations
2291 an optional constant time support for AES was added
2292 when building openssl for no-asm.
2293 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2294 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2295 At this time this feature is by default disabled.
2296 It will be enabled by default in 3.0.
2300 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2302 * Revert the change of EOF detection while reading in libssl to avoid
2303 regressions in applications depending on the current way of reporting
2304 the EOF. As the existing method is not fully accurate the change to
2305 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2306 branch and will be present in the 3.0 release.
2310 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2311 when primes for RSA keys are computed.
2312 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2313 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2314 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2315 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2316 This avoids possible fingerprinting of newly generated RSA modules.
2320 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2322 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2323 while reading in libssl then we would report an error back to the
2324 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2325 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2326 therefore give a hint as to what went wrong.
2330 * Check that ed25519 and ed448 are allowed by the security level. Previously
2331 signature algorithms not using an MD were not being checked that they were
2332 allowed by the security level.
2336 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2337 was not quite right. The behaviour was not consistent between resumption
2338 and normal handshakes, and also not quite consistent with historical
2339 behaviour. The behaviour in various scenarios has been clarified and
2340 it has been updated to make it match historical behaviour as closely as
2345 * *[VMS only]* The header files that the VMS compilers include automatically,
2346 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2347 that the C++ compiler doesn't understand. This is a shortcoming in the
2348 compiler, but can be worked around with `__cplusplus` guards.
2350 C++ applications that use OpenSSL libraries must be compiled using the
2351 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2352 functions. Otherwise, only functions with symbols of less than 31
2353 characters can be used, as the linker will not be able to successfully
2354 resolve symbols with longer names.
2358 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2359 The presence of this system service is determined at run-time.
2363 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2368 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2370 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2371 number generator (RNG). This was intended to include protection in the
2372 event of a fork() system call in order to ensure that the parent and child
2373 processes did not share the same RNG state. However, this protection was not
2374 being used in the default case.
2376 A partial mitigation for this issue is that the output from a high
2377 precision timer is mixed into the RNG state so the likelihood of a parent
2378 and child process sharing state is significantly reduced.
2380 If an application already calls OPENSSL_init_crypto() explicitly using
2381 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2384 *Matthias St. Pierre*
2386 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2387 used even when parsing explicit parameters, when loading a encoded key
2388 or calling `EC_GROUP_new_from_ecpkparameters()`/
2389 `EC_GROUP_new_from_ecparameters()`.
2390 This prevents bypass of security hardening and performance gains,
2391 especially for curves with specialized EC_METHODs.
2392 By default, if a key encoded with explicit parameters is loaded and later
2393 encoded, the output is still encoded with explicit parameters, even if
2394 internally a "named" EC_GROUP is used for computation.
2398 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2399 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2400 NULL. After this change, only the cofactor parameter can be NULL. It also
2401 does some minimal sanity checks on the passed order.
2406 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2407 An attack is simple, if the first CMS_recipientInfo is valid but the
2408 second CMS_recipientInfo is chosen ciphertext. If the second
2409 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2410 encryption key will be replaced by garbage, and the message cannot be
2411 decoded, but if the RSA decryption fails, the correct encryption key is
2412 used and the recipient will not notice the attack.
2413 As a work around for this potential attack the length of the decrypted
2414 key must be equal to the cipher default key length, in case the
2415 certificate is not given and all recipientInfo are tried out.
2416 The old behaviour can be re-enabled in the CMS code by setting the
2417 CMS_DEBUG_DECRYPT flag.
2422 * Early start up entropy quality from the DEVRANDOM seed source has been
2423 improved for older Linux systems. The RAND subsystem will wait for
2424 /dev/random to be producing output before seeding from /dev/urandom.
2425 The seeded state is stored for future library initialisations using
2426 a system global shared memory segment. The shared memory identifier
2427 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2428 the desired value. The default identifier is 114.
2432 * Correct the extended master secret constant on EBCDIC systems. Without this
2433 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2434 negotiate EMS will fail. Unfortunately this also means that TLS connections
2435 between EBCDIC systems with this fix, and EBCDIC systems without this
2436 fix will fail if they negotiate EMS.
2440 * Use Windows installation paths in the mingw builds
2442 Mingw isn't a POSIX environment per se, which means that Windows
2443 paths should be used for installation.
2448 * Changed DH_check to accept parameters with order q and 2q subgroups.
2449 With order 2q subgroups the bit 0 of the private key is not secret
2450 but DH_generate_key works around that by clearing bit 0 of the
2451 private key for those. This avoids leaking bit 0 of the private key.
2455 * Significantly reduce secure memory usage by the randomness pools.
2459 * Revert the DEVRANDOM_WAIT feature for Linux systems
2461 The DEVRANDOM_WAIT feature added a select() call to wait for the
2462 /dev/random device to become readable before reading from the
2463 /dev/urandom device.
2465 It turned out that this change had negative side effects on
2466 performance which were not acceptable. After some discussion it
2467 was decided to revert this feature and leave it up to the OS
2468 resp. the platform maintainer to ensure a proper initialization
2469 during early boot time.
2471 *Matthias St. Pierre*
2473 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2475 * Add build tests for C++. These are generated files that only do one
2476 thing, to include one public OpenSSL head file each. This tests that
2477 the public header files can be usefully included in a C++ application.
2479 This test isn't enabled by default. It can be enabled with the option
2480 'enable-buildtest-c++'.
2484 * Enable SHA3 pre-hashing for ECDSA and DSA.
2488 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2489 This changes the size when using the `genpkey` command when no size is given.
2490 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2491 generation commands to use 2048 bits by default.
2495 * Reorganize the manual pages to consistently have RETURN VALUES,
2496 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2497 util/fix-doc-nits accordingly.
2499 *Paul Yang, Joshua Lock*
2501 * Add the missing accessor EVP_PKEY_get0_engine()
2505 * Have commands like `s_client` and `s_server` output the signature scheme
2506 along with other cipher suite parameters when debugging.
2510 * Make OPENSSL_config() error agnostic again.
2514 * Do the error handling in RSA decryption constant time.
2518 * Prevent over long nonces in ChaCha20-Poly1305.
2520 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2521 for every encryption operation. RFC 7539 specifies that the nonce value
2522 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2523 and front pads the nonce with 0 bytes if it is less than 12
2524 bytes. However it also incorrectly allows a nonce to be set of up to 16
2525 bytes. In this case only the last 12 bytes are significant and any
2526 additional leading bytes are ignored.
2528 It is a requirement of using this cipher that nonce values are
2529 unique. Messages encrypted using a reused nonce value are susceptible to
2530 serious confidentiality and integrity attacks. If an application changes
2531 the default nonce length to be longer than 12 bytes and then makes a
2532 change to the leading bytes of the nonce expecting the new value to be a
2533 new unique nonce then such an application could inadvertently encrypt
2534 messages with a reused nonce.
2536 Additionally the ignored bytes in a long nonce are not covered by the
2537 integrity guarantee of this cipher. Any application that relies on the
2538 integrity of these ignored leading bytes of a long nonce may be further
2539 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2540 is safe because no such use sets such a long nonce value. However user
2541 applications that use this cipher directly and set a non-default nonce
2542 length to be longer than 12 bytes may be vulnerable.
2544 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2550 * Add DEVRANDOM_WAIT feature for Linux systems
2552 On older Linux systems where the getrandom() system call is not available,
2553 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2554 Contrary to getrandom(), the /dev/urandom device will not block during
2555 early boot when the kernel CSPRNG has not been seeded yet.
2557 To mitigate this known weakness, use select() to wait for /dev/random to
2558 become readable before reading from /dev/urandom.
2560 * Ensure that SM2 only uses SM3 as digest algorithm
2564 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2566 * Change the info callback signals for the start and end of a post-handshake
2567 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2568 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2569 confused by this and assume that a TLSv1.2 renegotiation has started. This
2570 can break KeyUpdate handling. Instead we no longer signal the start and end
2571 of a post handshake message exchange (although the messages themselves are
2572 still signalled). This could break some applications that were expecting
2573 the old signals. However without this KeyUpdate is not usable for many
2578 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2580 * Timing vulnerability in DSA signature generation
2582 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2583 timing side channel attack. An attacker could use variations in the signing
2584 algorithm to recover the private key.
2586 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2591 * Timing vulnerability in ECDSA signature generation
2593 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2594 timing side channel attack. An attacker could use variations in the signing
2595 algorithm to recover the private key.
2597 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2602 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2603 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2604 of two gigabytes and the error handling improved.
2606 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2607 categorized as a normal bug, not a security issue, because the DRBG reseeds
2608 automatically and is fully functional even without additional randomness
2609 provided by the application.
2611 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2613 * Add a new ClientHello callback. Provides a callback interface that gives
2614 the application the ability to adjust the nascent SSL object at the
2615 earliest stage of ClientHello processing, immediately after extensions have
2616 been collected but before they have been processed. In particular, this
2617 callback can adjust the supported TLS versions in response to the contents
2622 * Add SM2 base algorithm support.
2626 * s390x assembly pack: add (improved) hardware-support for the following
2627 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2628 aes-cfb/cfb8, aes-ecb.
2632 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2633 parameter is no longer accepted, as it leads to a corrupt table. NULL
2634 pem_str is reserved for alias entries only.
2638 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2639 step for prime curves. The new implementation is based on formulae from
2640 differential addition-and-doubling in homogeneous projective coordinates
2641 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2642 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2643 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2644 to work in projective coordinates.
2646 *Billy Bob Brumley, Nicola Tuveri*
2648 * Change generating and checking of primes so that the error rate of not
2649 being prime depends on the intended use based on the size of the input.
2650 For larger primes this will result in more rounds of Miller-Rabin.
2651 The maximal error rate for primes with more than 1080 bits is lowered
2654 *Kurt Roeckx, Annie Yousar*
2656 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2660 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2661 moving between systems, and to avoid confusion when a Windows build is
2662 done with mingw vs with MSVC. For POSIX installs, there's still a
2663 symlink or copy named 'tsget' to avoid that confusion as well.
2667 * Revert blinding in ECDSA sign and instead make problematic addition
2668 length-invariant. Switch even to fixed-length Montgomery multiplication.
2672 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2673 step for binary curves. The new implementation is based on formulae from
2674 differential addition-and-doubling in mixed Lopez-Dahab projective
2675 coordinates, modified to independently blind the operands.
2677 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2679 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2680 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2681 EC_METHODs to implement their own specialized "ladder step", to take
2682 advantage of more favorable coordinate systems or more efficient
2683 differential addition-and-doubling algorithms.
2685 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2687 * Modified the random device based seed sources to keep the relevant
2688 file descriptors open rather than reopening them on each access.
2689 This allows such sources to operate in a chroot() jail without
2690 the associated device nodes being available. This behaviour can be
2691 controlled using RAND_keep_random_devices_open().
2695 * Numerous side-channel attack mitigations have been applied. This may have
2696 performance impacts for some algorithms for the benefit of improved
2697 security. Specific changes are noted in this change log by their respective
2702 * AIX shared library support overhaul. Switch to AIX "natural" way of
2703 handling shared libraries, which means collecting shared objects of
2704 different versions and bitnesses in one common archive. This allows to
2705 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2706 doesn't affect the way 3rd party applications are linked, only how
2707 multi-version installation is managed.
2711 * Make ec_group_do_inverse_ord() more robust and available to other
2712 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2713 mitigations are applied to the fallback BN_mod_inverse().
2714 When using this function rather than BN_mod_inverse() directly, new
2715 EC cryptosystem implementations are then safer-by-default.
2719 * Add coordinate blinding for EC_POINT and implement projective
2720 coordinate blinding for generic prime curves as a countermeasure to
2721 chosen point SCA attacks.
2723 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2725 * Add blinding to ECDSA and DSA signatures to protect against side channel
2726 attacks discovered by Keegan Ryan (NCC Group).
2730 * Enforce checking in the `pkeyutl` command to ensure that the input
2731 length does not exceed the maximum supported digest length when performing
2732 a sign, verify or verifyrecover operation.
2736 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2737 I/O in combination with something like select() or poll() will hang. This
2738 can be turned off again using SSL_CTX_clear_mode().
2739 Many applications do not properly handle non-application data records, and
2740 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2741 around the problems in those applications, but can also break some.
2742 It's recommended to read the manpages about SSL_read(), SSL_write(),
2743 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2744 SSL_CTX_set_read_ahead() again.
2748 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2749 now allow empty (zero character) pass phrases.
2753 * Apply blinding to binary field modular inversion and remove patent
2754 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2758 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2759 binary and prime elliptic curves.
2763 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2764 constant time fixed point multiplication.
2768 * Revise elliptic curve scalar multiplication with timing attack
2769 defenses: ec_wNAF_mul redirects to a constant time implementation
2770 when computing fixed point and variable point multiplication (which
2771 in OpenSSL are mostly used with secret scalars in keygen, sign,
2772 ECDH derive operations).
2773 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2776 * Updated CONTRIBUTING
2780 * Updated DRBG / RAND to request nonce and additional low entropy
2781 randomness from the system.
2783 *Matthias St. Pierre*
2785 * Updated 'openssl rehash' to use OpenSSL consistent default.
2789 * Moved the load of the ssl_conf module to libcrypto, which helps
2790 loading engines that libssl uses before libssl is initialised.
2794 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2798 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2800 *Ingo Schwarze, Rich Salz*
2802 * Added output of accepting IP address and port for 'openssl s_server'
2806 * Added a new API for TLSv1.3 ciphersuites:
2807 SSL_CTX_set_ciphersuites()
2808 SSL_set_ciphersuites()
2812 * Memory allocation failures consistently add an error to the error
2817 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2818 in libcrypto when run as setuid/setgid.
2822 * Load any config file by default when libssl is used.
2826 * Added new public header file <openssl/rand_drbg.h> and documentation
2827 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2829 *Matthias St. Pierre*
2831 * QNX support removed (cannot find contributors to get their approval
2832 for the license change).
2836 * TLSv1.3 replay protection for early data has been implemented. See the
2837 SSL_read_early_data() man page for further details.
2841 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2842 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2843 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2844 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2845 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2846 configuration has been separated out. See the ciphers man page or the
2847 SSL_CTX_set_ciphersuites() man page for more information.
2851 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2852 in responder mode now supports the new "-multi" option, which
2853 spawns the specified number of child processes to handle OCSP
2854 requests. The "-timeout" option now also limits the OCSP
2855 responder's patience to wait to receive the full client request
2856 on a newly accepted connection. Child processes are respawned
2857 as needed, and the CA index file is automatically reloaded
2858 when changed. This makes it possible to run the "ocsp" responder
2859 as a long-running service, making the OpenSSL CA somewhat more
2860 feature-complete. In this mode, most diagnostic messages logged
2861 after entering the event loop are logged via syslog(3) rather than
2866 * Added support for X448 and Ed448. Heavily based on original work by
2871 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2872 objects loaded. This adds the functions OSSL_STORE_expect() and
2873 OSSL_STORE_find() as well as needed tools to construct searches and
2874 get the search data out of them.
2878 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2879 version of OpenSSL should review their configuration settings to ensure
2880 that they are still appropriate for TLSv1.3. For further information see:
2881 <https://wiki.openssl.org/index.php/TLS1.3>
2885 * Grand redesign of the OpenSSL random generator
2887 The default RAND method now utilizes an AES-CTR DRBG according to
2888 NIST standard SP 800-90Ar1. The new random generator is essentially
2889 a port of the default random generator from the OpenSSL FIPS 2.0
2890 object module. It is a hybrid deterministic random bit generator
2891 using an AES-CTR bit stream and which seeds and reseeds itself
2892 automatically using trusted system entropy sources.
2894 Some of its new features are:
2895 - Support for multiple DRBG instances with seed chaining.
2896 - The default RAND method makes use of a DRBG.
2897 - There is a public and private DRBG instance.
2898 - The DRBG instances are fork-safe.
2899 - Keep all global DRBG instances on the secure heap if it is enabled.
2900 - The public and private DRBG instance are per thread for lock free
2903 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2905 * Changed Configure so it only says what it does and doesn't dump
2906 so much data. Instead, ./configdata.pm should be used as a script
2907 to display all sorts of configuration data.
2911 * Added processing of "make variables" to Configure.
2915 * Added SHA512/224 and SHA512/256 algorithm support.
2919 * The last traces of Netware support, first removed in 1.1.0, have
2924 * Get rid of Makefile.shared, and in the process, make the processing
2925 of certain files (rc.obj, or the .def/.map/.opt files produced from
2926 the ordinal files) more visible and hopefully easier to trace and
2927 debug (or make silent).
2931 * Make it possible to have environment variable assignments as
2932 arguments to config / Configure.
2936 * Add multi-prime RSA (RFC 8017) support.
2940 * Add SM3 implemented according to GB/T 32905-2016
2941 *Jack Lloyd <jack.lloyd@ribose.com>,*
2942 *Ronald Tse <ronald.tse@ribose.com>,*
2943 *Erick Borsboom <erick.borsboom@ribose.com>*
2945 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2946 as documented in RFC6066.
2947 Based on a patch from Tomasz Moń
2949 *Filipe Raimundo da Silva*
2951 * Add SM4 implemented according to GB/T 32907-2016.
2952 *Jack Lloyd <jack.lloyd@ribose.com>,*
2953 *Ronald Tse <ronald.tse@ribose.com>,*
2954 *Erick Borsboom <erick.borsboom@ribose.com>*
2956 * Reimplement -newreq-nodes and ERR_error_string_n; the
2957 original author does not agree with the license change.
2961 * Add ARIA AEAD TLS support.
2965 * Some macro definitions to support VS6 have been removed. Visual
2966 Studio 6 has not worked since 1.1.0
2970 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2971 without clearing the errors.
2975 * Add "atfork" functions. If building on a system that without
2976 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2977 requirements. The RAND facility now uses/requires this.
2985 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2986 not possible to disable entirely. However, it's still possible to
2987 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2990 To disable, configure with 'no-ui-console'. 'no-ui' is still
2991 possible to use as an alias. Check at compile time with the
2992 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2993 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2997 * Add a STORE module, which implements a uniform and URI based reader of
2998 stores that can contain keys, certificates, CRLs and numerous other
2999 objects. The main API is loosely based on a few stdio functions,
3000 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3001 OSSL_STORE_error and OSSL_STORE_close.
3002 The implementation uses backends called "loaders" to implement arbitrary
3003 URI schemes. There is one built in "loader" for the 'file' scheme.
3007 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3008 then adjusted to work on FreeBSD 8.4 as well.
3009 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3010 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3014 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3015 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3016 error code calls like this:
3018 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3020 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3021 that can be encoded in C. For the foreseeable future, this will only
3024 *Richard Levitte and Tim Hudson*
3026 * Removed BSD cryptodev engine.
3030 * Add a build target 'build_all_generated', to build all generated files
3031 and only that. This can be used to prepare everything that requires
3032 things like perl for a system that lacks perl and then move everything
3033 to that system and do the rest of the build there.
3037 * In the UI interface, make it possible to duplicate the user data. This
3038 can be used by engines that need to retain the data for a longer time
3039 than just the call where this user data is passed.
3043 * Ignore the '-named_curve auto' value for compatibility of applications
3046 *Tomáš Mráz <tmraz@fedoraproject.org>*
3048 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3049 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3050 alerts across multiple records (some of which could be empty). In practice
3051 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3052 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3053 support this at all. Supporting it adds significant complexity to the
3054 record layer, and its removal is unlikely to cause interoperability
3059 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3060 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3061 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3066 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3067 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3069 *Richard Levitte, Andy Polyakov*
3071 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3076 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3077 platform rather than 'mingw'.
3081 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3082 success if they are asked to add an object which already exists
3083 in the store. This change cascades to other functions which load
3084 certificates and CRLs.
3088 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3089 facilitate stack unwinding even from assembly subroutines.
3093 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3094 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3098 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3099 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3100 which is the minimum version we support.
3104 * Certificate time validation (X509_cmp_time) enforces stricter
3105 compliance with RFC 5280. Fractional seconds and timezone offsets
3106 are no longer allowed.
3110 * Add support for ARIA
3114 * s_client will now send the Server Name Indication (SNI) extension by
3115 default unless the new "-noservername" option is used. The server name is
3116 based on the host provided to the "-connect" option unless overridden by
3117 using "-servername".
3121 * Add support for SipHash
3125 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3126 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3127 prevent issues where no progress is being made and the peer continually
3128 sends unrecognised record types, using up resources processing them.
3132 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3133 using the algorithm defined in
3134 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3138 * Heartbeat support has been removed; the ABI is changed for now.
3140 *Richard Levitte, Rich Salz*
3142 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3146 * The RSA "null" method, which was partially supported to avoid patent
3147 issues, has been replaced to always returns NULL.
3154 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3156 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3157 used even when parsing explicit parameters, when loading a encoded key
3158 or calling `EC_GROUP_new_from_ecpkparameters()`/
3159 `EC_GROUP_new_from_ecparameters()`.
3160 This prevents bypass of security hardening and performance gains,
3161 especially for curves with specialized EC_METHODs.
3162 By default, if a key encoded with explicit parameters is loaded and later
3163 encoded, the output is still encoded with explicit parameters, even if
3164 internally a "named" EC_GROUP is used for computation.
3168 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3169 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3170 NULL. After this change, only the cofactor parameter can be NULL. It also
3171 does some minimal sanity checks on the passed order.
3176 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3177 An attack is simple, if the first CMS_recipientInfo is valid but the
3178 second CMS_recipientInfo is chosen ciphertext. If the second
3179 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3180 encryption key will be replaced by garbage, and the message cannot be
3181 decoded, but if the RSA decryption fails, the correct encryption key is
3182 used and the recipient will not notice the attack.
3183 As a work around for this potential attack the length of the decrypted
3184 key must be equal to the cipher default key length, in case the
3185 certificate is not given and all recipientInfo are tried out.
3186 The old behaviour can be re-enabled in the CMS code by setting the
3187 CMS_DEBUG_DECRYPT flag.
3192 * Use Windows installation paths in the mingw builds
3194 Mingw isn't a POSIX environment per se, which means that Windows
3195 paths should be used for installation.
3200 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3202 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3203 This changes the size when using the `genpkey` command when no size is given.
3204 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3205 generation commands to use 2048 bits by default.
3209 * Prevent over long nonces in ChaCha20-Poly1305.
3211 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3212 for every encryption operation. RFC 7539 specifies that the nonce value
3213 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3214 and front pads the nonce with 0 bytes if it is less than 12
3215 bytes. However it also incorrectly allows a nonce to be set of up to 16
3216 bytes. In this case only the last 12 bytes are significant and any
3217 additional leading bytes are ignored.
3219 It is a requirement of using this cipher that nonce values are
3220 unique. Messages encrypted using a reused nonce value are susceptible to
3221 serious confidentiality and integrity attacks. If an application changes
3222 the default nonce length to be longer than 12 bytes and then makes a
3223 change to the leading bytes of the nonce expecting the new value to be a
3224 new unique nonce then such an application could inadvertently encrypt
3225 messages with a reused nonce.
3227 Additionally the ignored bytes in a long nonce are not covered by the
3228 integrity guarantee of this cipher. Any application that relies on the
3229 integrity of these ignored leading bytes of a long nonce may be further
3230 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3231 is safe because no such use sets such a long nonce value. However user
3232 applications that use this cipher directly and set a non-default nonce
3233 length to be longer than 12 bytes may be vulnerable.
3235 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3241 * Added SCA hardening for modular field inversion in EC_GROUP through
3242 a new dedicated field_inv() pointer in EC_METHOD.
3243 This also addresses a leakage affecting conversions from projective
3244 to affine coordinates.
3246 *Billy Bob Brumley, Nicola Tuveri*
3248 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3249 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3253 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3257 * Remove the 'dist' target and add a tarball building script. The
3258 'dist' target has fallen out of use, and it shouldn't be
3259 necessary to configure just to create a source distribution.
3263 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3265 * Timing vulnerability in DSA signature generation
3267 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3268 timing side channel attack. An attacker could use variations in the signing
3269 algorithm to recover the private key.
3271 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3276 * Timing vulnerability in ECDSA signature generation
3278 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3279 timing side channel attack. An attacker could use variations in the signing
3280 algorithm to recover the private key.
3282 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3287 * Add coordinate blinding for EC_POINT and implement projective
3288 coordinate blinding for generic prime curves as a countermeasure to
3289 chosen point SCA attacks.
3291 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3293 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3295 * Client DoS due to large DH parameter
3297 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3298 malicious server can send a very large prime value to the client. This will
3299 cause the client to spend an unreasonably long period of time generating a
3300 key for this prime resulting in a hang until the client has finished. This
3301 could be exploited in a Denial Of Service attack.
3303 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3308 * Cache timing vulnerability in RSA Key Generation
3310 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3311 a cache timing side channel attack. An attacker with sufficient access to
3312 mount cache timing attacks during the RSA key generation process could
3313 recover the private key.
3315 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3316 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3321 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3322 parameter is no longer accepted, as it leads to a corrupt table. NULL
3323 pem_str is reserved for alias entries only.
3327 * Revert blinding in ECDSA sign and instead make problematic addition
3328 length-invariant. Switch even to fixed-length Montgomery multiplication.
3332 * Change generating and checking of primes so that the error rate of not
3333 being prime depends on the intended use based on the size of the input.
3334 For larger primes this will result in more rounds of Miller-Rabin.
3335 The maximal error rate for primes with more than 1080 bits is lowered
3338 *Kurt Roeckx, Annie Yousar*
3340 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3344 * Add blinding to ECDSA and DSA signatures to protect against side channel
3345 attacks discovered by Keegan Ryan (NCC Group).
3349 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3350 now allow empty (zero character) pass phrases.
3354 * Certificate time validation (X509_cmp_time) enforces stricter
3355 compliance with RFC 5280. Fractional seconds and timezone offsets
3356 are no longer allowed.
3360 * Fixed a text canonicalisation bug in CMS
3362 Where a CMS detached signature is used with text content the text goes
3363 through a canonicalisation process first prior to signing or verifying a
3364 signature. This process strips trailing space at the end of lines, converts
3365 line terminators to CRLF and removes additional trailing line terminators
3366 at the end of a file. A bug in the canonicalisation process meant that
3367 some characters, such as form-feed, were incorrectly treated as whitespace
3368 and removed. This is contrary to the specification (RFC5485). This fix
3369 could mean that detached text data signed with an earlier version of
3370 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3371 signed with a fixed OpenSSL may fail to verify with an earlier version of
3372 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3373 and use the "-binary" flag (for the "cms" command line application) or set
3374 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3378 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3380 * Constructed ASN.1 types with a recursive definition could exceed the stack
3382 Constructed ASN.1 types with a recursive definition (such as can be found
3383 in PKCS7) could eventually exceed the stack given malicious input with
3384 excessive recursion. This could result in a Denial Of Service attack. There
3385 are no such structures used within SSL/TLS that come from untrusted sources
3386 so this is considered safe.
3388 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3394 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3396 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3397 effectively reduced to only comparing the least significant bit of each
3398 byte. This allows an attacker to forge messages that would be considered as
3399 authenticated in an amount of tries lower than that guaranteed by the
3400 security claims of the scheme. The module can only be compiled by the
3401 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3403 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3409 * Add a build target 'build_all_generated', to build all generated files
3410 and only that. This can be used to prepare everything that requires
3411 things like perl for a system that lacks perl and then move everything
3412 to that system and do the rest of the build there.
3416 * Backport SSL_OP_NO_RENGOTIATION
3418 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3419 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3420 changes this is no longer possible in 1.1.0. Therefore, the new
3421 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3422 1.1.0 to provide equivalent functionality.
3424 Note that if an application built against 1.1.0h headers (or above) is run
3425 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3426 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3430 * Removed the OS390-Unix config target. It relied on a script that doesn't
3435 * rsaz_1024_mul_avx2 overflow bug on x86_64
3437 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3438 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3439 Analysis suggests that attacks against RSA and DSA as a result of this
3440 defect would be very difficult to perform and are not believed likely.
3441 Attacks against DH1024 are considered just feasible, because most of the
3442 work necessary to deduce information about a private key may be performed
3443 offline. The amount of resources required for such an attack would be
3444 significant. However, for an attack on TLS to be meaningful, the server
3445 would have to share the DH1024 private key among multiple clients, which is
3446 no longer an option since CVE-2016-0701.
3448 This only affects processors that support the AVX2 but not ADX extensions
3449 like Intel Haswell (4th generation).
3451 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3452 was originally found via the OSS-Fuzz project.
3457 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3459 * bn_sqrx8x_internal carry bug on x86_64
3461 There is a carry propagating bug in the x86_64 Montgomery squaring
3462 procedure. No EC algorithms are affected. Analysis suggests that attacks
3463 against RSA and DSA as a result of this defect would be very difficult to
3464 perform and are not believed likely. Attacks against DH are considered just
3465 feasible (although very difficult) because most of the work necessary to
3466 deduce information about a private key may be performed offline. The amount
3467 of resources required for such an attack would be very significant and
3468 likely only accessible to a limited number of attackers. An attacker would
3469 additionally need online access to an unpatched system using the target
3470 private key in a scenario with persistent DH parameters and a private
3471 key that is shared between multiple clients.
3473 This only affects processors that support the BMI1, BMI2 and ADX extensions
3474 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3476 This issue was reported to OpenSSL by the OSS-Fuzz project.
3481 * Malformed X.509 IPAddressFamily could cause OOB read
3483 If an X.509 certificate has a malformed IPAddressFamily extension,
3484 OpenSSL could do a one-byte buffer overread. The most likely result
3485 would be an erroneous display of the certificate in text format.
3487 This issue was reported to OpenSSL by the OSS-Fuzz project.
3492 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3494 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3495 platform rather than 'mingw'.
3499 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3500 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3501 which is the minimum version we support.
3505 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3507 * Encrypt-Then-Mac renegotiation crash
3509 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3510 negotiated where it was not in the original handshake (or vice-versa) then
3511 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3512 and servers are affected.
3514 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3519 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3521 * Truncated packet could crash via OOB read
3523 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3524 cipher is being used, then a truncated packet can cause that host to
3525 perform an out-of-bounds read, usually resulting in a crash.
3527 This issue was reported to OpenSSL by Robert Święcki of Google.
3532 * Bad (EC)DHE parameters cause a client crash
3534 If a malicious server supplies bad parameters for a DHE or ECDHE key
3535 exchange then this can result in the client attempting to dereference a
3536 NULL pointer leading to a client crash. This could be exploited in a Denial
3539 This issue was reported to OpenSSL by Guido Vranken.
3544 * BN_mod_exp may produce incorrect results on x86_64
3546 There is a carry propagating bug in the x86_64 Montgomery squaring
3547 procedure. No EC algorithms are affected. Analysis suggests that attacks
3548 against RSA and DSA as a result of this defect would be very difficult to
3549 perform and are not believed likely. Attacks against DH are considered just
3550 feasible (although very difficult) because most of the work necessary to
3551 deduce information about a private key may be performed offline. The amount
3552 of resources required for such an attack would be very significant and
3553 likely only accessible to a limited number of attackers. An attacker would
3554 additionally need online access to an unpatched system using the target
3555 private key in a scenario with persistent DH parameters and a private
3556 key that is shared between multiple clients. For example this can occur by
3557 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3558 similar to CVE-2015-3193 but must be treated as a separate problem.
3560 This issue was reported to OpenSSL by the OSS-Fuzz project.
3565 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3567 * ChaCha20/Poly1305 heap-buffer-overflow
3569 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3570 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3571 crash. This issue is not considered to be exploitable beyond a DoS.
3573 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3578 * CMS Null dereference
3580 Applications parsing invalid CMS structures can crash with a NULL pointer
3581 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3582 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3583 structure callback if an attempt is made to free certain invalid encodings.
3584 Only CHOICE structures using a callback which do not handle NULL value are
3587 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3592 * Montgomery multiplication may produce incorrect results
3594 There is a carry propagating bug in the Broadwell-specific Montgomery
3595 multiplication procedure that handles input lengths divisible by, but
3596 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3597 and DH private keys are impossible. This is because the subroutine in
3598 question is not used in operations with the private key itself and an input
3599 of the attacker's direct choice. Otherwise the bug can manifest itself as
3600 transient authentication and key negotiation failures or reproducible
3601 erroneous outcome of public-key operations with specially crafted input.
3602 Among EC algorithms only Brainpool P-512 curves are affected and one
3603 presumably can attack ECDH key negotiation. Impact was not analyzed in
3604 detail, because pre-requisites for attack are considered unlikely. Namely
3605 multiple clients have to choose the curve in question and the server has to
3606 share the private key among them, neither of which is default behaviour.
3607 Even then only clients that chose the curve will be affected.
3609 This issue was publicly reported as transient failures and was not
3610 initially recognized as a security issue. Thanks to Richard Morgan for
3611 providing reproducible case.
3616 * Removed automatic addition of RPATH in shared libraries and executables,
3617 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3621 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3623 * Fix Use After Free for large message sizes
3625 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3626 message larger than approx 16k is received then the underlying buffer to
3627 store the incoming message is reallocated and moved. Unfortunately a
3628 dangling pointer to the old location is left which results in an attempt to
3629 write to the previously freed location. This is likely to result in a
3630 crash, however it could potentially lead to execution of arbitrary code.
3632 This issue only affects OpenSSL 1.1.0a.
3634 This issue was reported to OpenSSL by Robert Święcki.
3639 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3641 * OCSP Status Request extension unbounded memory growth
3643 A malicious client can send an excessively large OCSP Status Request
3644 extension. If that client continually requests renegotiation, sending a
3645 large OCSP Status Request extension each time, then there will be unbounded
3646 memory growth on the server. This will eventually lead to a Denial Of
3647 Service attack through memory exhaustion. Servers with a default
3648 configuration are vulnerable even if they do not support OCSP. Builds using
3649 the "no-ocsp" build time option are not affected.
3651 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3656 * SSL_peek() hang on empty record
3658 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3659 sends an empty record. This could be exploited by a malicious peer in a
3660 Denial Of Service attack.
3662 This issue was reported to OpenSSL by Alex Gaynor.
3667 * Excessive allocation of memory in tls_get_message_header() and
3668 dtls1_preprocess_fragment()
3670 A (D)TLS message includes 3 bytes for its length in the header for the
3671 message. This would allow for messages up to 16Mb in length. Messages of
3672 this length are excessive and OpenSSL includes a check to ensure that a
3673 peer is sending reasonably sized messages in order to avoid too much memory
3674 being consumed to service a connection. A flaw in the logic of version
3675 1.1.0 means that memory for the message is allocated too early, prior to
3676 the excessive message length check. Due to way memory is allocated in
3677 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3678 to service a connection. This could lead to a Denial of Service through
3679 memory exhaustion. However, the excessive message length check still takes
3680 place, and this would cause the connection to immediately fail. Assuming
3681 that the application calls SSL_free() on the failed connection in a timely
3682 manner then the 21Mb of allocated memory will then be immediately freed
3683 again. Therefore, the excessive memory allocation will be transitory in
3684 nature. This then means that there is only a security impact if:
3686 1) The application does not call SSL_free() in a timely manner in the event
3687 that the connection fails
3689 2) The application is working in a constrained environment where there is
3690 very little free memory
3692 3) The attacker initiates multiple connection attempts such that there are
3693 multiple connections in a state where memory has been allocated for the
3694 connection; SSL_free() has not yet been called; and there is insufficient
3695 memory to service the multiple requests.
3697 Except in the instance of (1) above any Denial Of Service is likely to be
3698 transitory because as soon as the connection fails the memory is
3699 subsequently freed again in the SSL_free() call. However there is an
3700 increased risk during this period of application crashes due to the lack of
3701 memory - which would then mean a more serious Denial of Service.
3703 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3704 (CVE-2016-6307 and CVE-2016-6308)
3708 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3709 had to be removed. Primary reason is that vendor assembler can't
3710 assemble our modules with -KPIC flag. As result it, assembly
3711 support, was not even available as option. But its lack means
3712 lack of side-channel resistant code, which is incompatible with
3713 security by todays standards. Fortunately gcc is readily available
3714 prepackaged option, which we firmly point at...
3718 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3720 * Windows command-line tool supports UTF-8 opt-in option for arguments
3721 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3722 (to any value) allows Windows user to access PKCS#12 file generated
3723 with Windows CryptoAPI and protected with non-ASCII password, as well
3724 as files generated under UTF-8 locale on Linux also protected with
3729 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3730 have been disabled by default and removed from DEFAULT, just like RC4.
3731 See the RC4 item below to re-enable both.
3735 * The method for finding the storage location for the Windows RAND seed file
3736 has changed. First we check %RANDFILE%. If that is not set then we check
3737 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3738 all else fails we fall back to C:\.
3742 * The EVP_EncryptUpdate() function has had its return type changed from void
3743 to int. A return of 0 indicates and error while a return of 1 indicates
3748 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3749 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3750 off the constant time implementation for RSA, DSA and DH have been made
3751 no-ops and deprecated.
3755 * Windows RAND implementation was simplified to only get entropy by
3756 calling CryptGenRandom(). Various other RAND-related tickets
3759 *Joseph Wylie Yandle, Rich Salz*
3761 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3762 and `OPENSSL_LH_`, respectively. The old names are available
3763 with API compatibility. They new names are now completely documented.
3767 * Unify TYPE_up_ref(obj) methods signature.
3768 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3769 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3770 int (instead of void) like all others TYPE_up_ref() methods.
3771 So now these methods also check the return value of CRYPTO_atomic_add(),
3772 and the validity of object reference counter.
3774 *fdasilvayy@gmail.com*
3776 * With Windows Visual Studio builds, the .pdb files are installed
3777 alongside the installed libraries and executables. For a static
3778 library installation, ossl_static.pdb is the associate compiler
3779 generated .pdb file to be used when linking programs.
3783 * Remove openssl.spec. Packaging files belong with the packagers.
3787 * Automatic Darwin/OSX configuration has had a refresh, it will now
3788 recognise x86_64 architectures automatically. You can still decide
3789 to build for a different bitness with the environment variable
3790 KERNEL_BITS (can be 32 or 64), for example:
3792 KERNEL_BITS=32 ./config
3796 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3797 256 bit AES and HMAC with SHA256.
3801 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3805 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3809 * To enable users to have their own config files and build file templates,
3810 Configure looks in the directory indicated by the environment variable
3811 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3812 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3813 name and is used as is.
3817 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3818 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3819 X509_CERT_FILE_CTX was removed.
3823 * "shared" builds are now the default. To create only static libraries use
3824 the "no-shared" Configure option.
3828 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3829 All of these option have not worked for some while and are fundamental
3834 * Make various cleanup routines no-ops and mark them as deprecated. Most
3835 global cleanup functions are no longer required because they are handled
3836 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3837 Explicitly de-initing can cause problems (e.g. where a library that uses
3838 OpenSSL de-inits, but an application is still using it). The affected
3839 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3840 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3841 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3842 COMP_zlib_cleanup().
3846 * --strict-warnings no longer enables runtime debugging options
3847 such as REF_DEBUG. Instead, debug options are automatically
3848 enabled with '--debug' builds.
3850 *Andy Polyakov, Emilia Käsper*
3852 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3853 have been moved out of the public header files. New functions for managing
3854 these have been added.
3858 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3859 objects have been moved out of the public header files. New
3860 functions for managing these have been added.
3864 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3865 have been moved out of the public header files. New functions for managing
3866 these have been added.
3870 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3871 moved out of the public header files. New functions for managing these
3876 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3880 * Removed the mk1mf build scripts.
3884 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3885 it is always safe to #include a header now.
3889 * Removed the aged BC-32 config and all its supporting scripts
3893 * Removed support for Ultrix, Netware, and OS/2.
3897 * Add support for HKDF.
3899 *Alessandro Ghedini*
3901 * Add support for blake2b and blake2s
3905 * Added support for "pipelining". Ciphers that have the
3906 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3907 encryptions/decryptions simultaneously. There are currently no built-in
3908 ciphers with this property but the expectation is that engines will be able
3909 to offer it to significantly improve throughput. Support has been extended
3910 into libssl so that multiple records for a single connection can be
3911 processed in one go (for >=TLS 1.1).
3915 * Added the AFALG engine. This is an async capable engine which is able to
3916 offload work to the Linux kernel. In this initial version it only supports
3917 AES128-CBC. The kernel must be version 4.1.0 or greater.
3921 * OpenSSL now uses a new threading API. It is no longer necessary to
3922 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3923 are two supported threading models: pthreads and windows threads. It is
3924 also possible to configure OpenSSL at compile time for "no-threads". The
3925 old threading API should no longer be used. The functions have been
3926 replaced with "no-op" compatibility macros.
3928 *Alessandro Ghedini, Matt Caswell*
3930 * Modify behavior of ALPN to invoke callback after SNI/servername
3931 callback, such that updates to the SSL_CTX affect ALPN.
3935 * Add SSL_CIPHER queries for authentication and key-exchange.
3939 * Changes to the DEFAULT cipherlist:
3940 - Prefer (EC)DHE handshakes over plain RSA.
3941 - Prefer AEAD ciphers over legacy ciphers.
3942 - Prefer ECDSA over RSA when both certificates are available.
3943 - Prefer TLSv1.2 ciphers/PRF.
3944 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3949 * Change the ECC default curve list to be this, in order: x25519,
3950 secp256r1, secp521r1, secp384r1.
3954 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3955 disabled by default. They can be re-enabled using the
3956 enable-weak-ssl-ciphers option to Configure.
3960 * If the server has ALPN configured, but supports no protocols that the
3961 client advertises, send a fatal "no_application_protocol" alert.
3962 This behaviour is SHALL in RFC 7301, though it isn't universally
3963 implemented by other servers.
3967 * Add X25519 support.
3968 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3969 for public and private key encoding using the format documented in
3970 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3971 key generation and key derivation.
3973 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3978 * Deprecate SRP_VBASE_get_by_user.
3979 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3980 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3981 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3982 seed, even if the seed is configured.
3984 Users should use SRP_VBASE_get1_by_user instead. Note that in
3985 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3986 also that even though configuring the SRP seed attempts to hide
3987 invalid usernames by continuing the handshake with fake
3988 credentials, this behaviour is not constant time and no strong
3989 guarantees are made that the handshake is indistinguishable from
3990 that of a valid user.
3994 * Configuration change; it's now possible to build dynamic engines
3995 without having to build shared libraries and vice versa. This
3996 only applies to the engines in `engines/`, those in `crypto/engine/`
3997 will always be built into libcrypto (i.e. "static").
3999 Building dynamic engines is enabled by default; to disable, use
4000 the configuration option "disable-dynamic-engine".
4002 The only requirements for building dynamic engines are the
4003 presence of the DSO module and building with position independent
4004 code, so they will also automatically be disabled if configuring
4005 with "disable-dso" or "disable-pic".
4007 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4008 are also taken away from openssl/opensslconf.h, as they are
4013 * Configuration change; if there is a known flag to compile
4014 position independent code, it will always be applied on the
4015 libcrypto and libssl object files, and never on the application
4016 object files. This means other libraries that use routines from
4017 libcrypto / libssl can be made into shared libraries regardless
4018 of how OpenSSL was configured.
4020 If this isn't desirable, the configuration options "disable-pic"
4021 or "no-pic" can be used to disable the use of PIC. This will
4022 also disable building shared libraries and dynamic engines.
4026 * Removed JPAKE code. It was experimental and has no wide use.
4030 * The INSTALL_PREFIX Makefile variable has been renamed to
4031 DESTDIR. That makes for less confusion on what this variable
4032 is for. Also, the configuration option --install_prefix is
4037 * Heartbeat for TLS has been removed and is disabled by default
4038 for DTLS; configure with enable-heartbeats. Code that uses the
4039 old #define's might need to be updated.
4041 *Emilia Käsper, Rich Salz*
4043 * Rename REF_CHECK to REF_DEBUG.
4047 * New "unified" build system
4049 The "unified" build system is aimed to be a common system for all
4050 platforms we support. With it comes new support for VMS.
4052 This system builds supports building in a different directory tree
4053 than the source tree. It produces one Makefile (for unix family
4054 or lookalikes), or one descrip.mms (for VMS).
4056 The source of information to make the Makefile / descrip.mms is
4057 small files called 'build.info', holding the necessary
4058 information for each directory with source to compile, and a
4059 template in Configurations, like unix-Makefile.tmpl or
4062 With this change, the library names were also renamed on Windows
4063 and on VMS. They now have names that are closer to the standard
4064 on Unix, and include the major version number, and in certain
4065 cases, the architecture they are built for. See "Notes on shared
4066 libraries" in INSTALL.
4068 We rely heavily on the perl module Text::Template.
4072 * Added support for auto-initialisation and de-initialisation of the library.
4073 OpenSSL no longer requires explicit init or deinit routines to be called,
4074 except in certain circumstances. See the OPENSSL_init_crypto() and
4075 OPENSSL_init_ssl() man pages for further information.
4079 * The arguments to the DTLSv1_listen function have changed. Specifically the
4080 "peer" argument is now expected to be a BIO_ADDR object.
4082 * Rewrite of BIO networking library. The BIO library lacked consistent
4083 support of IPv6, and adding it required some more extensive
4084 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4085 which hold all types of addresses and chains of address information.
4086 It also introduces a new API, with functions like BIO_socket,
4087 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4088 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4089 have been adapted accordingly.
4093 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4098 * CRIME protection: disable compression by default, even if OpenSSL is
4099 compiled with zlib enabled. Applications can still enable compression
4100 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4101 using the SSL_CONF library to configure compression.
4105 * The signature of the session callback configured with
4106 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4107 was explicitly marked as `const unsigned char*` instead of
4112 * Always DPURIFY. Remove the use of uninitialized memory in the
4113 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4117 * Removed many obsolete configuration items, including
4118 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4119 MD2_CHAR, MD2_INT, MD2_LONG
4121 IDEA_SHORT, IDEA_LONG
4122 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4124 *Rich Salz, with advice from Andy Polyakov*
4126 * Many BN internals have been moved to an internal header file.
4128 *Rich Salz with help from Andy Polyakov*
4130 * Configuration and writing out the results from it has changed.
4131 Files such as Makefile include/openssl/opensslconf.h and are now
4132 produced through general templates, such as Makefile.in and
4133 crypto/opensslconf.h.in and some help from the perl module
4136 Also, the center of configuration information is no longer
4137 Makefile. Instead, Configure produces a perl module in
4138 configdata.pm which holds most of the config data (in the hash
4139 table %config), the target data that comes from the target
4140 configuration in one of the `Configurations/*.conf` files (in
4145 * To clarify their intended purposes, the Configure options
4146 --prefix and --openssldir change their semantics, and become more
4147 straightforward and less interdependent.
4149 --prefix shall be used exclusively to give the location INSTALLTOP
4150 where programs, scripts, libraries, include files and manuals are
4151 going to be installed. The default is now /usr/local.
4153 --openssldir shall be used exclusively to give the default
4154 location OPENSSLDIR where certificates, private keys, CRLs are
4155 managed. This is also where the default openssl.cnf gets
4157 If the directory given with this option is a relative path, the
4158 values of both the --prefix value and the --openssldir value will
4159 be combined to become OPENSSLDIR.
4160 The default for --openssldir is INSTALLTOP/ssl.
4162 Anyone who uses --openssldir to specify where OpenSSL is to be
4163 installed MUST change to use --prefix instead.
4167 * The GOST engine was out of date and therefore it has been removed. An up
4168 to date GOST engine is now being maintained in an external repository.
4169 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4170 support for GOST ciphersuites (these are only activated if a GOST engine
4175 * EGD is no longer supported by default; use enable-egd when
4178 *Ben Kaduk and Rich Salz*
4180 * The distribution now has Makefile.in files, which are used to
4181 create Makefile's when Configure is run. *Configure must be run
4182 before trying to build now.*
4186 * The return value for SSL_CIPHER_description() for error conditions
4191 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4193 Obtaining and performing DNSSEC validation of TLSA records is
4194 the application's responsibility. The application provides
4195 the TLSA records of its choice to OpenSSL, and these are then
4196 used to authenticate the peer.
4198 The TLSA records need not even come from DNS. They can, for
4199 example, be used to implement local end-entity certificate or
4200 trust-anchor "pinning", where the "pin" data takes the form
4201 of TLSA records, which can augment or replace verification
4202 based on the usual WebPKI public certification authorities.
4206 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4207 continues to support deprecated interfaces in default builds.
4208 However, applications are strongly advised to compile their
4209 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4210 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4211 or the 1.1.0 releases.
4213 In environments in which all applications have been ported to
4214 not use any deprecated interfaces OpenSSL's Configure script
4215 should be used with the --api=1.1.0 option to entirely remove
4216 support for the deprecated features from the library and
4217 unconditionally disable them in the installed headers.
4218 Essentially the same effect can be achieved with the "no-deprecated"
4219 argument to Configure, except that this will always restrict
4220 the build to just the latest API, rather than a fixed API
4223 As applications are ported to future revisions of the API,
4224 they should update their compile-time OPENSSL_API_COMPAT define
4225 accordingly, but in most cases should be able to continue to
4226 compile with later releases.
4228 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4229 0x10000000L and 0x00908000L, respectively. However those
4230 versions did not support the OPENSSL_API_COMPAT feature, and
4231 so applications are not typically tested for explicit support
4232 of just the undeprecated features of either release.
4236 * Add support for setting the minimum and maximum supported protocol.
4237 It can bet set via the SSL_set_min_proto_version() and
4238 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4239 MaxProtocol. It's recommended to use the new APIs to disable
4240 protocols instead of disabling individual protocols using
4241 SSL_set_options() or SSL_CONF's Protocol. This change also
4242 removes support for disabling TLS 1.2 in the OpenSSL TLS
4243 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4247 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4251 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4252 and integrates ECDSA and ECDH functionality into EC. Implementations can
4253 now redirect key generation and no longer need to convert to or from
4256 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4257 include the ec.h header file instead.
4261 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4262 ciphers who are no longer supported and drops support the ephemeral RSA key
4263 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4267 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4268 opaque. For HMAC_CTX, the following constructors and destructors
4271 HMAC_CTX *HMAC_CTX_new(void);
4272 void HMAC_CTX_free(HMAC_CTX *ctx);
4274 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4275 destroy such methods has been added. See EVP_MD_meth_new(3) and
4276 EVP_CIPHER_meth_new(3) for documentation.
4279 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4280 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4281 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4282 an already created structure.
4283 2) For consistency with the majority of our object creators and
4284 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4285 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4286 for deprecated builds.
4290 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4291 cryptographic operations to be performed asynchronously as long as an
4292 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4293 further details. Libssl has also had this capability integrated with the
4294 introduction of the new mode SSL_MODE_ASYNC and associated error
4295 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4296 pages. This work was developed in partnership with Intel Corp.
4300 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4301 always enabled now. If you want to disable the support you should
4302 exclude it using the list of supported ciphers. This also means that the
4303 "-no_ecdhe" option has been removed from s_server.
4307 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4308 SSL_{CTX_}set1_curves() which can set a list.
4312 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4313 curve you want to support using SSL_{CTX_}set1_curves().
4317 * State machine rewrite. The state machine code has been significantly
4318 refactored in order to remove much duplication of code and solve issues
4319 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4320 further details). This change does have some associated API changes.
4321 Notably the SSL_state() function has been removed and replaced by
4322 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4323 SSL_set_state() has been removed altogether. The previous handshake states
4324 defined in ssl.h and ssl3.h have also been removed.
4328 * All instances of the string "ssleay" in the public API were replaced
4329 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4330 Some error codes related to internal RSA_eay API's were renamed.
4334 * The demo files in crypto/threads were moved to demo/threads.
4338 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4341 *Matt Caswell, Rich Salz*
4343 * New ASN.1 embed macro.
4345 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4346 structure is not allocated: it is part of the parent. That is instead of
4354 This reduces memory fragmentation and make it impossible to accidentally
4355 set a mandatory field to NULL.
4357 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4358 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4359 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4364 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4368 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4369 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4370 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4371 DES and RC4 ciphersuites.
4375 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4376 This changes the decoding behaviour for some invalid messages,
4377 though the change is mostly in the more lenient direction, and
4378 legacy behaviour is preserved as much as possible.
4382 * Fix no-stdio build.
4383 *David Woodhouse <David.Woodhouse@intel.com> and also*
4384 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4386 * New testing framework
4387 The testing framework has been largely rewritten and is now using
4388 perl and the perl modules Test::Harness and an extended variant of
4389 Test::More called OpenSSL::Test to do its work. All test scripts in
4390 test/ have been rewritten into test recipes, and all direct calls to
4391 executables in test/Makefile have become individual recipes using the
4392 simplified testing OpenSSL::Test::Simple.
4394 For documentation on our testing modules, do:
4396 perldoc test/testlib/OpenSSL/Test/Simple.pm
4397 perldoc test/testlib/OpenSSL/Test.pm
4401 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4402 are used; the latter aborts on memory leaks (usually checked on exit).
4403 Some undocumented "set malloc, etc., hooks" functions were removed
4404 and others were changed. All are now documented.
4408 * In DSA_generate_parameters_ex, if the provided seed is too short,
4411 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4413 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4414 from RFC4279, RFC4785, RFC5487, RFC5489.
4416 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4417 original RSA_PSK patch.
4421 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4422 era flag was never set throughout the codebase (only read). Also removed
4423 SSL3_FLAGS_POP_BUFFER which was only used if
4424 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4428 * Changed the default name options in the "ca", "crl", "req" and "x509"
4429 to be "oneline" instead of "compat".
4433 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4434 not aware of clients that still exhibit this bug, and the workaround
4435 hasn't been working properly for a while.
4439 * The return type of BIO_number_read() and BIO_number_written() as well as
4440 the corresponding num_read and num_write members in the BIO structure has
4441 changed from unsigned long to uint64_t. On platforms where an unsigned
4442 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4447 * Given the pervasive nature of TLS extensions it is inadvisable to run
4448 OpenSSL without support for them. It also means that maintaining
4449 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4450 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4454 * Removed support for the two export grade static DH ciphersuites
4455 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4456 were newly added (along with a number of other static DH ciphersuites) to
4457 1.0.2. However the two export ones have *never* worked since they were
4458 introduced. It seems strange in any case to be adding new export
4459 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4463 * Version negotiation has been rewritten. In particular SSLv23_method(),
4464 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4465 and turned into macros which simply call the new preferred function names
4466 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4467 should use the new names instead. Also as part of this change the ssl23.h
4468 header file has been removed.
4472 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4473 code and the associated standard is no longer considered fit-for-purpose.
4477 * RT2547 was closed. When generating a private key, try to make the
4478 output file readable only by the owner. This behavior change might
4479 be noticeable when interacting with other software.
4481 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4486 * Added HTTP GET support to the ocsp command.
4490 * Changed default digest for the dgst and enc commands from MD5 to
4495 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4499 * Added support for TLS extended master secret from
4500 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4501 initial patch which was a great help during development.
4505 * All libssl internal structures have been removed from the public header
4506 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4507 now redundant). Users should not attempt to access internal structures
4508 directly. Instead they should use the provided API functions.
4512 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4513 Access to deprecated functions can be re-enabled by running config with
4514 "enable-deprecated". In addition applications wishing to use deprecated
4515 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4516 will, by default, disable some transitive includes that previously existed
4517 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4521 * Added support for OCB mode. OpenSSL has been granted a patent license
4522 compatible with the OpenSSL license for use of OCB. Details are available
4523 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4524 for OCB can be removed by calling config with no-ocb.
4528 * SSLv2 support has been removed. It still supports receiving an SSLv2
4529 compatible client hello.
4533 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4534 done while fixing the error code for the key-too-small case.
4536 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4538 * CA.sh has been removed; use CA.pl instead.
4542 * Removed old DES API.
4546 * Remove various unsupported platforms:
4552 Sinix/ReliantUNIX RM400
4557 16-bit platforms such as WIN16
4561 * Clean up OPENSSL_NO_xxx #define's
4562 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4563 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4564 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4565 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4566 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4567 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4568 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4569 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4570 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4571 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4575 * Cleaned up dead code
4576 Remove all but one '#ifdef undef' which is to be looked at.
4580 * Clean up calling of xxx_free routines.
4581 Just like free(), fix most of the xxx_free routines to accept
4582 NULL. Remove the non-null checks from callers. Save much code.
4586 * Add secure heap for storage of private keys (when possible).
4587 Add BIO_s_secmem(), CBIGNUM, etc.
4588 Contributed by Akamai Technologies under our Corporate CLA.
4592 * Experimental support for a new, fast, unbiased prime candidate generator,
4593 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4595 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4597 * New output format NSS in the sess_id command line tool. This allows
4598 exporting the session id and the master key in NSS keylog format.
4600 *Martin Kaiser <martin@kaiser.cx>*
4602 * Harmonize version and its documentation. -f flag is used to display
4605 *mancha <mancha1@zoho.com>*
4607 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4608 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4610 *mancha <mancha1@zoho.com>*
4612 * Fix some double frees. These are not thought to be exploitable.
4614 *mancha <mancha1@zoho.com>*
4616 * A missing bounds check in the handling of the TLS heartbeat extension
4617 can be used to reveal up to 64k of memory to a connected client or
4620 Thanks for Neel Mehta of Google Security for discovering this bug and to
4621 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4622 preparing the fix ([CVE-2014-0160])
4624 *Adam Langley, Bodo Moeller*
4626 * Fix for the attack described in the paper "Recovering OpenSSL
4627 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4628 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4629 <http://eprint.iacr.org/2014/140>
4631 Thanks to Yuval Yarom and Naomi Benger for discovering this
4632 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4634 *Yuval Yarom and Naomi Benger*
4636 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4637 this fixes a limitation in previous versions of OpenSSL.
4641 * Experimental encrypt-then-mac support.
4643 Experimental support for encrypt then mac from
4644 draft-gutmann-tls-encrypt-then-mac-02.txt
4646 To enable it set the appropriate extension number (0x42 for the test
4647 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4649 For non-compliant peers (i.e. just about everything) this should have no
4652 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4656 * Add EVP support for key wrapping algorithms, to avoid problems with
4657 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4658 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4659 algorithms and include tests cases.
4663 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4668 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4669 MGF1 digest and OAEP label.
4673 * Make openssl verify return errors.
4675 *Chris Palmer <palmer@google.com> and Ben Laurie*
4677 * New function ASN1_TIME_diff to calculate the difference between two
4678 ASN1_TIME structures or one structure and the current time.
4682 * Update fips_test_suite to support multiple command line options. New
4683 test to induce all self test errors in sequence and check expected
4688 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4689 sign or verify all in one operation.
4693 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4694 test programs and fips_test_suite. Includes functionality to parse
4695 the minimal script output of fipsalgest.pl directly.
4699 * Add authorisation parameter to FIPS_module_mode_set().
4703 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4707 * Use separate DRBG fields for internal and external flags. New function
4708 FIPS_drbg_health_check() to perform on demand health checking. Add
4709 generation tests to fips_test_suite with reduced health check interval to
4710 demonstrate periodic health checking. Add "nodh" option to
4711 fips_test_suite to skip very slow DH test.
4715 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4720 * More extensive health check for DRBG checking many more failure modes.
4721 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4722 combination: call this in fips_test_suite.
4726 * Add support for canonical generation of DSA parameter 'g'. See
4729 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4730 POST to handle HMAC cases.
4734 * Add functions FIPS_module_version() and FIPS_module_version_text()
4735 to return numerical and string versions of the FIPS module number.
4739 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4740 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4741 outside the validated module in the FIPS capable OpenSSL.
4745 * Minor change to DRBG entropy callback semantics. In some cases
4746 there is no multiple of the block length between min_len and
4747 max_len. Allow the callback to return more than max_len bytes
4748 of entropy but discard any extra: it is the callback's responsibility
4749 to ensure that the extra data discarded does not impact the
4750 requested amount of entropy.
4754 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4755 information in FIPS186-3, SP800-57 and SP800-131A.
4759 * CCM support via EVP. Interface is very similar to GCM case except we
4760 must supply all data in one chunk (i.e. no update, final) and the
4761 message length must be supplied if AAD is used. Add algorithm test
4766 * Initial version of POST overhaul. Add POST callback to allow the status
4767 of POST to be monitored and/or failures induced. Modify fips_test_suite
4768 to use callback. Always run all selftests even if one fails.
4772 * XTS support including algorithm test driver in the fips_gcmtest program.
4773 Note: this does increase the maximum key length from 32 to 64 bytes but
4774 there should be no binary compatibility issues as existing applications
4775 will never use XTS mode.
4779 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4780 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4781 performs algorithm blocking for unapproved PRNG types. Also do not
4782 set PRNG type in FIPS_mode_set(): leave this to the application.
4783 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4784 the standard OpenSSL PRNG: set additional data to a date time vector.
4788 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4789 This shouldn't present any incompatibility problems because applications
4790 shouldn't be using these directly and any that are will need to rethink
4791 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4795 * Extensive self tests and health checking required by SP800-90 DRBG.
4796 Remove strength parameter from FIPS_drbg_instantiate and always
4797 instantiate at maximum supported strength.
4801 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4805 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4809 * New function DH_compute_key_padded() to compute a DH key and pad with
4810 leading zeroes if needed: this complies with SP800-56A et al.
4814 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4815 anything, incomplete, subject to change and largely untested at present.
4819 * Modify fipscanisteronly build option to only build the necessary object
4820 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4824 * Add experimental option FIPSSYMS to give all symbols in
4825 fipscanister.o and FIPS or fips prefix. This will avoid
4826 conflicts with future versions of OpenSSL. Add perl script
4827 util/fipsas.pl to preprocess assembly language source files
4828 and rename any affected symbols.
4832 * Add selftest checks and algorithm block of non-fips algorithms in
4833 FIPS mode. Remove DES2 from selftests.
4837 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4838 return internal method without any ENGINE dependencies. Add new
4839 tiny fips sign and verify functions.
4843 * New build option no-ec2m to disable characteristic 2 code.
4847 * New build option "fipscanisteronly". This only builds fipscanister.o
4848 and (currently) associated fips utilities. Uses the file Makefile.fips
4849 instead of Makefile.org as the prototype.
4853 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4854 Update fips_gcmtest to use IV generator.
4858 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4859 setting output buffer to NULL. The `*Final` function must be
4860 called although it will not retrieve any additional data. The tag
4861 can be set or retrieved with a ctrl. The IV length is by default 12
4862 bytes (96 bits) but can be set to an alternative value. If the IV
4863 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4868 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4869 underlying do_cipher function handles all cipher semantics itself
4870 including padding and finalisation. This is useful if (for example)
4871 an ENGINE cipher handles block padding itself. The behaviour of
4872 do_cipher is subtly changed if this flag is set: the return value
4873 is the number of characters written to the output buffer (zero is
4874 no longer an error code) or a negative error code. Also if the
4875 input buffer is NULL and length 0 finalisation should be performed.
4879 * If a candidate issuer certificate is already part of the constructed
4880 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4884 * Improve forward-security support: add functions
4886 void SSL_CTX_set_not_resumable_session_callback(
4887 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4888 void SSL_set_not_resumable_session_callback(
4889 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4891 for use by SSL/TLS servers; the callback function will be called whenever a
4892 new session is created, and gets to decide whether the session may be
4893 cached to make it resumable (return 0) or not (return 1). (As by the
4894 SSL/TLS protocol specifications, the session_id sent by the server will be
4895 empty to indicate that the session is not resumable; also, the server will
4896 not generate RFC 4507 (RFC 5077) session tickets.)
4898 A simple reasonable callback implementation is to return is_forward_secure.
4899 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4900 by the SSL/TLS server library, indicating whether it can provide forward
4903 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4905 * New -verify_name option in command line utilities to set verification
4910 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4911 Add CMAC pkey methods.
4915 * Experimental renegotiation in s_server -www mode. If the client
4916 browses /reneg connection is renegotiated. If /renegcert it is
4917 renegotiated requesting a certificate.
4921 * Add an "external" session cache for debugging purposes to s_server. This
4922 should help trace issues which normally are only apparent in deployed
4923 multi-process servers.
4927 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4928 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4929 BIO_set_cipher() and some obscure PEM functions were changed so they
4930 can now return an error. The RAND changes required a change to the
4931 RAND_METHOD structure.
4935 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4936 a gcc attribute to warn if the result of a function is ignored. This
4937 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4938 whose return value is often ignored.
4942 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4943 These allow SCTs (signed certificate timestamps) to be requested and
4944 validated when establishing a connection.
4946 *Rob Percival <robpercival@google.com>*
4951 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4953 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4954 used even when parsing explicit parameters, when loading a encoded key
4955 or calling `EC_GROUP_new_from_ecpkparameters()`/
4956 `EC_GROUP_new_from_ecparameters()`.
4957 This prevents bypass of security hardening and performance gains,
4958 especially for curves with specialized EC_METHODs.
4959 By default, if a key encoded with explicit parameters is loaded and later
4960 encoded, the output is still encoded with explicit parameters, even if
4961 internally a "named" EC_GROUP is used for computation.
4965 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4966 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4967 NULL. After this change, only the cofactor parameter can be NULL. It also
4968 does some minimal sanity checks on the passed order.
4973 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4974 An attack is simple, if the first CMS_recipientInfo is valid but the
4975 second CMS_recipientInfo is chosen ciphertext. If the second
4976 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4977 encryption key will be replaced by garbage, and the message cannot be
4978 decoded, but if the RSA decryption fails, the correct encryption key is
4979 used and the recipient will not notice the attack.
4980 As a work around for this potential attack the length of the decrypted
4981 key must be equal to the cipher default key length, in case the
4982 certificate is not given and all recipientInfo are tried out.
4983 The old behaviour can be re-enabled in the CMS code by setting the
4984 CMS_DEBUG_DECRYPT flag.
4989 * Document issue with installation paths in diverse Windows builds
4991 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4992 binaries and run-time config file.
4997 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4999 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5000 This changes the size when using the `genpkey` command when no size is given.
5001 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5002 generation commands to use 2048 bits by default.
5006 * Add FIPS support for Android Arm 64-bit
5008 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5009 Module in Version 2.0.10. For some reason, the corresponding target
5010 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5011 built with FIPS support on Android Arm 64-bit. This omission has been
5014 *Matthias St. Pierre*
5016 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5018 * 0-byte record padding oracle
5020 If an application encounters a fatal protocol error and then calls
5021 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5022 then OpenSSL can respond differently to the calling application if a 0 byte
5023 record is received with invalid padding compared to if a 0 byte record is
5024 received with an invalid MAC. If the application then behaves differently
5025 based on that in a way that is detectable to the remote peer, then this
5026 amounts to a padding oracle that could be used to decrypt data.
5028 In order for this to be exploitable "non-stitched" ciphersuites must be in
5029 use. Stitched ciphersuites are optimised implementations of certain
5030 commonly used ciphersuites. Also the application must call SSL_shutdown()
5031 twice even if a protocol error has occurred (applications should not do
5032 this but some do anyway).
5034 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5035 Aviram, with additional investigation by Steven Collison and Andrew
5036 Hourselt. It was reported to OpenSSL on 10th December 2018.
5041 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5045 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5047 * Microarchitecture timing vulnerability in ECC scalar multiplication
5049 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5050 shown to be vulnerable to a microarchitecture timing side channel attack.
5051 An attacker with sufficient access to mount local timing attacks during
5052 ECDSA signature generation could recover the private key.
5054 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5055 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5061 * Timing vulnerability in DSA signature generation
5063 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5064 timing side channel attack. An attacker could use variations in the signing
5065 algorithm to recover the private key.
5067 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5072 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5073 Module, accidentally introduced while backporting security fixes from the
5074 development branch and hindering the use of ECC in FIPS mode.
5078 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5080 * Client DoS due to large DH parameter
5082 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5083 malicious server can send a very large prime value to the client. This will
5084 cause the client to spend an unreasonably long period of time generating a
5085 key for this prime resulting in a hang until the client has finished. This
5086 could be exploited in a Denial Of Service attack.
5088 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5093 * Cache timing vulnerability in RSA Key Generation
5095 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5096 a cache timing side channel attack. An attacker with sufficient access to
5097 mount cache timing attacks during the RSA key generation process could
5098 recover the private key.
5100 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5101 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5106 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5107 parameter is no longer accepted, as it leads to a corrupt table. NULL
5108 pem_str is reserved for alias entries only.
5112 * Revert blinding in ECDSA sign and instead make problematic addition
5113 length-invariant. Switch even to fixed-length Montgomery multiplication.
5117 * Change generating and checking of primes so that the error rate of not
5118 being prime depends on the intended use based on the size of the input.
5119 For larger primes this will result in more rounds of Miller-Rabin.
5120 The maximal error rate for primes with more than 1080 bits is lowered
5123 *Kurt Roeckx, Annie Yousar*
5125 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5129 * Add blinding to ECDSA and DSA signatures to protect against side channel
5130 attacks discovered by Keegan Ryan (NCC Group).
5134 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5135 now allow empty (zero character) pass phrases.
5139 * Certificate time validation (X509_cmp_time) enforces stricter
5140 compliance with RFC 5280. Fractional seconds and timezone offsets
5141 are no longer allowed.
5145 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5147 * Constructed ASN.1 types with a recursive definition could exceed the stack
5149 Constructed ASN.1 types with a recursive definition (such as can be found
5150 in PKCS7) could eventually exceed the stack given malicious input with
5151 excessive recursion. This could result in a Denial Of Service attack. There
5152 are no such structures used within SSL/TLS that come from untrusted sources
5153 so this is considered safe.
5155 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5161 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5163 * Read/write after SSL object in error state
5165 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5166 mechanism. The intent was that if a fatal error occurred during a handshake
5167 then OpenSSL would move into the error state and would immediately fail if
5168 you attempted to continue the handshake. This works as designed for the
5169 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5170 SSL_connect()), however due to a bug it does not work correctly if
5171 SSL_read() or SSL_write() is called directly. In that scenario, if the
5172 handshake fails then a fatal error will be returned in the initial function
5173 call. If SSL_read()/SSL_write() is subsequently called by the application
5174 for the same SSL object then it will succeed and the data is passed without
5175 being decrypted/encrypted directly from the SSL/TLS record layer.
5177 In order to exploit this issue an application bug would have to be present
5178 that resulted in a call to SSL_read()/SSL_write() being issued after having
5179 already received a fatal error.
5181 This issue was reported to OpenSSL by David Benjamin (Google).
5186 * rsaz_1024_mul_avx2 overflow bug on x86_64
5188 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5189 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5190 Analysis suggests that attacks against RSA and DSA as a result of this
5191 defect would be very difficult to perform and are not believed likely.
5192 Attacks against DH1024 are considered just feasible, because most of the
5193 work necessary to deduce information about a private key may be performed
5194 offline. The amount of resources required for such an attack would be
5195 significant. However, for an attack on TLS to be meaningful, the server
5196 would have to share the DH1024 private key among multiple clients, which is
5197 no longer an option since CVE-2016-0701.
5199 This only affects processors that support the AVX2 but not ADX extensions
5200 like Intel Haswell (4th generation).
5202 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5203 was originally found via the OSS-Fuzz project.
5208 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5210 * bn_sqrx8x_internal carry bug on x86_64
5212 There is a carry propagating bug in the x86_64 Montgomery squaring
5213 procedure. No EC algorithms are affected. Analysis suggests that attacks
5214 against RSA and DSA as a result of this defect would be very difficult to
5215 perform and are not believed likely. Attacks against DH are considered just
5216 feasible (although very difficult) because most of the work necessary to
5217 deduce information about a private key may be performed offline. The amount
5218 of resources required for such an attack would be very significant and
5219 likely only accessible to a limited number of attackers. An attacker would
5220 additionally need online access to an unpatched system using the target
5221 private key in a scenario with persistent DH parameters and a private
5222 key that is shared between multiple clients.
5224 This only affects processors that support the BMI1, BMI2 and ADX extensions
5225 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5227 This issue was reported to OpenSSL by the OSS-Fuzz project.
5232 * Malformed X.509 IPAddressFamily could cause OOB read
5234 If an X.509 certificate has a malformed IPAddressFamily extension,
5235 OpenSSL could do a one-byte buffer overread. The most likely result
5236 would be an erroneous display of the certificate in text format.
5238 This issue was reported to OpenSSL by the OSS-Fuzz project.
5242 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5244 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5245 platform rather than 'mingw'.
5249 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5251 * Truncated packet could crash via OOB read
5253 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5254 cipher is being used, then a truncated packet can cause that host to
5255 perform an out-of-bounds read, usually resulting in a crash.
5257 This issue was reported to OpenSSL by Robert Święcki of Google.
5262 * BN_mod_exp may produce incorrect results on x86_64
5264 There is a carry propagating bug in the x86_64 Montgomery squaring
5265 procedure. No EC algorithms are affected. Analysis suggests that attacks
5266 against RSA and DSA as a result of this defect would be very difficult to
5267 perform and are not believed likely. Attacks against DH are considered just
5268 feasible (although very difficult) because most of the work necessary to
5269 deduce information about a private key may be performed offline. The amount
5270 of resources required for such an attack would be very significant and
5271 likely only accessible to a limited number of attackers. An attacker would
5272 additionally need online access to an unpatched system using the target
5273 private key in a scenario with persistent DH parameters and a private
5274 key that is shared between multiple clients. For example this can occur by
5275 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5276 similar to CVE-2015-3193 but must be treated as a separate problem.
5278 This issue was reported to OpenSSL by the OSS-Fuzz project.
5283 * Montgomery multiplication may produce incorrect results
5285 There is a carry propagating bug in the Broadwell-specific Montgomery
5286 multiplication procedure that handles input lengths divisible by, but
5287 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5288 and DH private keys are impossible. This is because the subroutine in
5289 question is not used in operations with the private key itself and an input
5290 of the attacker's direct choice. Otherwise the bug can manifest itself as
5291 transient authentication and key negotiation failures or reproducible
5292 erroneous outcome of public-key operations with specially crafted input.
5293 Among EC algorithms only Brainpool P-512 curves are affected and one
5294 presumably can attack ECDH key negotiation. Impact was not analyzed in
5295 detail, because pre-requisites for attack are considered unlikely. Namely
5296 multiple clients have to choose the curve in question and the server has to
5297 share the private key among them, neither of which is default behaviour.
5298 Even then only clients that chose the curve will be affected.
5300 This issue was publicly reported as transient failures and was not
5301 initially recognized as a security issue. Thanks to Richard Morgan for
5302 providing reproducible case.
5307 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5308 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5309 prevent issues where no progress is being made and the peer continually
5310 sends unrecognised record types, using up resources processing them.
5314 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5316 * Missing CRL sanity check
5318 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5319 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5320 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5322 This issue only affects the OpenSSL 1.0.2i
5327 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5329 * OCSP Status Request extension unbounded memory growth
5331 A malicious client can send an excessively large OCSP Status Request
5332 extension. If that client continually requests renegotiation, sending a
5333 large OCSP Status Request extension each time, then there will be unbounded
5334 memory growth on the server. This will eventually lead to a Denial Of
5335 Service attack through memory exhaustion. Servers with a default
5336 configuration are vulnerable even if they do not support OCSP. Builds using
5337 the "no-ocsp" build time option are not affected.
5339 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5344 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5347 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5353 * OOB write in MDC2_Update()
5355 An overflow can occur in MDC2_Update() either if called directly or
5356 through the EVP_DigestUpdate() function using MDC2. If an attacker
5357 is able to supply very large amounts of input data after a previous
5358 call to EVP_EncryptUpdate() with a partial block then a length check
5359 can overflow resulting in a heap corruption.
5361 The amount of data needed is comparable to SIZE_MAX which is impractical
5364 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5369 * Malformed SHA512 ticket DoS
5371 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5372 DoS attack where a malformed ticket will result in an OOB read which will
5375 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5376 a custom server callback and ticket lookup mechanism.
5378 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5383 * OOB write in BN_bn2dec()
5385 The function BN_bn2dec() does not check the return value of BN_div_word().
5386 This can cause an OOB write if an application uses this function with an
5387 overly large BIGNUM. This could be a problem if an overly large certificate
5388 or CRL is printed out from an untrusted source. TLS is not affected because
5389 record limits will reject an oversized certificate before it is parsed.
5391 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5396 * OOB read in TS_OBJ_print_bio()
5398 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5399 the total length the OID text representation would use and not the amount
5400 of data written. This will result in OOB reads when large OIDs are
5403 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5408 * Pointer arithmetic undefined behaviour
5410 Avoid some undefined pointer arithmetic
5412 A common idiom in the codebase is to check limits in the following manner:
5415 Where "p" points to some malloc'd data of SIZE bytes and
5418 "len" here could be from some externally supplied data (e.g. from a TLS
5421 The rules of C pointer arithmetic are such that "p + len" is only well
5422 defined where len <= SIZE. Therefore the above idiom is actually
5423 undefined behaviour.
5425 For example this could cause problems if some malloc implementation
5426 provides an address for "p" such that "p + len" actually overflows for
5427 values of len that are too big and therefore p + len < limit.
5429 This issue was reported to OpenSSL by Guido Vranken
5434 * Constant time flag not preserved in DSA signing
5436 Operations in the DSA signing algorithm should run in constant time in
5437 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5438 implementation means that a non-constant time codepath is followed for
5439 certain operations. This has been demonstrated through a cache-timing
5440 attack to be sufficient for an attacker to recover the private DSA key.
5442 This issue was reported by César Pereida (Aalto University), Billy Brumley
5443 (Tampere University of Technology), and Yuval Yarom (The University of
5444 Adelaide and NICTA).
5449 * DTLS buffered message DoS
5451 In a DTLS connection where handshake messages are delivered out-of-order
5452 those messages that OpenSSL is not yet ready to process will be buffered
5453 for later use. Under certain circumstances, a flaw in the logic means that
5454 those messages do not get removed from the buffer even though the handshake
5455 has been completed. An attacker could force up to approx. 15 messages to
5456 remain in the buffer when they are no longer required. These messages will
5457 be cleared when the DTLS connection is closed. The default maximum size for
5458 a message is 100k. Therefore, the attacker could force an additional 1500k
5459 to be consumed per connection. By opening many simultaneous connections an
5460 attacker could cause a DoS attack through memory exhaustion.
5462 This issue was reported to OpenSSL by Quan Luo.
5467 * DTLS replay protection DoS
5469 A flaw in the DTLS replay attack protection mechanism means that records
5470 that arrive for future epochs update the replay protection "window" before
5471 the MAC for the record has been validated. This could be exploited by an
5472 attacker by sending a record for the next epoch (which does not have to
5473 decrypt or have a valid MAC), with a very large sequence number. This means
5474 that all subsequent legitimate packets are dropped causing a denial of
5475 service for a specific DTLS connection.
5477 This issue was reported to OpenSSL by the OCAP audit team.
5482 * Certificate message OOB reads
5484 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5485 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5486 theoretical DoS risk but this has not been observed in practice on common
5489 The messages affected are client certificate, client certificate request
5490 and server certificate. As a result the attack can only be performed
5491 against a client or a server which enables client authentication.
5493 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5498 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5500 * Prevent padding oracle in AES-NI CBC MAC check
5502 A MITM attacker can use a padding oracle attack to decrypt traffic
5503 when the connection uses an AES CBC cipher and the server support
5506 This issue was introduced as part of the fix for Lucky 13 padding
5507 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5508 constant time by making sure that always the same bytes are read and
5509 compared against either the MAC or padding bytes. But it no longer
5510 checked that there was enough data to have both the MAC and padding
5513 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5517 * Fix EVP_EncodeUpdate overflow
5519 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5520 Base64 encoding of binary data. If an attacker is able to supply very large
5521 amounts of input data then a length check can overflow resulting in a heap
5524 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5525 the `PEM_write_bio*` family of functions. These are mainly used within the
5526 OpenSSL command line applications, so any application which processes data
5527 from an untrusted source and outputs it as a PEM file should be considered
5528 vulnerable to this issue. User applications that call these APIs directly
5529 with large amounts of untrusted data may also be vulnerable.
5531 This issue was reported by Guido Vranken.
5536 * Fix EVP_EncryptUpdate overflow
5538 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5539 is able to supply very large amounts of input data after a previous call to
5540 EVP_EncryptUpdate() with a partial block then a length check can overflow
5541 resulting in a heap corruption. Following an analysis of all OpenSSL
5542 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5543 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5544 the first called function after an EVP_EncryptInit(), and therefore that
5545 specific call must be safe. The second form is where the length passed to
5546 EVP_EncryptUpdate() can be seen from the code to be some small value and
5547 therefore there is no possibility of an overflow. Since all instances are
5548 one of these two forms, it is believed that there can be no overflows in
5549 internal code due to this problem. It should be noted that
5550 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5551 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5552 of these calls have also been analysed too and it is believed there are no
5553 instances in internal usage where an overflow could occur.
5555 This issue was reported by Guido Vranken.
5560 * Prevent ASN.1 BIO excessive memory allocation
5562 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5563 a short invalid encoding can cause allocation of large amounts of memory
5564 potentially consuming excessive resources or exhausting memory.
5566 Any application parsing untrusted data through d2i BIO functions is
5567 affected. The memory based functions such as d2i_X509() are *not* affected.
5568 Since the memory based functions are used by the TLS library, TLS
5569 applications are not affected.
5571 This issue was reported by Brian Carpenter.
5578 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5579 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5580 in arbitrary stack data being returned in the buffer.
5582 This issue was reported by Guido Vranken.
5587 * Modify behavior of ALPN to invoke callback after SNI/servername
5588 callback, such that updates to the SSL_CTX affect ALPN.
5592 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5597 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5598 methods are enabled and ssl2 is disabled the methods return NULL.
5602 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5604 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5605 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5606 provide any "EXPORT" or "LOW" strength ciphers.
5610 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5611 is by default disabled at build-time. Builds that are not configured with
5612 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5613 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5614 will need to explicitly call either of:
5616 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5618 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5620 as appropriate. Even if either of those is used, or the application
5621 explicitly uses the version-specific SSLv2_method() or its client and
5622 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5623 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5624 ciphers, and SSLv2 56-bit DES are no longer available.
5629 * Fix a double-free in DSA code
5631 A double free bug was discovered when OpenSSL parses malformed DSA private
5632 keys and could lead to a DoS attack or memory corruption for applications
5633 that receive DSA private keys from untrusted sources. This scenario is
5636 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5642 * Disable SRP fake user seed to address a server memory leak.
5644 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5646 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5647 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5648 was changed to ignore the "fake user" SRP seed, even if the seed
5651 Users should use SRP_VBASE_get1_by_user instead. Note that in
5652 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5653 also that even though configuring the SRP seed attempts to hide
5654 invalid usernames by continuing the handshake with fake
5655 credentials, this behaviour is not constant time and no strong
5656 guarantees are made that the handshake is indistinguishable from
5657 that of a valid user.
5662 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5664 In the BN_hex2bn function the number of hex digits is calculated using an
5665 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5666 large values of `i` this can result in `bn_expand` not allocating any
5667 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5668 field as NULL leading to a subsequent NULL ptr deref. For very large values
5669 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5670 In this case memory is allocated to the internal BIGNUM data field, but it
5671 is insufficiently sized leading to heap corruption. A similar issue exists
5672 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5673 is ever called by user applications with very large untrusted hex/dec data.
5674 This is anticipated to be a rare occurrence.
5676 All OpenSSL internal usage of these functions use data that is not expected
5677 to be untrusted, e.g. config file data or application command line
5678 arguments. If user developed applications generate config file data based
5679 on untrusted data then it is possible that this could also lead to security
5680 consequences. This is also anticipated to be rare.
5682 This issue was reported to OpenSSL by Guido Vranken.
5687 * Fix memory issues in `BIO_*printf` functions
5689 The internal `fmtstr` function used in processing a "%s" format string in
5690 the `BIO_*printf` functions could overflow while calculating the length of a
5691 string and cause an OOB read when printing very long strings.
5693 Additionally the internal `doapr_outch` function can attempt to write to an
5694 OOB memory location (at an offset from the NULL pointer) in the event of a
5695 memory allocation failure. In 1.0.2 and below this could be caused where
5696 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5697 could be in processing a very long "%s" format string. Memory leaks can
5700 The first issue may mask the second issue dependent on compiler behaviour.
5701 These problems could enable attacks where large amounts of untrusted data
5702 is passed to the `BIO_*printf` functions. If applications use these functions
5703 in this way then they could be vulnerable. OpenSSL itself uses these
5704 functions when printing out human-readable dumps of ASN.1 data. Therefore
5705 applications that print this data could be vulnerable if the data is from
5706 untrusted sources. OpenSSL command line applications could also be
5707 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5708 as command line arguments.
5710 Libssl is not considered directly vulnerable. Additionally certificates etc
5711 received via remote connections via libssl are also unlikely to be able to
5712 trigger these issues because of message size limits enforced within libssl.
5714 This issue was reported to OpenSSL Guido Vranken.
5719 * Side channel attack on modular exponentiation
5721 A side-channel attack was found which makes use of cache-bank conflicts on
5722 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5723 of RSA keys. The ability to exploit this issue is limited as it relies on
5724 an attacker who has control of code in a thread running on the same
5725 hyper-threaded core as the victim thread which is performing decryptions.
5727 This issue was reported to OpenSSL by Yuval Yarom, The University of
5728 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5729 Nadia Heninger, University of Pennsylvania with more information at
5730 <http://cachebleed.info>.
5735 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5736 if no keysize is specified with default_bits. This fixes an
5737 omission in an earlier change that changed all RSA/DSA key generation
5738 commands to use 2048 bits by default.
5742 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5744 * DH small subgroups
5746 Historically OpenSSL only ever generated DH parameters based on "safe"
5747 primes. More recently (in version 1.0.2) support was provided for
5748 generating X9.42 style parameter files such as those required for RFC 5114
5749 support. The primes used in such files may not be "safe". Where an
5750 application is using DH configured with parameters based on primes that are
5751 not "safe" then an attacker could use this fact to find a peer's private
5752 DH exponent. This attack requires that the attacker complete multiple
5753 handshakes in which the peer uses the same private DH exponent. For example
5754 this could be used to discover a TLS server's private DH exponent if it's
5755 reusing the private DH exponent or it's using a static DH ciphersuite.
5757 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5758 TLS. It is not on by default. If the option is not set then the server
5759 reuses the same private DH exponent for the life of the server process and
5760 would be vulnerable to this attack. It is believed that many popular
5761 applications do set this option and would therefore not be at risk.
5763 The fix for this issue adds an additional check where a "q" parameter is
5764 available (as is the case in X9.42 based parameters). This detects the
5765 only known attack, and is the only possible defense for static DH
5766 ciphersuites. This could have some performance impact.
5768 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5769 default and cannot be disabled. This could have some performance impact.
5771 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5776 * SSLv2 doesn't block disabled ciphers
5778 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5779 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5780 been disabled, provided that the SSLv2 protocol was not also disabled via
5783 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5784 and Sebastian Schinzel.
5789 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5791 * BN_mod_exp may produce incorrect results on x86_64
5793 There is a carry propagating bug in the x86_64 Montgomery squaring
5794 procedure. No EC algorithms are affected. Analysis suggests that attacks
5795 against RSA and DSA as a result of this defect would be very difficult to
5796 perform and are not believed likely. Attacks against DH are considered just
5797 feasible (although very difficult) because most of the work necessary to
5798 deduce information about a private key may be performed offline. The amount
5799 of resources required for such an attack would be very significant and
5800 likely only accessible to a limited number of attackers. An attacker would
5801 additionally need online access to an unpatched system using the target
5802 private key in a scenario with persistent DH parameters and a private
5803 key that is shared between multiple clients. For example this can occur by
5804 default in OpenSSL DHE based SSL/TLS ciphersuites.
5806 This issue was reported to OpenSSL by Hanno Böck.
5811 * Certificate verify crash with missing PSS parameter
5813 The signature verification routines will crash with a NULL pointer
5814 dereference if presented with an ASN.1 signature using the RSA PSS
5815 algorithm and absent mask generation function parameter. Since these
5816 routines are used to verify certificate signature algorithms this can be
5817 used to crash any certificate verification operation and exploited in a
5818 DoS attack. Any application which performs certificate verification is
5819 vulnerable including OpenSSL clients and servers which enable client
5822 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5827 * X509_ATTRIBUTE memory leak
5829 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5830 memory. This structure is used by the PKCS#7 and CMS routines so any
5831 application which reads PKCS#7 or CMS data from untrusted sources is
5832 affected. SSL/TLS is not affected.
5834 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5840 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5841 This changes the decoding behaviour for some invalid messages,
5842 though the change is mostly in the more lenient direction, and
5843 legacy behaviour is preserved as much as possible.
5847 * In DSA_generate_parameters_ex, if the provided seed is too short,
5850 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5852 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5854 * Alternate chains certificate forgery
5856 During certificate verification, OpenSSL will attempt to find an
5857 alternative certificate chain if the first attempt to build such a chain
5858 fails. An error in the implementation of this logic can mean that an
5859 attacker could cause certain checks on untrusted certificates to be
5860 bypassed, such as the CA flag, enabling them to use a valid leaf
5861 certificate to act as a CA and "issue" an invalid certificate.
5863 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5868 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5870 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5871 incompatibility in the handling of HMAC. The previous ABI has now been
5876 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5878 * Malformed ECParameters causes infinite loop
5880 When processing an ECParameters structure OpenSSL enters an infinite loop
5881 if the curve specified is over a specially malformed binary polynomial
5884 This can be used to perform denial of service against any
5885 system which processes public keys, certificate requests or
5886 certificates. This includes TLS clients and TLS servers with
5887 client authentication enabled.
5889 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5894 * Exploitable out-of-bounds read in X509_cmp_time
5896 X509_cmp_time does not properly check the length of the ASN1_TIME
5897 string and can read a few bytes out of bounds. In addition,
5898 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5901 An attacker can use this to craft malformed certificates and CRLs of
5902 various sizes and potentially cause a segmentation fault, resulting in
5903 a DoS on applications that verify certificates or CRLs. TLS clients
5904 that verify CRLs are affected. TLS clients and servers with client
5905 authentication enabled may be affected if they use custom verification
5908 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5909 independently by Hanno Böck.
5914 * PKCS7 crash with missing EnvelopedContent
5916 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5917 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5918 with missing content and trigger a NULL pointer dereference on parsing.
5920 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5921 structures from untrusted sources are affected. OpenSSL clients and
5922 servers are not affected.
5924 This issue was reported to OpenSSL by Michal Zalewski (Google).
5929 * CMS verify infinite loop with unknown hash function
5931 When verifying a signedData message the CMS code can enter an infinite loop
5932 if presented with an unknown hash function OID. This can be used to perform
5933 denial of service against any system which verifies signedData messages using
5935 This issue was reported to OpenSSL by Johannes Bauer.
5940 * Race condition handling NewSessionTicket
5942 If a NewSessionTicket is received by a multi-threaded client when attempting to
5943 reuse a previous ticket then a race condition can occur potentially leading to
5944 a double free of the ticket data.
5949 * Only support 256-bit or stronger elliptic curves with the
5950 'ecdh_auto' setting (server) or by default (client). Of supported
5951 curves, prefer P-256 (both).
5955 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5957 * ClientHello sigalgs DoS fix
5959 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5960 invalid signature algorithms extension a NULL pointer dereference will
5961 occur. This can be exploited in a DoS attack against the server.
5963 This issue was was reported to OpenSSL by David Ramos of Stanford
5967 *Stephen Henson and Matt Caswell*
5969 * Multiblock corrupted pointer fix
5971 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5972 feature only applies on 64 bit x86 architecture platforms that support AES
5973 NI instructions. A defect in the implementation of "multiblock" can cause
5974 OpenSSL's internal write buffer to become incorrectly set to NULL when
5975 using non-blocking IO. Typically, when the user application is using a
5976 socket BIO for writing, this will only result in a failed connection.
5977 However if some other BIO is used then it is likely that a segmentation
5978 fault will be triggered, thus enabling a potential DoS attack.
5980 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5985 * Segmentation fault in DTLSv1_listen fix
5987 The DTLSv1_listen function is intended to be stateless and processes the
5988 initial ClientHello from many peers. It is common for user code to loop
5989 over the call to DTLSv1_listen until a valid ClientHello is received with
5990 an associated cookie. A defect in the implementation of DTLSv1_listen means
5991 that state is preserved in the SSL object from one invocation to the next
5992 that can lead to a segmentation fault. Errors processing the initial
5993 ClientHello can trigger this scenario. An example of such an error could be
5994 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5997 This issue was reported to OpenSSL by Per Allansson.
6002 * Segmentation fault in ASN1_TYPE_cmp fix
6004 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6005 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6006 certificate signature algorithm consistency this can be used to crash any
6007 certificate verification operation and exploited in a DoS attack. Any
6008 application which performs certificate verification is vulnerable including
6009 OpenSSL clients and servers which enable client authentication.
6014 * Segmentation fault for invalid PSS parameters fix
6016 The signature verification routines will crash with a NULL pointer
6017 dereference if presented with an ASN.1 signature using the RSA PSS
6018 algorithm and invalid parameters. Since these routines are used to verify
6019 certificate signature algorithms this can be used to crash any
6020 certificate verification operation and exploited in a DoS attack. Any
6021 application which performs certificate verification is vulnerable including
6022 OpenSSL clients and servers which enable client authentication.
6024 This issue was was reported to OpenSSL by Brian Carpenter.
6029 * ASN.1 structure reuse memory corruption fix
6031 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6032 memory corruption via an invalid write. Such reuse is and has been
6033 strongly discouraged and is believed to be rare.
6035 Applications that parse structures containing CHOICE or ANY DEFINED BY
6036 components may be affected. Certificate parsing (d2i_X509 and related
6037 functions) are however not affected. OpenSSL clients and servers are
6043 * PKCS7 NULL pointer dereferences fix
6045 The PKCS#7 parsing code does not handle missing outer ContentInfo
6046 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6047 missing content and trigger a NULL pointer dereference on parsing.
6049 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6050 otherwise parse PKCS#7 structures from untrusted sources are
6051 affected. OpenSSL clients and servers are not affected.
6053 This issue was reported to OpenSSL by Michal Zalewski (Google).
6058 * DoS via reachable assert in SSLv2 servers fix
6060 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6061 servers that both support SSLv2 and enable export cipher suites by sending
6062 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6064 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6065 (OpenSSL development team).
6070 * Empty CKE with client auth and DHE fix
6072 If client auth is used then a server can seg fault in the event of a DHE
6073 ciphersuite being selected and a zero length ClientKeyExchange message
6074 being sent by the client. This could be exploited in a DoS attack.
6079 * Handshake with unseeded PRNG fix
6081 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6082 with an unseeded PRNG. The conditions are:
6083 - The client is on a platform where the PRNG has not been seeded
6084 automatically, and the user has not seeded manually
6085 - A protocol specific client method version has been used (i.e. not
6086 SSL_client_methodv23)
6087 - A ciphersuite is used that does not require additional random data from
6088 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6090 If the handshake succeeds then the client random that has been used will
6091 have been generated from a PRNG with insufficient entropy and therefore the
6092 output may be predictable.
6094 For example using the following command with an unseeded openssl will
6095 succeed on an unpatched platform:
6097 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6102 * Use After Free following d2i_ECPrivatekey error fix
6104 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6105 could cause a use after free condition. This, in turn, could cause a double
6106 free in several private key parsing functions (such as d2i_PrivateKey
6107 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6108 for applications that receive EC private keys from untrusted
6109 sources. This scenario is considered rare.
6111 This issue was discovered by the BoringSSL project and fixed in their
6117 * X509_to_X509_REQ NULL pointer deref fix
6119 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6120 the certificate key is invalid. This function is rarely used in practice.
6122 This issue was discovered by Brian Carpenter.
6127 * Removed the export ciphers from the DEFAULT ciphers
6131 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6133 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6134 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6135 So far those who have to target multiple platforms would compromise
6136 and argue that binary targeting say ARMv5 would still execute on
6137 ARMv8. "Universal" build resolves this compromise by providing
6138 near-optimal performance even on newer platforms.
6142 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6143 (other platforms pending).
6145 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6147 * Add support for the SignedCertificateTimestampList certificate and
6148 OCSP response extensions from RFC6962.
6152 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6153 for corner cases. (Certain input points at infinity could lead to
6154 bogus results, with non-infinity inputs mapped to infinity too.)
6158 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6159 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6160 common cases are optimized and there still is room for further
6161 improvements. Vector Permutation AES for Altivec is also added.
6165 * Add support for little-endian ppc64 Linux target.
6167 *Marcelo Cerri (IBM)*
6169 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6170 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6171 are optimized and there still is room for further improvements.
6172 Both 32- and 64-bit modes are supported.
6174 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6176 * Improved ARMv7 NEON support.
6180 * Support for SPARC Architecture 2011 crypto extensions, first
6181 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6182 SHA256/512, MD5, GHASH and modular exponentiation.
6184 *Andy Polyakov, David Miller*
6186 * Accelerated modular exponentiation for Intel processors, a.k.a.
6189 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6191 * Support for new and upcoming Intel processors, including AVX2,
6192 BMI and SHA ISA extensions. This includes additional "stitched"
6193 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6196 This work was sponsored by Intel Corp.
6200 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6201 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6202 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6206 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6207 this fixes a limitation in previous versions of OpenSSL.
6211 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6212 MGF1 digest and OAEP label.
6216 * Add EVP support for key wrapping algorithms, to avoid problems with
6217 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6218 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6219 algorithms and include tests cases.
6223 * Add functions to allocate and set the fields of an ECDSA_METHOD
6226 *Douglas E. Engert, Steve Henson*
6228 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6229 difference in days and seconds between two tm or ASN1_TIME structures.
6233 * Add -rev test option to s_server to just reverse order of characters
6234 received by client and send back to server. Also prints an abbreviated
6235 summary of the connection parameters.
6239 * New option -brief for s_client and s_server to print out a brief summary
6240 of connection parameters.
6244 * Add callbacks for arbitrary TLS extensions.
6246 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6248 * New option -crl_download in several openssl utilities to download CRLs
6249 from CRLDP extension in certificates.
6253 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6257 * New function X509_CRL_diff to generate a delta CRL from the difference
6258 of two full CRLs. Add support to "crl" utility.
6262 * New functions to set lookup_crls function and to retrieve
6263 X509_STORE from X509_STORE_CTX.
6267 * Print out deprecated issuer and subject unique ID fields in
6272 * Extend OCSP I/O functions so they can be used for simple general purpose
6273 HTTP as well as OCSP. New wrapper function which can be used to download
6274 CRLs using the OCSP API.
6278 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6282 * `SSL_CONF*` functions. These provide a common framework for application
6283 configuration using configuration files or command lines.
6287 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6288 message callback and prints the results. Needs compile time option
6289 "enable-ssl-trace". New options to s_client and s_server to enable
6294 * New ctrl and macro to retrieve supported points extensions.
6295 Print out extension in s_server and s_client.
6299 * New functions to retrieve certificate signature and signature
6304 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6309 * New Suite B modes for TLS code. These use and enforce the requirements
6310 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6311 only use Suite B curves. The Suite B modes can be set by using the
6312 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6316 * New chain verification flags for Suite B levels of security. Check
6317 algorithms are acceptable when flags are set in X509_verify_cert.
6321 * Make tls1_check_chain return a set of flags indicating checks passed
6322 by a certificate chain. Add additional tests to handle client
6323 certificates: checks for matching certificate type and issuer name
6328 * If an attempt is made to use a signature algorithm not in the peer
6329 preference list abort the handshake. If client has no suitable
6330 signature algorithms in response to a certificate request do not
6331 use the certificate.
6335 * If server EC tmp key is not in client preference list abort handshake.
6339 * Add support for certificate stores in CERT structure. This makes it
6340 possible to have different stores per SSL structure or one store in
6341 the parent SSL_CTX. Include distinct stores for certificate chain
6342 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6343 to build and store a certificate chain in CERT structure: returning
6344 an error if the chain cannot be built: this will allow applications
6345 to test if a chain is correctly configured.
6347 Note: if the CERT based stores are not set then the parent SSL_CTX
6348 store is used to retain compatibility with existing behaviour.
6352 * New function ssl_set_client_disabled to set a ciphersuite disabled
6353 mask based on the current session, check mask when sending client
6354 hello and checking the requested ciphersuite.
6358 * New ctrls to retrieve and set certificate types in a certificate
6359 request message. Print out received values in s_client. If certificate
6360 types is not set with custom values set sensible values based on
6361 supported signature algorithms.
6365 * Support for distinct client and server supported signature algorithms.
6369 * Add certificate callback. If set this is called whenever a certificate
6370 is required by client or server. An application can decide which
6371 certificate chain to present based on arbitrary criteria: for example
6372 supported signature algorithms. Add very simple example to s_server.
6373 This fixes many of the problems and restrictions of the existing client
6374 certificate callback: for example you can now clear an existing
6375 certificate and specify the whole chain.
6379 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6380 the certificate can be used for (if anything). Set valid_flags field
6381 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6382 to have similar checks in it.
6384 Add new "cert_flags" field to CERT structure and include a "strict mode".
6385 This enforces some TLS certificate requirements (such as only permitting
6386 certificate signature algorithms contained in the supported algorithms
6387 extension) which some implementations ignore: this option should be used
6388 with caution as it could cause interoperability issues.
6392 * Update and tidy signature algorithm extension processing. Work out
6393 shared signature algorithms based on preferences and peer algorithms
6394 and print them out in s_client and s_server. Abort handshake if no
6395 shared signature algorithms.
6399 * Add new functions to allow customised supported signature algorithms
6400 for SSL and SSL_CTX structures. Add options to s_client and s_server
6405 * New function SSL_certs_clear() to delete all references to certificates
6406 from an SSL structure. Before this once a certificate had been added
6407 it couldn't be removed.
6411 * Integrate hostname, email address and IP address checking with certificate
6412 verification. New verify options supporting checking in openssl utility.
6416 * Fixes and wildcard matching support to hostname and email checking
6417 functions. Add manual page.
6419 *Florian Weimer (Red Hat Product Security Team)*
6421 * New functions to check a hostname email or IP address against a
6422 certificate. Add options x509 utility to print results of checks against
6427 * Fix OCSP checking.
6429 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6431 * Initial experimental support for explicitly trusted non-root CAs.
6432 OpenSSL still tries to build a complete chain to a root but if an
6433 intermediate CA has a trust setting included that is used. The first
6434 setting is used: whether to trust (e.g., -addtrust option to the x509
6439 * Add -trusted_first option which attempts to find certificates in the
6440 trusted store even if an untrusted chain is also supplied.
6444 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6445 platform support for Linux and Android.
6449 * Support for linux-x32, ILP32 environment in x86_64 framework.
6453 * Experimental multi-implementation support for FIPS capable OpenSSL.
6454 When in FIPS mode the approved implementations are used as normal,
6455 when not in FIPS mode the internal unapproved versions are used instead.
6456 This means that the FIPS capable OpenSSL isn't forced to use the
6457 (often lower performance) FIPS implementations outside FIPS mode.
6461 * Transparently support X9.42 DH parameters when calling
6462 PEM_read_bio_DHparameters. This means existing applications can handle
6463 the new parameter format automatically.
6467 * Initial experimental support for X9.42 DH parameter format: mainly
6468 to support use of 'q' parameter for RFC5114 parameters.
6472 * Add DH parameters from RFC5114 including test data to dhtest.
6476 * Support for automatic EC temporary key parameter selection. If enabled
6477 the most preferred EC parameters are automatically used instead of
6478 hardcoded fixed parameters. Now a server just has to call:
6479 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6480 support ECDH and use the most appropriate parameters.
6484 * Enhance and tidy EC curve and point format TLS extension code. Use
6485 static structures instead of allocation if default values are used.
6486 New ctrls to set curves we wish to support and to retrieve shared curves.
6487 Print out shared curves in s_server. New options to s_server and s_client
6488 to set list of supported curves.
6492 * New ctrls to retrieve supported signature algorithms and
6493 supported curve values as an array of NIDs. Extend openssl utility
6494 to print out received values.
6498 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6499 between NIDs and the more common NIST names such as "P-256". Enhance
6500 ecparam utility and ECC method to recognise the NIST names for curves.
6504 * Enhance SSL/TLS certificate chain handling to support different
6505 chains for each certificate instead of one chain in the parent SSL_CTX.
6509 * Support for fixed DH ciphersuite client authentication: where both
6510 server and client use DH certificates with common parameters.
6514 * Support for fixed DH ciphersuites: those requiring DH server
6519 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6521 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6522 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6523 X509_CINF_get_signature were reverted post internal team review.
6528 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6530 * OCSP Status Request extension unbounded memory growth
6532 A malicious client can send an excessively large OCSP Status Request
6533 extension. If that client continually requests renegotiation, sending a
6534 large OCSP Status Request extension each time, then there will be unbounded
6535 memory growth on the server. This will eventually lead to a Denial Of
6536 Service attack through memory exhaustion. Servers with a default
6537 configuration are vulnerable even if they do not support OCSP. Builds using
6538 the "no-ocsp" build time option are not affected.
6540 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6545 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6548 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6554 * OOB write in MDC2_Update()
6556 An overflow can occur in MDC2_Update() either if called directly or
6557 through the EVP_DigestUpdate() function using MDC2. If an attacker
6558 is able to supply very large amounts of input data after a previous
6559 call to EVP_EncryptUpdate() with a partial block then a length check
6560 can overflow resulting in a heap corruption.
6562 The amount of data needed is comparable to SIZE_MAX which is impractical
6565 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6570 * Malformed SHA512 ticket DoS
6572 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6573 DoS attack where a malformed ticket will result in an OOB read which will
6576 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6577 a custom server callback and ticket lookup mechanism.
6579 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6584 * OOB write in BN_bn2dec()
6586 The function BN_bn2dec() does not check the return value of BN_div_word().
6587 This can cause an OOB write if an application uses this function with an
6588 overly large BIGNUM. This could be a problem if an overly large certificate
6589 or CRL is printed out from an untrusted source. TLS is not affected because
6590 record limits will reject an oversized certificate before it is parsed.
6592 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6597 * OOB read in TS_OBJ_print_bio()
6599 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6600 the total length the OID text representation would use and not the amount
6601 of data written. This will result in OOB reads when large OIDs are
6604 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6609 * Pointer arithmetic undefined behaviour
6611 Avoid some undefined pointer arithmetic
6613 A common idiom in the codebase is to check limits in the following manner:
6616 Where "p" points to some malloc'd data of SIZE bytes and
6619 "len" here could be from some externally supplied data (e.g. from a TLS
6622 The rules of C pointer arithmetic are such that "p + len" is only well
6623 defined where len <= SIZE. Therefore, the above idiom is actually
6624 undefined behaviour.
6626 For example this could cause problems if some malloc implementation
6627 provides an address for "p" such that "p + len" actually overflows for
6628 values of len that are too big and therefore p + len < limit.
6630 This issue was reported to OpenSSL by Guido Vranken
6635 * Constant time flag not preserved in DSA signing
6637 Operations in the DSA signing algorithm should run in constant time in
6638 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6639 implementation means that a non-constant time codepath is followed for
6640 certain operations. This has been demonstrated through a cache-timing
6641 attack to be sufficient for an attacker to recover the private DSA key.
6643 This issue was reported by César Pereida (Aalto University), Billy Brumley
6644 (Tampere University of Technology), and Yuval Yarom (The University of
6645 Adelaide and NICTA).
6650 * DTLS buffered message DoS
6652 In a DTLS connection where handshake messages are delivered out-of-order
6653 those messages that OpenSSL is not yet ready to process will be buffered
6654 for later use. Under certain circumstances, a flaw in the logic means that
6655 those messages do not get removed from the buffer even though the handshake
6656 has been completed. An attacker could force up to approx. 15 messages to
6657 remain in the buffer when they are no longer required. These messages will
6658 be cleared when the DTLS connection is closed. The default maximum size for
6659 a message is 100k. Therefore, the attacker could force an additional 1500k
6660 to be consumed per connection. By opening many simultaneous connections an
6661 attacker could cause a DoS attack through memory exhaustion.
6663 This issue was reported to OpenSSL by Quan Luo.
6668 * DTLS replay protection DoS
6670 A flaw in the DTLS replay attack protection mechanism means that records
6671 that arrive for future epochs update the replay protection "window" before
6672 the MAC for the record has been validated. This could be exploited by an
6673 attacker by sending a record for the next epoch (which does not have to
6674 decrypt or have a valid MAC), with a very large sequence number. This means
6675 that all subsequent legitimate packets are dropped causing a denial of
6676 service for a specific DTLS connection.
6678 This issue was reported to OpenSSL by the OCAP audit team.
6683 * Certificate message OOB reads
6685 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6686 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6687 theoretical DoS risk but this has not been observed in practice on common
6690 The messages affected are client certificate, client certificate request
6691 and server certificate. As a result the attack can only be performed
6692 against a client or a server which enables client authentication.
6694 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6699 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6701 * Prevent padding oracle in AES-NI CBC MAC check
6703 A MITM attacker can use a padding oracle attack to decrypt traffic
6704 when the connection uses an AES CBC cipher and the server support
6707 This issue was introduced as part of the fix for Lucky 13 padding
6708 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6709 constant time by making sure that always the same bytes are read and
6710 compared against either the MAC or padding bytes. But it no longer
6711 checked that there was enough data to have both the MAC and padding
6714 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6719 * Fix EVP_EncodeUpdate overflow
6721 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6722 Base64 encoding of binary data. If an attacker is able to supply very large
6723 amounts of input data then a length check can overflow resulting in a heap
6726 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6727 the `PEM_write_bio*` family of functions. These are mainly used within the
6728 OpenSSL command line applications, so any application which processes data
6729 from an untrusted source and outputs it as a PEM file should be considered
6730 vulnerable to this issue. User applications that call these APIs directly
6731 with large amounts of untrusted data may also be vulnerable.
6733 This issue was reported by Guido Vranken.
6738 * Fix EVP_EncryptUpdate overflow
6740 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6741 is able to supply very large amounts of input data after a previous call to
6742 EVP_EncryptUpdate() with a partial block then a length check can overflow
6743 resulting in a heap corruption. Following an analysis of all OpenSSL
6744 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6745 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6746 the first called function after an EVP_EncryptInit(), and therefore that
6747 specific call must be safe. The second form is where the length passed to
6748 EVP_EncryptUpdate() can be seen from the code to be some small value and
6749 therefore there is no possibility of an overflow. Since all instances are
6750 one of these two forms, it is believed that there can be no overflows in
6751 internal code due to this problem. It should be noted that
6752 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6753 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6754 of these calls have also been analysed too and it is believed there are no
6755 instances in internal usage where an overflow could occur.
6757 This issue was reported by Guido Vranken.
6762 * Prevent ASN.1 BIO excessive memory allocation
6764 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6765 a short invalid encoding can casuse allocation of large amounts of memory
6766 potentially consuming excessive resources or exhausting memory.
6768 Any application parsing untrusted data through d2i BIO functions is
6769 affected. The memory based functions such as d2i_X509() are *not* affected.
6770 Since the memory based functions are used by the TLS library, TLS
6771 applications are not affected.
6773 This issue was reported by Brian Carpenter.
6780 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6781 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6782 in arbitrary stack data being returned in the buffer.
6784 This issue was reported by Guido Vranken.
6789 * Modify behavior of ALPN to invoke callback after SNI/servername
6790 callback, such that updates to the SSL_CTX affect ALPN.
6794 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6799 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6800 methods are enabled and ssl2 is disabled the methods return NULL.
6804 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6806 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6807 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6808 provide any "EXPORT" or "LOW" strength ciphers.
6812 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6813 is by default disabled at build-time. Builds that are not configured with
6814 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6815 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6816 will need to explicitly call either of:
6818 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6820 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6822 as appropriate. Even if either of those is used, or the application
6823 explicitly uses the version-specific SSLv2_method() or its client and
6824 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6825 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6826 ciphers, and SSLv2 56-bit DES are no longer available.
6831 * Fix a double-free in DSA code
6833 A double free bug was discovered when OpenSSL parses malformed DSA private
6834 keys and could lead to a DoS attack or memory corruption for applications
6835 that receive DSA private keys from untrusted sources. This scenario is
6838 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6844 * Disable SRP fake user seed to address a server memory leak.
6846 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6848 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6849 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6850 was changed to ignore the "fake user" SRP seed, even if the seed
6853 Users should use SRP_VBASE_get1_by_user instead. Note that in
6854 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6855 also that even though configuring the SRP seed attempts to hide
6856 invalid usernames by continuing the handshake with fake
6857 credentials, this behaviour is not constant time and no strong
6858 guarantees are made that the handshake is indistinguishable from
6859 that of a valid user.
6864 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6866 In the BN_hex2bn function the number of hex digits is calculated using an
6867 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6868 large values of `i` this can result in `bn_expand` not allocating any
6869 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6870 field as NULL leading to a subsequent NULL ptr deref. For very large values
6871 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6872 In this case memory is allocated to the internal BIGNUM data field, but it
6873 is insufficiently sized leading to heap corruption. A similar issue exists
6874 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6875 is ever called by user applications with very large untrusted hex/dec data.
6876 This is anticipated to be a rare occurrence.
6878 All OpenSSL internal usage of these functions use data that is not expected
6879 to be untrusted, e.g. config file data or application command line
6880 arguments. If user developed applications generate config file data based
6881 on untrusted data then it is possible that this could also lead to security
6882 consequences. This is also anticipated to be rare.
6884 This issue was reported to OpenSSL by Guido Vranken.
6889 * Fix memory issues in `BIO_*printf` functions
6891 The internal `fmtstr` function used in processing a "%s" format string in
6892 the `BIO_*printf` functions could overflow while calculating the length of a
6893 string and cause an OOB read when printing very long strings.
6895 Additionally the internal `doapr_outch` function can attempt to write to an
6896 OOB memory location (at an offset from the NULL pointer) in the event of a
6897 memory allocation failure. In 1.0.2 and below this could be caused where
6898 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6899 could be in processing a very long "%s" format string. Memory leaks can
6902 The first issue may mask the second issue dependent on compiler behaviour.
6903 These problems could enable attacks where large amounts of untrusted data
6904 is passed to the `BIO_*printf` functions. If applications use these functions
6905 in this way then they could be vulnerable. OpenSSL itself uses these
6906 functions when printing out human-readable dumps of ASN.1 data. Therefore
6907 applications that print this data could be vulnerable if the data is from
6908 untrusted sources. OpenSSL command line applications could also be
6909 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6910 as command line arguments.
6912 Libssl is not considered directly vulnerable. Additionally certificates etc
6913 received via remote connections via libssl are also unlikely to be able to
6914 trigger these issues because of message size limits enforced within libssl.
6916 This issue was reported to OpenSSL Guido Vranken.
6921 * Side channel attack on modular exponentiation
6923 A side-channel attack was found which makes use of cache-bank conflicts on
6924 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6925 of RSA keys. The ability to exploit this issue is limited as it relies on
6926 an attacker who has control of code in a thread running on the same
6927 hyper-threaded core as the victim thread which is performing decryptions.
6929 This issue was reported to OpenSSL by Yuval Yarom, The University of
6930 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6931 Nadia Heninger, University of Pennsylvania with more information at
6932 <http://cachebleed.info>.
6937 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6938 if no keysize is specified with default_bits. This fixes an
6939 omission in an earlier change that changed all RSA/DSA key generation
6940 commands to use 2048 bits by default.
6944 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6946 * Protection for DH small subgroup attacks
6948 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6949 switched on by default and cannot be disabled. This could have some
6954 * SSLv2 doesn't block disabled ciphers
6956 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6957 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6958 been disabled, provided that the SSLv2 protocol was not also disabled via
6961 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6962 and Sebastian Schinzel.
6967 * Reject DH handshakes with parameters shorter than 1024 bits.
6971 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6973 * Certificate verify crash with missing PSS parameter
6975 The signature verification routines will crash with a NULL pointer
6976 dereference if presented with an ASN.1 signature using the RSA PSS
6977 algorithm and absent mask generation function parameter. Since these
6978 routines are used to verify certificate signature algorithms this can be
6979 used to crash any certificate verification operation and exploited in a
6980 DoS attack. Any application which performs certificate verification is
6981 vulnerable including OpenSSL clients and servers which enable client
6984 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6989 * X509_ATTRIBUTE memory leak
6991 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6992 memory. This structure is used by the PKCS#7 and CMS routines so any
6993 application which reads PKCS#7 or CMS data from untrusted sources is
6994 affected. SSL/TLS is not affected.
6996 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7002 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7003 This changes the decoding behaviour for some invalid messages,
7004 though the change is mostly in the more lenient direction, and
7005 legacy behaviour is preserved as much as possible.
7009 * In DSA_generate_parameters_ex, if the provided seed is too short,
7010 use a random seed, as already documented.
7012 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7014 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7016 * Alternate chains certificate forgery
7018 During certificate verfification, OpenSSL will attempt to find an
7019 alternative certificate chain if the first attempt to build such a chain
7020 fails. An error in the implementation of this logic can mean that an
7021 attacker could cause certain checks on untrusted certificates to be
7022 bypassed, such as the CA flag, enabling them to use a valid leaf
7023 certificate to act as a CA and "issue" an invalid certificate.
7025 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7031 * Race condition handling PSK identify hint
7033 If PSK identity hints are received by a multi-threaded client then
7034 the values are wrongly updated in the parent SSL_CTX structure. This can
7035 result in a race condition potentially leading to a double free of the
7041 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7043 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7044 incompatibility in the handling of HMAC. The previous ABI has now been
7047 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7049 * Malformed ECParameters causes infinite loop
7051 When processing an ECParameters structure OpenSSL enters an infinite loop
7052 if the curve specified is over a specially malformed binary polynomial
7055 This can be used to perform denial of service against any
7056 system which processes public keys, certificate requests or
7057 certificates. This includes TLS clients and TLS servers with
7058 client authentication enabled.
7060 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7065 * Exploitable out-of-bounds read in X509_cmp_time
7067 X509_cmp_time does not properly check the length of the ASN1_TIME
7068 string and can read a few bytes out of bounds. In addition,
7069 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7072 An attacker can use this to craft malformed certificates and CRLs of
7073 various sizes and potentially cause a segmentation fault, resulting in
7074 a DoS on applications that verify certificates or CRLs. TLS clients
7075 that verify CRLs are affected. TLS clients and servers with client
7076 authentication enabled may be affected if they use custom verification
7079 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7080 independently by Hanno Böck.
7085 * PKCS7 crash with missing EnvelopedContent
7087 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7088 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7089 with missing content and trigger a NULL pointer dereference on parsing.
7091 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7092 structures from untrusted sources are affected. OpenSSL clients and
7093 servers are not affected.
7095 This issue was reported to OpenSSL by Michal Zalewski (Google).
7100 * CMS verify infinite loop with unknown hash function
7102 When verifying a signedData message the CMS code can enter an infinite loop
7103 if presented with an unknown hash function OID. This can be used to perform
7104 denial of service against any system which verifies signedData messages using
7106 This issue was reported to OpenSSL by Johannes Bauer.
7111 * Race condition handling NewSessionTicket
7113 If a NewSessionTicket is received by a multi-threaded client when attempting to
7114 reuse a previous ticket then a race condition can occur potentially leading to
7115 a double free of the ticket data.
7120 * Reject DH handshakes with parameters shorter than 768 bits.
7122 *Kurt Roeckx and Emilia Kasper*
7124 * dhparam: generate 2048-bit parameters by default.
7126 *Kurt Roeckx and Emilia Kasper*
7128 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7130 * Segmentation fault in ASN1_TYPE_cmp fix
7132 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7133 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7134 certificate signature algorithm consistency this can be used to crash any
7135 certificate verification operation and exploited in a DoS attack. Any
7136 application which performs certificate verification is vulnerable including
7137 OpenSSL clients and servers which enable client authentication.
7142 * ASN.1 structure reuse memory corruption fix
7144 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7145 memory corruption via an invalid write. Such reuse is and has been
7146 strongly discouraged and is believed to be rare.
7148 Applications that parse structures containing CHOICE or ANY DEFINED BY
7149 components may be affected. Certificate parsing (d2i_X509 and related
7150 functions) are however not affected. OpenSSL clients and servers are
7156 * PKCS7 NULL pointer dereferences fix
7158 The PKCS#7 parsing code does not handle missing outer ContentInfo
7159 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7160 missing content and trigger a NULL pointer dereference on parsing.
7162 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7163 otherwise parse PKCS#7 structures from untrusted sources are
7164 affected. OpenSSL clients and servers are not affected.
7166 This issue was reported to OpenSSL by Michal Zalewski (Google).
7171 * DoS via reachable assert in SSLv2 servers fix
7173 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7174 servers that both support SSLv2 and enable export cipher suites by sending
7175 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7177 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7178 (OpenSSL development team).
7183 * Use After Free following d2i_ECPrivatekey error fix
7185 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7186 could cause a use after free condition. This, in turn, could cause a double
7187 free in several private key parsing functions (such as d2i_PrivateKey
7188 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7189 for applications that receive EC private keys from untrusted
7190 sources. This scenario is considered rare.
7192 This issue was discovered by the BoringSSL project and fixed in their
7198 * X509_to_X509_REQ NULL pointer deref fix
7200 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7201 the certificate key is invalid. This function is rarely used in practice.
7203 This issue was discovered by Brian Carpenter.
7208 * Removed the export ciphers from the DEFAULT ciphers
7212 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7214 * Build fixes for the Windows and OpenVMS platforms
7216 *Matt Caswell and Richard Levitte*
7218 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7220 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7221 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7222 dereference. This could lead to a Denial Of Service attack. Thanks to
7223 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7228 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7229 dtls1_buffer_record function under certain conditions. In particular this
7230 could occur if an attacker sent repeated DTLS records with the same
7231 sequence number but for the next epoch. The memory leak could be exploited
7232 by an attacker in a Denial of Service attack through memory exhaustion.
7233 Thanks to Chris Mueller for reporting this issue.
7238 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7239 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7240 method would be set to NULL which could later result in a NULL pointer
7241 dereference. Thanks to Frank Schmirler for reporting this issue.
7246 * Abort handshake if server key exchange message is omitted for ephemeral
7249 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7250 reporting this issue.
7255 * Remove non-export ephemeral RSA code on client and server. This code
7256 violated the TLS standard by allowing the use of temporary RSA keys in
7257 non-export ciphersuites and could be used by a server to effectively
7258 downgrade the RSA key length used to a value smaller than the server
7259 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7260 INRIA or reporting this issue.
7265 * Fixed issue where DH client certificates are accepted without verification.
7266 An OpenSSL server will accept a DH certificate for client authentication
7267 without the certificate verify message. This effectively allows a client to
7268 authenticate without the use of a private key. This only affects servers
7269 which trust a client certificate authority which issues certificates
7270 containing DH keys: these are extremely rare and hardly ever encountered.
7271 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7277 * Ensure that the session ID context of an SSL is updated when its
7278 SSL_CTX is updated via SSL_set_SSL_CTX.
7280 The session ID context is typically set from the parent SSL_CTX,
7281 and can vary with the CTX.
7285 * Fix various certificate fingerprint issues.
7287 By using non-DER or invalid encodings outside the signed portion of a
7288 certificate the fingerprint can be changed without breaking the signature.
7289 Although no details of the signed portion of the certificate can be changed
7290 this can cause problems with some applications: e.g. those using the
7291 certificate fingerprint for blacklists.
7293 1. Reject signatures with non zero unused bits.
7295 If the BIT STRING containing the signature has non zero unused bits reject
7296 the signature. All current signature algorithms require zero unused bits.
7298 2. Check certificate algorithm consistency.
7300 Check the AlgorithmIdentifier inside TBS matches the one in the
7301 certificate signature. NB: this will result in signature failure
7302 errors for some broken certificates.
7304 Thanks to Konrad Kraszewski from Google for reporting this issue.
7306 3. Check DSA/ECDSA signatures use DER.
7308 Re-encode DSA/ECDSA signatures and compare with the original received
7309 signature. Return an error if there is a mismatch.
7311 This will reject various cases including garbage after signature
7312 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7313 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7314 (negative or with leading zeroes).
7316 Further analysis was conducted and fixes were developed by Stephen Henson
7317 of the OpenSSL core team.
7323 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7324 results on some platforms, including x86_64. This bug occurs at random
7325 with a very low probability, and is not known to be exploitable in any
7326 way, though its exact impact is difficult to determine. Thanks to Pieter
7327 Wuille (Blockstream) who reported this issue and also suggested an initial
7328 fix. Further analysis was conducted by the OpenSSL development team and
7329 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7330 the OpenSSL core team.
7335 * Do not resume sessions on the server if the negotiated protocol
7336 version does not match the session's version. Resuming with a different
7337 version, while not strictly forbidden by the RFC, is of questionable
7338 sanity and breaks all known clients.
7340 *David Benjamin, Emilia Käsper*
7342 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7343 early CCS messages during renegotiation. (Note that because
7344 renegotiation is encrypted, this early CCS was not exploitable.)
7348 * Tighten client-side session ticket handling during renegotiation:
7349 ensure that the client only accepts a session ticket if the server sends
7350 the extension anew in the ServerHello. Previously, a TLS client would
7351 reuse the old extension state and thus accept a session ticket if one was
7352 announced in the initial ServerHello.
7354 Similarly, ensure that the client requires a session ticket if one
7355 was advertised in the ServerHello. Previously, a TLS client would
7356 ignore a missing NewSessionTicket message.
7360 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7364 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7365 sends a carefully crafted handshake message, to cause OpenSSL to fail
7366 to free up to 64k of memory causing a memory leak. This could be
7367 exploited in a Denial Of Service attack. This issue affects OpenSSL
7368 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7369 whether SRTP is used or configured. Implementations of OpenSSL that
7370 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7372 The fix was developed by the OpenSSL team.
7377 * Session Ticket Memory Leak.
7379 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7380 integrity of that ticket is first verified. In the event of a session
7381 ticket integrity check failing, OpenSSL will fail to free memory
7382 causing a memory leak. By sending a large number of invalid session
7383 tickets an attacker could exploit this issue in a Denial Of Service
7389 * Build option no-ssl3 is incomplete.
7391 When OpenSSL is configured with "no-ssl3" as a build option, servers
7392 could accept and complete an SSL 3.0 handshake, and clients could be
7393 configured to send them.
7396 *Akamai and the OpenSSL team*
7398 * Add support for TLS_FALLBACK_SCSV.
7399 Client applications doing fallback retries should call
7400 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7403 *Adam Langley, Bodo Moeller*
7405 * Add additional DigestInfo checks.
7407 Re-encode DigestInto in DER and check against the original when
7408 verifying RSA signature: this will reject any improperly encoded
7409 DigestInfo structures.
7411 Note: this is a precautionary measure and no attacks are currently known.
7415 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7417 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7418 SRP code can be overrun an internal buffer. Add sanity check that
7419 g, A, B < N to SRP code.
7421 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7422 Group for discovering this issue.
7427 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7428 TLS 1.0 instead of higher protocol versions when the ClientHello message
7429 is badly fragmented. This allows a man-in-the-middle attacker to force a
7430 downgrade to TLS 1.0 even if both the server and the client support a
7431 higher protocol version, by modifying the client's TLS records.
7433 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7434 researching this issue.
7439 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7440 to a denial of service attack. A malicious server can crash the client
7441 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7442 ciphersuite and sending carefully crafted handshake messages.
7444 Thanks to Felix Gröbert (Google) for discovering and researching this
7450 * By sending carefully crafted DTLS packets an attacker could cause openssl
7451 to leak memory. This can be exploited through a Denial of Service attack.
7452 Thanks to Adam Langley for discovering and researching this issue.
7457 * An attacker can force openssl to consume large amounts of memory whilst
7458 processing DTLS handshake messages. This can be exploited through a
7459 Denial of Service attack.
7460 Thanks to Adam Langley for discovering and researching this issue.
7465 * An attacker can force an error condition which causes openssl to crash
7466 whilst processing DTLS packets due to memory being freed twice. This
7467 can be exploited through a Denial of Service attack.
7468 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7474 * If a multithreaded client connects to a malicious server using a resumed
7475 session and the server sends an ec point format extension it could write
7476 up to 255 bytes to freed memory.
7478 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7484 * A malicious server can crash an OpenSSL client with a null pointer
7485 dereference (read) by specifying an SRP ciphersuite even though it was not
7486 properly negotiated with the client. This can be exploited through a
7487 Denial of Service attack.
7489 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7490 discovering and researching this issue.
7495 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7496 X509_name_oneline, X509_name_print_ex et al. to leak some information
7497 from the stack. Applications may be affected if they echo pretty printing
7498 output to the attacker.
7500 Thanks to Ivan Fratric (Google) for discovering this issue.
7503 *Emilia Käsper, and Steve Henson*
7505 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7506 for corner cases. (Certain input points at infinity could lead to
7507 bogus results, with non-infinity inputs mapped to infinity too.)
7511 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7513 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7514 handshake can force the use of weak keying material in OpenSSL
7515 SSL/TLS clients and servers.
7517 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7518 researching this issue. ([CVE-2014-0224])
7520 *KIKUCHI Masashi, Steve Henson*
7522 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7523 OpenSSL DTLS client the code can be made to recurse eventually crashing
7526 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7529 *Imre Rad, Steve Henson*
7531 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7532 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7533 client or server. This is potentially exploitable to run arbitrary
7534 code on a vulnerable client or server.
7536 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7538 *Jüri Aedla, Steve Henson*
7540 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7541 are subject to a denial of service attack.
7543 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7544 this issue. ([CVE-2014-3470])
7546 *Felix Gröbert, Ivan Fratric, Steve Henson*
7548 * Harmonize version and its documentation. -f flag is used to display
7551 *mancha <mancha1@zoho.com>*
7553 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7554 in i2d_ECPrivateKey.
7556 *mancha <mancha1@zoho.com>*
7558 * Fix some double frees. These are not thought to be exploitable.
7560 *mancha <mancha1@zoho.com>*
7562 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7564 * A missing bounds check in the handling of the TLS heartbeat extension
7565 can be used to reveal up to 64k of memory to a connected client or
7568 Thanks for Neel Mehta of Google Security for discovering this bug and to
7569 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7570 preparing the fix ([CVE-2014-0160])
7572 *Adam Langley, Bodo Moeller*
7574 * Fix for the attack described in the paper "Recovering OpenSSL
7575 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7576 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7577 <http://eprint.iacr.org/2014/140>
7579 Thanks to Yuval Yarom and Naomi Benger for discovering this
7580 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7582 *Yuval Yarom and Naomi Benger*
7584 * TLS pad extension: draft-agl-tls-padding-03
7586 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7587 TLS client Hello record length value would otherwise be > 255 and
7588 less that 512 pad with a dummy extension containing zeroes so it
7589 is at least 512 bytes long.
7591 *Adam Langley, Steve Henson*
7593 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7595 * Fix for TLS record tampering bug. A carefully crafted invalid
7596 handshake could crash OpenSSL with a NULL pointer exception.
7597 Thanks to Anton Johansson for reporting this issues.
7600 * Keep original DTLS digest and encryption contexts in retransmission
7601 structures so we can use the previous session parameters if they need
7602 to be resent. ([CVE-2013-6450])
7606 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7607 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7608 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7609 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7610 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7611 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7613 *Rob Stradling, Adam Langley*
7615 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7617 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7618 supporting platforms or when small records were transferred.
7620 *Andy Polyakov, Steve Henson*
7622 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7624 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7626 This addresses the flaw in CBC record processing discovered by
7627 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7628 at: <http://www.isg.rhul.ac.uk/tls/>
7630 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7631 Security Group at Royal Holloway, University of London
7632 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7633 Emilia Käsper for the initial patch.
7636 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7638 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7639 ciphersuites which can be exploited in a denial of service attack.
7640 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7641 and detecting this bug and to Wolfgang Ettlinger
7642 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7647 * Return an error when checking OCSP signatures when key is NULL.
7648 This fixes a DoS attack. ([CVE-2013-0166])
7652 * Make openssl verify return errors.
7654 *Chris Palmer <palmer@google.com> and Ben Laurie*
7656 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7657 the right response is stapled. Also change SSL_get_certificate()
7658 so it returns the certificate actually sent.
7659 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7661 *Rob Stradling <rob.stradling@comodo.com>*
7663 * Fix possible deadlock when decoding public keys.
7667 * Don't use TLS 1.0 record version number in initial client hello
7672 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7674 * Sanity check record length before skipping explicit IV in TLS
7675 1.2, 1.1 and DTLS to fix DoS attack.
7677 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7678 fuzzing as a service testing platform.
7683 * Initialise tkeylen properly when encrypting CMS messages.
7684 Thanks to Solar Designer of Openwall for reporting this issue.
7688 * In FIPS mode don't try to use composite ciphers as they are not
7693 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7695 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7696 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7697 mean any application compiled against OpenSSL 1.0.0 headers setting
7698 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7699 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7700 0x10000000L Any application which was previously compiled against
7701 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7702 will need to be recompiled as a result. Letting be results in
7703 inability to disable specifically TLS 1.1 and in client context,
7704 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7708 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7709 disable just protocol X, but all protocols above X *if* there are
7710 protocols *below* X still enabled. In more practical terms it means
7711 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7712 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7713 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7718 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7720 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7721 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7722 in CRYPTO_realloc_clean.
7724 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7725 issue and to Adam Langley <agl@chromium.org> for fixing it.
7728 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7730 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7734 * Workarounds for some broken servers that "hang" if a client hello
7735 record length exceeds 255 bytes.
7737 1. Do not use record version number > TLS 1.0 in initial client
7738 hello: some (but not all) hanging servers will now work.
7739 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7740 the number of ciphers sent in the client hello. This should be
7741 set to an even number, such as 50, for example by passing:
7742 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7743 Most broken servers should now work.
7744 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7745 TLS 1.2 client support entirely.
7749 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7753 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7755 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7756 STRING form instead of a DigestInfo.
7760 * The format used for MDC2 RSA signatures is inconsistent between EVP
7761 and the RSA_sign/RSA_verify functions. This was made more apparent when
7762 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7763 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7764 the correct format in RSA_verify so both forms transparently work.
7768 * Some servers which support TLS 1.0 can choke if we initially indicate
7769 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7770 encrypted premaster secret. As a workaround use the maximum permitted
7771 client version in client hello, this should keep such servers happy
7772 and still work with previous versions of OpenSSL.
7776 * Add support for TLS/DTLS heartbeats.
7778 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7780 * Add support for SCTP.
7782 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7784 * Improved PRNG seeding for VOS.
7786 *Paul Green <Paul.Green@stratus.com>*
7788 * Extensive assembler packs updates, most notably:
7790 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7791 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7792 - x86_64: bit-sliced AES implementation;
7793 - ARM: NEON support, contemporary platforms optimizations;
7794 - s390x: z196 support;
7795 - `*`: GHASH and GF(2^m) multiplication implementations;
7799 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7800 (removal of unnecessary code)
7802 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7804 * Add TLS key material exporter from RFC 5705.
7808 * Add DTLS-SRTP negotiation from RFC 5764.
7812 * Add Next Protocol Negotiation,
7813 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7814 disabled with a no-npn flag to config or Configure. Code donated
7817 *Adam Langley <agl@google.com> and Ben Laurie*
7819 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7820 NIST-P256, NIST-P521, with constant-time single point multiplication on
7821 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7822 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7823 Code made available under Apache License version 2.0.
7825 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7826 line to include this in your build of OpenSSL, and run "make depend" (or
7827 "make update"). This enables the following EC_METHODs:
7829 EC_GFp_nistp224_method()
7830 EC_GFp_nistp256_method()
7831 EC_GFp_nistp521_method()
7833 EC_GROUP_new_by_curve_name() will automatically use these (while
7834 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7837 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7839 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7840 all platforms. Move ssize_t definition from e_os.h to the public
7841 header file e_os2.h as it now appears in public header file cms.h
7845 * New -sigopt option to the ca, req and x509 utilities. Additional
7846 signature parameters can be passed using this option and in
7851 * Add RSA PSS signing function. This will generate and set the
7852 appropriate AlgorithmIdentifiers for PSS based on those in the
7853 corresponding EVP_MD_CTX structure. No application support yet.
7857 * Support for companion algorithm specific ASN1 signing routines.
7858 New function ASN1_item_sign_ctx() signs a pre-initialised
7859 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7860 the appropriate parameters.
7864 * Add new algorithm specific ASN1 verification initialisation function
7865 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7866 handling will be the same no matter what EVP_PKEY_METHOD is used.
7867 Add a PSS handler to support verification of PSS signatures: checked
7868 against a number of sample certificates.
7872 * Add signature printing for PSS. Add PSS OIDs.
7874 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7876 * Add algorithm specific signature printing. An individual ASN1 method
7877 can now print out signatures instead of the standard hex dump.
7879 More complex signatures (e.g. PSS) can print out more meaningful
7880 information. Include DSA version that prints out the signature
7885 * Password based recipient info support for CMS library: implementing
7890 * Split password based encryption into PBES2 and PBKDF2 functions. This
7891 neatly separates the code into cipher and PBE sections and is required
7892 for some algorithms that split PBES2 into separate pieces (such as
7893 password based CMS).
7897 * Session-handling fixes:
7898 - Fix handling of connections that are resuming with a session ID,
7899 but also support Session Tickets.
7900 - Fix a bug that suppressed issuing of a new ticket if the client
7901 presented a ticket with an expired session.
7902 - Try to set the ticket lifetime hint to something reasonable.
7903 - Make tickets shorter by excluding irrelevant information.
7904 - On the client side, don't ignore renewed tickets.
7906 *Adam Langley, Bodo Moeller (Google)*
7908 * Fix PSK session representation.
7912 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7914 This work was sponsored by Intel.
7918 * Add GCM support to TLS library. Some custom code is needed to split
7919 the IV between the fixed (from PRF) and explicit (from TLS record)
7920 portions. This adds all GCM ciphersuites supported by RFC5288 and
7921 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7922 add a special AESGCM string for GCM only.
7926 * Expand range of ctrls for AES GCM. Permit setting invocation
7927 field on decrypt and retrieval of invocation field only on encrypt.
7931 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7932 As required by RFC5289 these ciphersuites cannot be used if for
7933 versions of TLS earlier than 1.2.
7937 * For FIPS capable OpenSSL interpret a NULL default public key method
7938 as unset and return the appropriate default but do *not* set the default.
7939 This means we can return the appropriate method in applications that
7940 switch between FIPS and non-FIPS modes.
7944 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7945 ENGINE is used then we cannot handle that in the FIPS module so we
7946 keep original code iff non-FIPS operations are allowed.
7950 * Add -attime option to openssl utilities.
7952 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7954 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7958 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7959 FIPS EC methods unconditionally for now.
7963 * New build option no-ec2m to disable characteristic 2 code.
7967 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7968 all cases can be covered as some introduce binary incompatibilities.
7972 * Redirect RSA operations to FIPS module including keygen,
7973 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7977 * Add similar low-level API blocking to ciphers.
7981 * low-level digest APIs are not approved in FIPS mode: any attempt
7982 to use these will cause a fatal error. Applications that *really* want
7983 to use them can use the `private_*` version instead.
7987 * Redirect cipher operations to FIPS module for FIPS builds.
7991 * Redirect digest operations to FIPS module for FIPS builds.
7995 * Update build system to add "fips" flag which will link in fipscanister.o
7996 for static and shared library builds embedding a signature if needed.
8000 * Output TLS supported curves in preference order instead of numerical
8001 order. This is currently hardcoded for the highest order curves first.
8002 This should be configurable so applications can judge speed vs strength.
8006 * Add TLS v1.2 server support for client authentication.
8010 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8015 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8016 FIPS modules versions.
8020 * Add TLS v1.2 client side support for client authentication. Keep cache
8021 of handshake records longer as we don't know the hash algorithm to use
8022 until after the certificate request message is received.
8026 * Initial TLS v1.2 client support. Add a default signature algorithms
8027 extension including all the algorithms we support. Parse new signature
8028 format in client key exchange. Relax some ECC signing restrictions for
8029 TLS v1.2 as indicated in RFC5246.
8033 * Add server support for TLS v1.2 signature algorithms extension. Switch
8034 to new signature format when needed using client digest preference.
8035 All server ciphersuites should now work correctly in TLS v1.2. No client
8036 support yet and no support for client certificates.
8040 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8041 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8042 ciphersuites. At present only RSA key exchange ciphersuites work with
8043 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8044 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8045 and version checking.
8049 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8050 with this defined it will not be affected by any changes to ssl internal
8051 structures. Add several utility functions to allow openssl application
8052 to work with OPENSSL_NO_SSL_INTERN defined.
8056 * A long standing patch to add support for SRP from EdelWeb (Peter
8057 Sylvester and Christophe Renou) was integrated.
8058 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8059 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8062 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8066 * Permit abbreviated handshakes when renegotiating using the function
8067 SSL_renegotiate_abbreviated().
8069 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8071 * Add call to ENGINE_register_all_complete() to
8072 ENGINE_load_builtin_engines(), so some implementations get used
8073 automatically instead of needing explicit application support.
8077 * Add support for TLS key exporter as described in RFC5705.
8079 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8081 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8082 a few changes are required:
8084 Add SSL_OP_NO_TLSv1_1 flag.
8085 Add TLSv1_1 methods.
8086 Update version checking logic to handle version 1.1.
8087 Add explicit IV handling (ported from DTLS code).
8088 Add command line options to s_client/s_server.
8095 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8097 * X509_ATTRIBUTE memory leak
8099 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8100 memory. This structure is used by the PKCS#7 and CMS routines so any
8101 application which reads PKCS#7 or CMS data from untrusted sources is
8102 affected. SSL/TLS is not affected.
8104 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8110 * Race condition handling PSK identify hint
8112 If PSK identity hints are received by a multi-threaded client then
8113 the values are wrongly updated in the parent SSL_CTX structure. This can
8114 result in a race condition potentially leading to a double free of the
8120 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8122 * Malformed ECParameters causes infinite loop
8124 When processing an ECParameters structure OpenSSL enters an infinite loop
8125 if the curve specified is over a specially malformed binary polynomial
8128 This can be used to perform denial of service against any
8129 system which processes public keys, certificate requests or
8130 certificates. This includes TLS clients and TLS servers with
8131 client authentication enabled.
8133 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8138 * Exploitable out-of-bounds read in X509_cmp_time
8140 X509_cmp_time does not properly check the length of the ASN1_TIME
8141 string and can read a few bytes out of bounds. In addition,
8142 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8145 An attacker can use this to craft malformed certificates and CRLs of
8146 various sizes and potentially cause a segmentation fault, resulting in
8147 a DoS on applications that verify certificates or CRLs. TLS clients
8148 that verify CRLs are affected. TLS clients and servers with client
8149 authentication enabled may be affected if they use custom verification
8152 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8153 independently by Hanno Böck.
8158 * PKCS7 crash with missing EnvelopedContent
8160 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8161 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8162 with missing content and trigger a NULL pointer dereference on parsing.
8164 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8165 structures from untrusted sources are affected. OpenSSL clients and
8166 servers are not affected.
8168 This issue was reported to OpenSSL by Michal Zalewski (Google).
8173 * CMS verify infinite loop with unknown hash function
8175 When verifying a signedData message the CMS code can enter an infinite loop
8176 if presented with an unknown hash function OID. This can be used to perform
8177 denial of service against any system which verifies signedData messages using
8179 This issue was reported to OpenSSL by Johannes Bauer.
8184 * Race condition handling NewSessionTicket
8186 If a NewSessionTicket is received by a multi-threaded client when attempting to
8187 reuse a previous ticket then a race condition can occur potentially leading to
8188 a double free of the ticket data.
8193 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8195 * Segmentation fault in ASN1_TYPE_cmp fix
8197 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8198 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8199 certificate signature algorithm consistency this can be used to crash any
8200 certificate verification operation and exploited in a DoS attack. Any
8201 application which performs certificate verification is vulnerable including
8202 OpenSSL clients and servers which enable client authentication.
8207 * ASN.1 structure reuse memory corruption fix
8209 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8210 memory corruption via an invalid write. Such reuse is and has been
8211 strongly discouraged and is believed to be rare.
8213 Applications that parse structures containing CHOICE or ANY DEFINED BY
8214 components may be affected. Certificate parsing (d2i_X509 and related
8215 functions) are however not affected. OpenSSL clients and servers are
8221 * PKCS7 NULL pointer dereferences fix
8223 The PKCS#7 parsing code does not handle missing outer ContentInfo
8224 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8225 missing content and trigger a NULL pointer dereference on parsing.
8227 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8228 otherwise parse PKCS#7 structures from untrusted sources are
8229 affected. OpenSSL clients and servers are not affected.
8231 This issue was reported to OpenSSL by Michal Zalewski (Google).
8236 * DoS via reachable assert in SSLv2 servers fix
8238 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8239 servers that both support SSLv2 and enable export cipher suites by sending
8240 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8242 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8243 (OpenSSL development team).
8248 * Use After Free following d2i_ECPrivatekey error fix
8250 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8251 could cause a use after free condition. This, in turn, could cause a double
8252 free in several private key parsing functions (such as d2i_PrivateKey
8253 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8254 for applications that receive EC private keys from untrusted
8255 sources. This scenario is considered rare.
8257 This issue was discovered by the BoringSSL project and fixed in their
8263 * X509_to_X509_REQ NULL pointer deref fix
8265 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8266 the certificate key is invalid. This function is rarely used in practice.
8268 This issue was discovered by Brian Carpenter.
8273 * Removed the export ciphers from the DEFAULT ciphers
8277 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8279 * Build fixes for the Windows and OpenVMS platforms
8281 *Matt Caswell and Richard Levitte*
8283 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8285 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8286 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8287 dereference. This could lead to a Denial Of Service attack. Thanks to
8288 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8293 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8294 dtls1_buffer_record function under certain conditions. In particular this
8295 could occur if an attacker sent repeated DTLS records with the same
8296 sequence number but for the next epoch. The memory leak could be exploited
8297 by an attacker in a Denial of Service attack through memory exhaustion.
8298 Thanks to Chris Mueller for reporting this issue.
8303 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8304 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8305 method would be set to NULL which could later result in a NULL pointer
8306 dereference. Thanks to Frank Schmirler for reporting this issue.
8311 * Abort handshake if server key exchange message is omitted for ephemeral
8314 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8315 reporting this issue.
8320 * Remove non-export ephemeral RSA code on client and server. This code
8321 violated the TLS standard by allowing the use of temporary RSA keys in
8322 non-export ciphersuites and could be used by a server to effectively
8323 downgrade the RSA key length used to a value smaller than the server
8324 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8325 INRIA or reporting this issue.
8330 * Fixed issue where DH client certificates are accepted without verification.
8331 An OpenSSL server will accept a DH certificate for client authentication
8332 without the certificate verify message. This effectively allows a client to
8333 authenticate without the use of a private key. This only affects servers
8334 which trust a client certificate authority which issues certificates
8335 containing DH keys: these are extremely rare and hardly ever encountered.
8336 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8342 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8343 results on some platforms, including x86_64. This bug occurs at random
8344 with a very low probability, and is not known to be exploitable in any
8345 way, though its exact impact is difficult to determine. Thanks to Pieter
8346 Wuille (Blockstream) who reported this issue and also suggested an initial
8347 fix. Further analysis was conducted by the OpenSSL development team and
8348 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8349 the OpenSSL core team.
8354 * Fix various certificate fingerprint issues.
8356 By using non-DER or invalid encodings outside the signed portion of a
8357 certificate the fingerprint can be changed without breaking the signature.
8358 Although no details of the signed portion of the certificate can be changed
8359 this can cause problems with some applications: e.g. those using the
8360 certificate fingerprint for blacklists.
8362 1. Reject signatures with non zero unused bits.
8364 If the BIT STRING containing the signature has non zero unused bits reject
8365 the signature. All current signature algorithms require zero unused bits.
8367 2. Check certificate algorithm consistency.
8369 Check the AlgorithmIdentifier inside TBS matches the one in the
8370 certificate signature. NB: this will result in signature failure
8371 errors for some broken certificates.
8373 Thanks to Konrad Kraszewski from Google for reporting this issue.
8375 3. Check DSA/ECDSA signatures use DER.
8377 Reencode DSA/ECDSA signatures and compare with the original received
8378 signature. Return an error if there is a mismatch.
8380 This will reject various cases including garbage after signature
8381 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8382 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8383 (negative or with leading zeroes).
8385 Further analysis was conducted and fixes were developed by Stephen Henson
8386 of the OpenSSL core team.
8392 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8394 * Session Ticket Memory Leak.
8396 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8397 integrity of that ticket is first verified. In the event of a session
8398 ticket integrity check failing, OpenSSL will fail to free memory
8399 causing a memory leak. By sending a large number of invalid session
8400 tickets an attacker could exploit this issue in a Denial Of Service
8406 * Build option no-ssl3 is incomplete.
8408 When OpenSSL is configured with "no-ssl3" as a build option, servers
8409 could accept and complete an SSL 3.0 handshake, and clients could be
8410 configured to send them.
8413 *Akamai and the OpenSSL team*
8415 * Add support for TLS_FALLBACK_SCSV.
8416 Client applications doing fallback retries should call
8417 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8420 *Adam Langley, Bodo Moeller*
8422 * Add additional DigestInfo checks.
8424 Reencode DigestInto in DER and check against the original when
8425 verifying RSA signature: this will reject any improperly encoded
8426 DigestInfo structures.
8428 Note: this is a precautionary measure and no attacks are currently known.
8432 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8434 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8435 to a denial of service attack. A malicious server can crash the client
8436 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8437 ciphersuite and sending carefully crafted handshake messages.
8439 Thanks to Felix Gröbert (Google) for discovering and researching this
8445 * By sending carefully crafted DTLS packets an attacker could cause openssl
8446 to leak memory. This can be exploited through a Denial of Service attack.
8447 Thanks to Adam Langley for discovering and researching this issue.
8452 * An attacker can force openssl to consume large amounts of memory whilst
8453 processing DTLS handshake messages. This can be exploited through a
8454 Denial of Service attack.
8455 Thanks to Adam Langley for discovering and researching this issue.
8460 * An attacker can force an error condition which causes openssl to crash
8461 whilst processing DTLS packets due to memory being freed twice. This
8462 can be exploited through a Denial of Service attack.
8463 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8469 * If a multithreaded client connects to a malicious server using a resumed
8470 session and the server sends an ec point format extension it could write
8471 up to 255 bytes to freed memory.
8473 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8479 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8480 X509_name_oneline, X509_name_print_ex et al. to leak some information
8481 from the stack. Applications may be affected if they echo pretty printing
8482 output to the attacker.
8484 Thanks to Ivan Fratric (Google) for discovering this issue.
8487 *Emilia Käsper, and Steve Henson*
8489 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8490 for corner cases. (Certain input points at infinity could lead to
8491 bogus results, with non-infinity inputs mapped to infinity too.)
8495 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8497 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8498 handshake can force the use of weak keying material in OpenSSL
8499 SSL/TLS clients and servers.
8501 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8502 researching this issue. ([CVE-2014-0224])
8504 *KIKUCHI Masashi, Steve Henson*
8506 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8507 OpenSSL DTLS client the code can be made to recurse eventually crashing
8510 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8513 *Imre Rad, Steve Henson*
8515 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8516 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8517 client or server. This is potentially exploitable to run arbitrary
8518 code on a vulnerable client or server.
8520 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8522 *Jüri Aedla, Steve Henson*
8524 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8525 are subject to a denial of service attack.
8527 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8528 this issue. ([CVE-2014-3470])
8530 *Felix Gröbert, Ivan Fratric, Steve Henson*
8532 * Harmonize version and its documentation. -f flag is used to display
8535 *mancha <mancha1@zoho.com>*
8537 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8538 in i2d_ECPrivateKey.
8540 *mancha <mancha1@zoho.com>*
8542 * Fix some double frees. These are not thought to be exploitable.
8544 *mancha <mancha1@zoho.com>*
8546 * Fix for the attack described in the paper "Recovering OpenSSL
8547 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8548 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8549 <http://eprint.iacr.org/2014/140>
8551 Thanks to Yuval Yarom and Naomi Benger for discovering this
8552 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8554 *Yuval Yarom and Naomi Benger*
8556 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8558 * Keep original DTLS digest and encryption contexts in retransmission
8559 structures so we can use the previous session parameters if they need
8560 to be resent. ([CVE-2013-6450])
8564 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8565 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8566 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8567 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8568 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8569 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8571 *Rob Stradling, Adam Langley*
8573 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8575 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8577 This addresses the flaw in CBC record processing discovered by
8578 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8579 at: <http://www.isg.rhul.ac.uk/tls/>
8581 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8582 Security Group at Royal Holloway, University of London
8583 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8584 Emilia Käsper for the initial patch.
8587 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8589 * Return an error when checking OCSP signatures when key is NULL.
8590 This fixes a DoS attack. ([CVE-2013-0166])
8594 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8595 the right response is stapled. Also change SSL_get_certificate()
8596 so it returns the certificate actually sent.
8597 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8598 (This is a backport)
8600 *Rob Stradling <rob.stradling@comodo.com>*
8602 * Fix possible deadlock when decoding public keys.
8606 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8608 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8611 * Sanity check record length before skipping explicit IV in DTLS
8614 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8615 fuzzing as a service testing platform.
8620 * Initialise tkeylen properly when encrypting CMS messages.
8621 Thanks to Solar Designer of Openwall for reporting this issue.
8625 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8627 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8628 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8629 in CRYPTO_realloc_clean.
8631 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8632 issue and to Adam Langley <agl@chromium.org> for fixing it.
8635 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8637 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8639 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8640 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8641 content decryption and always return the same error. Note: this attack
8642 needs on average 2^20 messages so it only affects automated senders. The
8643 old behaviour can be re-enabled in the CMS code by setting the
8644 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8645 an MMA defence is not necessary.
8646 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8647 this issue. ([CVE-2012-0884])
8651 * Fix CVE-2011-4619: make sure we really are receiving a
8652 client hello before rejecting multiple SGC restarts. Thanks to
8653 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8657 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8659 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8660 Thanks to Antonio Martin, Enterprise Secure Access Research and
8661 Development, Cisco Systems, Inc. for discovering this bug and
8662 preparing a fix. ([CVE-2012-0050])
8666 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8668 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8669 of the Vaudenay padding oracle attack on CBC mode encryption
8670 which enables an efficient plaintext recovery attack against
8671 the OpenSSL implementation of DTLS. Their attack exploits timing
8672 differences arising during decryption processing. A research
8673 paper describing this attack can be found at:
8674 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8675 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8676 Security Group at Royal Holloway, University of London
8677 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8678 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8679 for preparing the fix. ([CVE-2011-4108])
8681 *Robin Seggelmann, Michael Tuexen*
8683 * Clear bytes used for block padding of SSL 3.0 records.
8686 *Adam Langley (Google)*
8688 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8689 Kadianakis <desnacked@gmail.com> for discovering this issue and
8690 Adam Langley for preparing the fix. ([CVE-2011-4619])
8692 *Adam Langley (Google)*
8694 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8696 *Andrey Kulikov <amdeich@gmail.com>*
8698 * Prevent malformed RFC3779 data triggering an assertion failure.
8699 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8700 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8702 *Rob Austein <sra@hactrn.net>*
8704 * Improved PRNG seeding for VOS.
8706 *Paul Green <Paul.Green@stratus.com>*
8708 * Fix ssl_ciph.c set-up race.
8710 *Adam Langley (Google)*
8712 * Fix spurious failures in ecdsatest.c.
8714 *Emilia Käsper (Google)*
8716 * Fix the BIO_f_buffer() implementation (which was mixing different
8717 interpretations of the `..._len` fields).
8719 *Adam Langley (Google)*
8721 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8722 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8723 threads won't reuse the same blinding coefficients.
8725 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8726 lock to call BN_BLINDING_invert_ex, and avoids one use of
8727 BN_BLINDING_update for each BN_BLINDING structure (previously,
8728 the last update always remained unused).
8730 *Emilia Käsper (Google)*
8732 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8734 *Bob Buckholz (Google)*
8736 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8738 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8739 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8741 *Kaspar Brand <ossl@velox.ch>*
8743 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8744 for multi-threaded use of ECDH. ([CVE-2011-3210])
8746 *Adam Langley (Google)*
8748 * Fix x509_name_ex_d2i memory leak on bad inputs.
8752 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8753 signature public key algorithm by using OID xref utilities instead.
8754 Before this you could only use some ECC ciphersuites with SHA1 only.
8758 * Add protection against ECDSA timing attacks as mentioned in the paper
8759 by Billy Bob Brumley and Nicola Tuveri, see:
8760 <http://eprint.iacr.org/2011/232.pdf>
8762 *Billy Bob Brumley and Nicola Tuveri*
8764 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8766 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8768 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8770 * Fix bug in string printing code: if *any* escaping is enabled we must
8771 escape the escape character (backslash) or the resulting string is
8776 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8778 * Disable code workaround for ancient and obsolete Netscape browsers
8779 and servers: an attacker can use it in a ciphersuite downgrade attack.
8780 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8784 * Fixed J-PAKE implementation error, originally discovered by
8785 Sebastien Martini, further info and confirmation from Stefan
8786 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8790 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8792 * Fix extension code to avoid race conditions which can result in a buffer
8793 overrun vulnerability: resumed sessions must not be modified as they can
8794 be shared by multiple threads. CVE-2010-3864
8798 * Fix WIN32 build system to correctly link an ENGINE directory into
8803 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8805 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8808 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8810 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8812 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8813 context. The operation can be customised via the ctrl mechanism in
8814 case ENGINEs want to include additional functionality.
8818 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8822 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8823 output hashes compatible with older versions of OpenSSL.
8825 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8827 * Fix compression algorithm handling: if resuming a session use the
8828 compression algorithm of the resumed session instead of determining
8829 it from client hello again. Don't allow server to change algorithm.
8833 * Add load_crls() function to commands tidying load_certs() too. Add option
8834 to verify utility to allow additional CRLs to be included.
8838 * Update OCSP request code to permit adding custom headers to the request:
8839 some responders need this.
8843 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8846 *Julia Lawall <julia@diku.dk>*
8848 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8849 needlessly dereferenced structures, used obsolete functions and
8850 didn't handle all updated verify codes correctly.
8854 * Disable MD2 in the default configuration.
8858 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8859 indicate the initial BIO being pushed or popped. This makes it possible
8860 to determine whether the BIO is the one explicitly called or as a result
8861 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8862 it handles reference counts correctly and doesn't zero out the I/O bio
8863 when it is not being explicitly popped. WARNING: applications which
8864 included workarounds for the old buggy behaviour will need to be modified
8865 or they could free up already freed BIOs.
8869 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8870 renaming to all platforms (within the 0.9.8 branch, this was
8871 done conditionally on Netware platforms to avoid a name clash).
8873 *Guenter <lists@gknw.net>*
8875 * Add ECDHE and PSK support to DTLS.
8877 *Michael Tuexen <tuexen@fh-muenster.de>*
8879 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8884 * Add "missing" function EVP_MD_flags() (without this the only way to
8885 retrieve a digest flags is by accessing the structure directly. Update
8886 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8887 or cipher is registered as in the "from" argument. Print out all
8888 registered digests in the dgst usage message instead of manually
8889 attempting to work them out.
8893 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8894 this allows the use of compression and extensions. Change default cipher
8895 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8896 by default unless an application cipher string requests it.
8900 * Alter match criteria in PKCS12_parse(). It used to try to use local
8901 key ids to find matching certificates and keys but some PKCS#12 files
8902 don't follow the (somewhat unwritten) rules and this strategy fails.
8903 Now just gather all certificates together and the first private key
8904 then look for the first certificate that matches the key.
8908 * Support use of registered digest and cipher names for dgst and cipher
8909 commands instead of having to add each one as a special case. So now
8916 openssl dgst -sha256 foo
8918 and this works for ENGINE based algorithms too.
8922 * Update Gost ENGINE to support parameter files.
8924 *Victor B. Wagner <vitus@cryptocom.ru>*
8926 * Support GeneralizedTime in ca utility.
8928 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8930 * Enhance the hash format used for certificate directory links. The new
8931 form uses the canonical encoding (meaning equivalent names will work
8932 even if they aren't identical) and uses SHA1 instead of MD5. This form
8933 is incompatible with the older format and as a result c_rehash should
8934 be used to rebuild symbolic links.
8938 * Make PKCS#8 the default write format for private keys, replacing the
8939 traditional format. This form is standardised, more secure and doesn't
8940 include an implicit MD5 dependency.
8944 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8945 committed to OpenSSL should pass this lot as a minimum.
8949 * Add session ticket override functionality for use by EAP-FAST.
8951 *Jouni Malinen <j@w1.fi>*
8953 * Modify HMAC functions to return a value. Since these can be implemented
8954 in an ENGINE errors can occur.
8958 * Type-checked OBJ_bsearch_ex.
8962 * Type-checked OBJ_bsearch. Also some constification necessitated
8963 by type-checking. Still to come: TXT_DB, bsearch(?),
8964 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8969 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8970 seconds to a tm structure directly, instead of going through OS
8971 specific date routines. This avoids any issues with OS routines such
8972 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8973 and X509_time_adj_ex() to cover the extended range. The existing
8974 X509_time_adj() is still usable and will no longer have any date issues.
8978 * Delta CRL support. New use deltas option which will attempt to locate
8979 and search any appropriate delta CRLs available.
8981 This work was sponsored by Google.
8985 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8986 code and add additional score elements. Validate alternate CRL paths
8987 as part of the CRL checking and indicate a new error "CRL path validation
8988 error" in this case. Applications wanting additional details can use
8989 the verify callback and check the new "parent" field. If this is not
8990 NULL CRL path validation is taking place. Existing applications won't
8991 see this because it requires extended CRL support which is off by
8994 This work was sponsored by Google.
8998 * Support for freshest CRL extension.
9000 This work was sponsored by Google.
9004 * Initial indirect CRL support. Currently only supported in the CRLs
9005 passed directly and not via lookup. Process certificate issuer
9006 CRL entry extension and lookup CRL entries by bother issuer name
9007 and serial number. Check and process CRL issuer entry in IDP extension.
9009 This work was sponsored by Google.
9013 * Add support for distinct certificate and CRL paths. The CRL issuer
9014 certificate is validated separately in this case. Only enabled if
9015 an extended CRL support flag is set: this flag will enable additional
9016 CRL functionality in future.
9018 This work was sponsored by Google.
9022 * Add support for policy mappings extension.
9024 This work was sponsored by Google.
9028 * Fixes to pathlength constraint, self issued certificate handling,
9029 policy processing to align with RFC3280 and PKITS tests.
9031 This work was sponsored by Google.
9035 * Support for name constraints certificate extension. DN, email, DNS
9036 and URI types are currently supported.
9038 This work was sponsored by Google.
9042 * To cater for systems that provide a pointer-based thread ID rather
9043 than numeric, deprecate the current numeric thread ID mechanism and
9044 replace it with a structure and associated callback type. This
9045 mechanism allows a numeric "hash" to be extracted from a thread ID in
9046 either case, and on platforms where pointers are larger than 'long',
9047 mixing is done to help ensure the numeric 'hash' is usable even if it
9048 can't be guaranteed unique. The default mechanism is to use "&errno"
9049 as a pointer-based thread ID to distinguish between threads.
9051 Applications that want to provide their own thread IDs should now use
9052 CRYPTO_THREADID_set_callback() to register a callback that will call
9053 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9055 Note that ERR_remove_state() is now deprecated, because it is tied
9056 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9057 to free the current thread's error state should be replaced by
9058 ERR_remove_thread_state(NULL).
9060 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9061 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9062 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9063 application was previously providing a numeric thread callback that
9064 was inappropriate for distinguishing threads, then uniqueness might
9065 have been obtained with &errno that happened immediately in the
9066 intermediate development versions of OpenSSL; this is no longer the
9067 case, the numeric thread callback will now override the automatic use
9070 *Geoff Thorpe, with help from Bodo Moeller*
9072 * Initial support for different CRL issuing certificates. This covers a
9073 simple case where the self issued certificates in the chain exist and
9074 the real CRL issuer is higher in the existing chain.
9076 This work was sponsored by Google.
9080 * Removed effectively defunct crypto/store from the build.
9084 * Revamp of STACK to provide stronger type-checking. Still to come:
9085 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9086 ASN1_STRING, CONF_VALUE.
9090 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9091 RAM on SSL connections. This option can save about 34k per idle SSL.
9095 * Revamp of LHASH to provide stronger type-checking. Still to come:
9096 STACK, TXT_DB, bsearch, qsort.
9100 * Initial support for Cryptographic Message Syntax (aka CMS) based
9101 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9102 support for data, signedData, compressedData, digestedData and
9103 encryptedData, envelopedData types included. Scripts to check against
9104 RFC4134 examples draft and interop and consistency checks of many
9105 content types and variants.
9109 * Add options to enc utility to support use of zlib compression BIO.
9113 * Extend mk1mf to support importing of options and assembly language
9114 files from Configure script, currently only included in VC-WIN32.
9115 The assembly language rules can now optionally generate the source
9116 files from the associated perl scripts.
9120 * Implement remaining functionality needed to support GOST ciphersuites.
9121 Interop testing has been performed using CryptoPro implementations.
9123 *Victor B. Wagner <vitus@cryptocom.ru>*
9125 * s390x assembler pack.
9129 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9134 * Implement Opaque PRF Input TLS extension as specified in
9135 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9136 official specification yet and no extension type assignment by
9137 IANA exists, this extension (for now) will have to be explicitly
9138 enabled when building OpenSSL by providing the extension number
9139 to use. For example, specify an option
9141 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9143 to the "config" or "Configure" script to enable the extension,
9144 assuming extension number 0x9527 (which is a completely arbitrary
9145 and unofficial assignment based on the MD5 hash of the Internet
9146 Draft). Note that by doing so, you potentially lose
9147 interoperability with other TLS implementations since these might
9148 be using the same extension number for other purposes.
9150 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9151 opaque PRF input value to use in the handshake. This will create
9152 an internal copy of the length-'len' string at 'src', and will
9153 return non-zero for success.
9155 To get more control and flexibility, provide a callback function
9158 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9159 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9163 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9166 Callback function 'cb' will be called in handshakes, and is
9167 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9168 Argument 'arg' is for application purposes (the value as given to
9169 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9170 be provided to the callback function). The callback function
9171 has to return non-zero to report success: usually 1 to use opaque
9172 PRF input just if possible, or 2 to enforce use of the opaque PRF
9173 input. In the latter case, the library will abort the handshake
9174 if opaque PRF input is not successfully negotiated.
9176 Arguments 'peerinput' and 'len' given to the callback function
9177 will always be NULL and 0 in the case of a client. A server will
9178 see the client's opaque PRF input through these variables if
9179 available (NULL and 0 otherwise). Note that if the server
9180 provides an opaque PRF input, the length must be the same as the
9181 length of the client's opaque PRF input.
9183 Note that the callback function will only be called when creating
9184 a new session (session resumption can resume whatever was
9185 previously negotiated), and will not be called in SSL 2.0
9186 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9187 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9188 for applications that need to enforce opaque PRF input.
9192 * Update ssl code to support digests other than SHA1+MD5 for handshake
9195 *Victor B. Wagner <vitus@cryptocom.ru>*
9197 * Add RFC4507 support to OpenSSL. This includes the corrections in
9198 RFC4507bis. The encrypted ticket format is an encrypted encoded
9199 SSL_SESSION structure, that way new session features are automatically
9202 If a client application caches session in an SSL_SESSION structure
9203 support is transparent because tickets are now stored in the encoded
9206 The SSL_CTX structure automatically generates keys for ticket
9207 protection in servers so again support should be possible
9208 with no application modification.
9210 If a client or server wishes to disable RFC4507 support then the option
9211 SSL_OP_NO_TICKET can be set.
9213 Add a TLS extension debugging callback to allow the contents of any client
9214 or server extensions to be examined.
9216 This work was sponsored by Google.
9220 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9221 OpenSSL should now compile cleanly on gcc 4.2
9223 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9225 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9226 support including streaming MAC support: this is required for GOST
9227 ciphersuite support.
9229 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9231 * Add option -stream to use PKCS#7 streaming in smime utility. New
9232 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9233 to output in BER and PEM format.
9237 * Experimental support for use of HMAC via EVP_PKEY interface. This
9238 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9239 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9240 ENGINE support for HMAC keys which are unextractable. New -mac and
9241 -macopt options to dgst utility.
9245 * New option -sigopt to dgst utility. Update dgst to use
9246 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9247 alternative signing parameters such as X9.31 or PSS in the dgst
9252 * Change ssl_cipher_apply_rule(), the internal function that does
9253 the work each time a ciphersuite string requests enabling
9254 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9255 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9256 the order of disabled ciphersuites such that those ciphersuites
9257 that most recently went from enabled to disabled not only stay
9258 in order with respect to each other, but also have higher priority
9259 than other disabled ciphersuites the next time ciphersuites are
9262 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9263 the same ciphersuites as with "HIGH" alone, but in a specific
9264 order where the PSK ciphersuites come first (since they are the
9265 most recently disabled ciphersuites when "HIGH" is parsed).
9267 Also, change ssl_create_cipher_list() (using this new
9268 functionality) such that between otherwise identical
9269 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9274 * Change ssl_create_cipher_list() so that it automatically
9275 arranges the ciphersuites in reasonable order before starting
9276 to process the rule string. Thus, the definition for "DEFAULT"
9277 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9278 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9279 This makes it much easier to arrive at a reasonable default order
9280 in applications for which anonymous ciphers are OK (meaning
9281 that you can't actually use DEFAULT).
9283 *Bodo Moeller; suggested by Victor Duchovni*
9285 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9286 processing) into multiple integers instead of setting
9287 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9288 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9289 (These masks as well as the individual bit definitions are hidden
9290 away into the non-exported interface ssl/ssl_locl.h, so this
9291 change to the definition of the SSL_CIPHER structure shouldn't
9292 affect applications.) This give us more bits for each of these
9293 categories, so there is no longer a need to coagulate AES128 and
9294 AES256 into a single algorithm bit, and to coagulate Camellia128
9295 and Camellia256 into a single algorithm bit, which has led to all
9298 Thus, among other things, the kludge introduced in 0.9.7m and
9299 0.9.8e for masking out AES256 independently of AES128 or masking
9300 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9302 With the change, we also introduce new ciphersuite aliases that
9303 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9308 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9309 Use the leftmost N bytes of the signature input if the input is
9310 larger than the prime q (with N being the size in bytes of q).
9314 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9315 it yet and it is largely untested.
9319 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9323 * Initial incomplete changes to avoid need for function casts in OpenSSL
9324 some compilers (gcc 4.2 and later) reject their use. Safestack is
9325 reimplemented. Update ASN1 to avoid use of legacy functions.
9329 * Win32/64 targets are linked with Winsock2.
9333 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9334 to external functions. This can be used to increase CRL handling
9335 efficiency especially when CRLs are very large by (for example) storing
9336 the CRL revoked certificates in a database.
9340 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9341 new CRLs added to a directory can be used. New command line option
9342 -verify_return_error to s_client and s_server. This causes real errors
9343 to be returned by the verify callback instead of carrying on no matter
9344 what. This reflects the way a "real world" verify callback would behave.
9348 * GOST engine, supporting several GOST algorithms and public key formats.
9349 Kindly donated by Cryptocom.
9353 * Partial support for Issuing Distribution Point CRL extension. CRLs
9354 partitioned by DP are handled but no indirect CRL or reason partitioning
9355 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9356 selected via a scoring technique which handles IDP and AKID in CRLs.
9360 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9361 will ultimately be used for all verify operations: this will remove the
9362 X509_STORE dependency on certificate verification and allow alternative
9363 lookup methods. X509_STORE based implementations of these two callbacks.
9367 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9368 Modify get_crl() to find a valid (unexpired) CRL if possible.
9372 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9373 this would be called X509_CRL_cmp() but that name is already used by
9374 a function that just compares CRL issuer names. Cache several CRL
9375 extensions in X509_CRL structure and cache CRLDP in X509.
9379 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9380 this maps equivalent X509_NAME structures into a consistent structure.
9381 Name comparison can then be performed rapidly using memcmp().
9385 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9390 * Allow digests to supply their own micalg string for S/MIME type using
9391 the ctrl EVP_MD_CTRL_MICALG.
9395 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9396 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9397 ctrl. It can then customise the structure before and/or after signing
9402 * New function OBJ_add_sigid() to allow application defined signature OIDs
9403 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9404 to free up any added signature OIDs.
9408 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9409 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9410 digest and cipher tables. New options added to openssl utility:
9411 list-message-digest-algorithms and list-cipher-algorithms.
9415 * Change the array representation of binary polynomials: the list
9416 of degrees of non-zero coefficients is now terminated with -1.
9417 Previously it was terminated with 0, which was also part of the
9418 value; thus, the array representation was not applicable to
9419 polynomials where t^0 has coefficient zero. This change makes
9420 the array representation useful in a more general context.
9424 * Various modifications and fixes to SSL/TLS cipher string
9425 handling. For ECC, the code now distinguishes between fixed ECDH
9426 with RSA certificates on the one hand and with ECDSA certificates
9427 on the other hand, since these are separate ciphersuites. The
9428 unused code for Fortezza ciphersuites has been removed.
9430 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9431 (not "ECDHE"). For consistency with the code for DH
9432 certificates, use of ECDH certificates is now considered ECDH
9433 authentication, not RSA or ECDSA authentication (the latter is
9434 merely the CA's signing algorithm and not actively used in the
9437 The temporary ciphersuite alias "ECCdraft" is no longer
9438 available, and ECC ciphersuites are no longer excluded from "ALL"
9439 and "DEFAULT". The following aliases now exist for RFC 4492
9440 ciphersuites, most of these by analogy with the DH case:
9442 kECDHr - ECDH cert, signed with RSA
9443 kECDHe - ECDH cert, signed with ECDSA
9444 kECDH - ECDH cert (signed with either RSA or ECDSA)
9445 kEECDH - ephemeral ECDH
9446 ECDH - ECDH cert or ephemeral ECDH
9452 AECDH - anonymous ECDH
9453 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9457 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9458 Use correct micalg parameters depending on digest(s) in signed message.
9462 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9463 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9467 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9468 an engine to register a method. Add ENGINE lookups for methods and
9469 functional reference processing.
9473 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9474 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9479 * New -resign option to smime utility. This adds one or more signers
9480 to an existing PKCS#7 signedData structure. Also -md option to use an
9481 alternative message digest algorithm for signing.
9485 * Tidy up PKCS#7 routines and add new functions to make it easier to
9486 create PKCS7 structures containing multiple signers. Update smime
9487 application to support multiple signers.
9491 * New -macalg option to pkcs12 utility to allow setting of an alternative
9496 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9497 Reorganize PBE internals to lookup from a static table using NIDs,
9498 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9499 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9500 PRF which will be automatically used with PBES2.
9504 * Replace the algorithm specific calls to generate keys in "req" with the
9509 * Update PKCS#7 enveloped data routines to use new API. This is now
9510 supported by any public key method supporting the encrypt operation. A
9511 ctrl is added to allow the public key algorithm to examine or modify
9512 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9517 * Add a ctrl to asn1 method to allow a public key algorithm to express
9518 a default digest type to use. In most cases this will be SHA1 but some
9519 algorithms (such as GOST) need to specify an alternative digest. The
9520 return value indicates how strong the preference is 1 means optional and
9521 2 is mandatory (that is it is the only supported type). Modify
9522 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9523 use the default md. Update openssl utilities to use the default digest
9524 type for signing if it is not explicitly indicated.
9528 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9529 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9530 signing method from the key type. This effectively removes the link
9531 between digests and public key types.
9535 * Add an OID cross reference table and utility functions. Its purpose is to
9536 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9537 rsaEncryption. This will allow some of the algorithm specific hackery
9538 needed to use the correct OID to be removed.
9542 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9543 structures for PKCS7_sign(). They are now set up by the relevant public
9548 * Add provisional EC pkey method with support for ECDSA and ECDH.
9552 * Add support for key derivation (agreement) in the API, DH method and
9557 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9558 public and private key formats. As a side effect these add additional
9559 command line functionality not previously available: DSA signatures can be
9560 generated and verified using pkeyutl and DH key support and generation in
9567 *Oliver Tappe <zooey@hirschkaefer.de>*
9569 * New make target "install_html_docs" installs HTML renditions of the
9572 *Oliver Tappe <zooey@hirschkaefer.de>*
9574 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9575 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9576 support key and parameter generation and add initial key generation
9577 functionality for RSA.
9581 * Add functions for main EVP_PKEY_method operations. The undocumented
9582 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9583 `EVP_PKEY_{encrypt,decrypt}_old`.
9587 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9588 key API, doesn't do much yet.
9592 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9593 public key algorithms. New option to openssl utility:
9594 "list-public-key-algorithms" to print out info.
9598 * Implement the Supported Elliptic Curves Extension for
9599 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9603 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9604 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9608 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9609 utilities such as rsa, dsa, dsaparam etc except they process any key
9614 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9615 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9616 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9621 * Initial support for pluggable public key ASN1.
9622 De-spaghettify the public key ASN1 handling. Move public and private
9623 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9624 algorithm specific handling to a single module within the relevant
9625 algorithm directory. Add functions to allow (near) opaque processing
9626 of public and private key structures.
9630 * Implement the Supported Point Formats Extension for
9631 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9635 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9636 for the psk identity [hint] and the psk callback functions to the
9637 SSL_SESSION, SSL and SSL_CTX structure.
9640 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9644 SSL_CTX_use_psk_identity_hint
9645 SSL_get_psk_identity_hint
9646 SSL_get_psk_identity
9647 SSL_use_psk_identity_hint
9649 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9651 * Add RFC 3161 compliant time stamp request creation, response generation
9652 and response verification functionality.
9654 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9656 * Add initial support for TLS extensions, specifically for the server_name
9657 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9658 have new members for a hostname. The SSL data structure has an
9659 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9660 stored in that context to allow for session resumption, even after the
9661 SSL has been switched to a new SSL_CTX in reaction to a client's
9662 server_name extension.
9664 New functions (subject to change):
9666 SSL_get_servername()
9667 SSL_get_servername_type()
9670 New CTRL codes and macros (subject to change):
9672 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9673 - SSL_CTX_set_tlsext_servername_callback()
9674 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9675 - SSL_CTX_set_tlsext_servername_arg()
9676 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9678 openssl s_client has a new '-servername ...' option.
9680 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9681 '-key2 ...', '-servername_fatal' (subject to change). This allows
9682 testing the HostName extension for a specific single hostname ('-cert'
9683 and '-key' remain fallbacks for handshakes without HostName
9684 negotiation). If the unrecognized_name alert has to be sent, this by
9685 default is a warning; it becomes fatal with the '-servername_fatal'
9688 *Peter Sylvester, Remy Allais, Christophe Renou*
9690 * Whirlpool hash implementation is added.
9694 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9695 bn(64,32). Because of instruction set limitations it doesn't have
9696 any negative impact on performance. This was done mostly in order
9697 to make it possible to share assembler modules, such as bn_mul_mont
9698 implementations, between 32- and 64-bit builds without hassle.
9702 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9703 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9708 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9709 dedicated Montgomery multiplication procedure, is introduced.
9710 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9711 "64-bit" performance on certain 32-bit targets.
9715 * New option SSL_OP_NO_COMP to disable use of compression selectively
9716 in SSL structures. New SSL ctrl to set maximum send fragment size.
9717 Save memory by setting the I/O buffer sizes dynamically instead of
9718 using the maximum available value.
9722 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9723 in addition to the text details.
9727 * Very, very preliminary EXPERIMENTAL support for printing of general
9728 ASN1 structures. This currently produces rather ugly output and doesn't
9729 handle several customised structures at all.
9733 * Integrated support for PVK file format and some related formats such
9734 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9735 these in the 'rsa' and 'dsa' utilities.
9739 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9743 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9744 place for the (very old) "NETSCAPE" format certificates which are now
9745 handled using new ASN1 code equivalents.
9749 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9750 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9751 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9755 * Modify CRL distribution points extension code to print out previously
9756 unsupported fields. Enhance extension setting code to allow setting of
9761 * Add print and set support for Issuing Distribution Point CRL extension.
9765 * Change 'Configure' script to enable Camellia by default.
9772 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9774 * When rejecting SSL/TLS records due to an incorrect version number, never
9775 update s->server with a new major version number. As of
9776 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9777 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9778 the previous behavior could result in a read attempt at NULL when
9779 receiving specific incorrect SSL/TLS records once record payload
9780 protection is active. ([CVE-2010-0740])
9782 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9784 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9785 could be crashed if the relevant tables were not present (e.g. chrooted).
9787 *Tomas Hoger <thoger@redhat.com>*
9789 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9791 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9793 *Martin Olsson, Neel Mehta*
9795 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9796 accommodate for stack sorting, always a write lock!).
9800 * On some versions of WIN32 Heap32Next is very slow. This can cause
9801 excessive delays in the RAND_poll(): over a minute. As a workaround
9802 include a time check in the inner Heap32Next loop too.
9806 * The code that handled flushing of data in SSL/TLS originally used the
9807 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9808 the problem outlined in PR#1949. The fix suggested there however can
9809 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9810 of Apache). So instead simplify the code to flush unconditionally.
9811 This should be fine since flushing with no data to flush is a no op.
9815 * Handle TLS versions 2.0 and later properly and correctly use the
9816 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9817 off ancient servers have a habit of sticking around for a while...
9821 * Modify compression code so it frees up structures without using the
9822 ex_data callbacks. This works around a problem where some applications
9823 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9824 restarting) then use compression (e.g. SSL with compression) later.
9825 This results in significant per-connection memory leaks and
9826 has caused some security issues including CVE-2008-1678 and
9831 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9832 change when encrypting or decrypting.
9836 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9837 connect and renegotiate with servers which do not support RI.
9838 Until RI is more widely deployed this option is enabled by default.
9842 * Add "missing" ssl ctrls to clear options and mode.
9846 * If client attempts to renegotiate and doesn't support RI respond with
9847 a no_renegotiation alert as required by RFC5746. Some renegotiating
9848 TLS clients will continue a connection gracefully when they receive
9849 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9850 waiting for a server hello which it will never receive. Now we treat a
9851 received no_renegotiation alert as a fatal error. This is because
9852 applications requesting a renegotiation might well expect it to succeed
9853 and would have no code in place to handle the server denying it so the
9854 only safe thing to do is to terminate the connection.
9858 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9859 peer supports secure renegotiation and 0 otherwise. Print out peer
9860 renegotiation support in s_client/s_server.
9864 * Replace the highly broken and deprecated SPKAC certification method with
9865 the updated NID creation version. This should correctly handle UTF8.
9869 * Implement RFC5746. Re-enable renegotiation but require the extension
9870 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9871 turns out to be a bad idea. It has been replaced by
9872 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9873 SSL_CTX_set_options(). This is really not recommended unless you
9874 know what you are doing.
9876 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9878 * Fixes to stateless session resumption handling. Use initial_ctx when
9879 issuing and attempting to decrypt tickets in case it has changed during
9880 servername handling. Use a non-zero length session ID when attempting
9881 stateless session resumption: this makes it possible to determine if
9882 a resumption has occurred immediately after receiving server hello
9883 (several places in OpenSSL subtly assume this) instead of later in
9888 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9889 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9890 fixes for a few places where the return code is not checked
9893 *Julia Lawall <julia@diku.dk>*
9895 * Add --strict-warnings option to Configure script to include devteam
9896 warnings in other configurations.
9900 * Add support for --libdir option and LIBDIR variable in makefiles. This
9901 makes it possible to install openssl libraries in locations which
9902 have names other than "lib", for example "/usr/lib64" which some
9905 *Steve Henson, based on patch from Jeremy Utley*
9907 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9908 X690 8.9.12 and can produce some misleading textual output of OIDs.
9910 *Steve Henson, reported by Dan Kaminsky*
9912 * Delete MD2 from algorithm tables. This follows the recommendation in
9913 several standards that it is not used in new applications due to
9914 several cryptographic weaknesses. For binary compatibility reasons
9915 the MD2 API is still compiled in by default.
9919 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9924 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9925 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9928 *Guenter <lists@gknw.net>*
9930 * Fix the server certificate chain building code to use X509_verify_cert(),
9931 it used to have an ad-hoc builder which was unable to cope with anything
9932 other than a simple chain.
9934 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9936 * Don't check self signed certificate signatures in X509_verify_cert()
9937 by default (a flag can override this): it just wastes time without
9938 adding any security. As a useful side effect self signed root CAs
9939 with non-FIPS digests are now usable in FIPS mode.
9943 * In dtls1_process_out_of_seq_message() the check if the current message
9944 is already buffered was missing. For every new message was memory
9945 allocated, allowing an attacker to perform an denial of service attack
9946 with sending out of seq handshake messages until there is no memory
9947 left. Additionally every future message was buffered, even if the
9948 sequence number made no sense and would be part of another handshake.
9949 So only messages with sequence numbers less than 10 in advance will be
9950 buffered. ([CVE-2009-1378])
9952 *Robin Seggelmann, discovered by Daniel Mentz*
9954 * Records are buffered if they arrive with a future epoch to be
9955 processed after finishing the corresponding handshake. There is
9956 currently no limitation to this buffer allowing an attacker to perform
9957 a DOS attack with sending records with future epochs until there is no
9958 memory left. This patch adds the pqueue_size() function to determine
9959 the size of a buffer and limits the record buffer to 100 entries.
9962 *Robin Seggelmann, discovered by Daniel Mentz*
9964 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9965 parent structure is freed. ([CVE-2009-1379])
9969 * Handle non-blocking I/O properly in SSL_shutdown() call.
9971 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9973 * Add `2.5.4.*` OIDs
9975 *Ilya O. <vrghost@gmail.com>*
9977 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9979 * Disable renegotiation completely - this fixes a severe security
9980 problem ([CVE-2009-3555]) at the cost of breaking all
9981 renegotiation. Renegotiation can be re-enabled by setting
9982 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9983 run-time. This is really not recommended unless you know what
9988 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9990 * Don't set val to NULL when freeing up structures, it is freed up by
9991 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9992 zeroing past the valid field. ([CVE-2009-0789])
9994 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9996 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9997 checked correctly. This would allow some invalid signed attributes to
9998 appear to verify correctly. ([CVE-2009-0591])
10000 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10002 * Reject UniversalString and BMPString types with invalid lengths. This
10003 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10004 a legal length. ([CVE-2009-0590])
10008 * Set S/MIME signing as the default purpose rather than setting it
10009 unconditionally. This allows applications to override it at the store
10014 * Permit restricted recursion of ASN1 strings. This is needed in practice
10015 to handle some structures.
10019 * Improve efficiency of mem_gets: don't search whole buffer each time
10022 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10024 * New -hex option for openssl rand.
10028 * Print out UTF8String and NumericString when parsing ASN1.
10032 * Support NumericString type for name components.
10036 * Allow CC in the environment to override the automatically chosen
10037 compiler. Note that nothing is done to ensure flags work with the
10042 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10044 * Properly check EVP_VerifyFinal() and similar return values
10047 *Ben Laurie, Bodo Moeller, Google Security Team*
10049 * Enable TLS extensions by default.
10053 * Allow the CHIL engine to be loaded, whether the application is
10054 multithreaded or not. (This does not release the developer from the
10055 obligation to set up the dynamic locking callbacks.)
10057 *Sander Temme <sander@temme.net>*
10059 * Use correct exit code if there is an error in dgst command.
10061 *Steve Henson; problem pointed out by Roland Dirlewanger*
10063 * Tweak Configure so that you need to say "experimental-jpake" to enable
10064 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10068 * Add experimental JPAKE support, including demo authentication in
10069 s_client and s_server.
10073 * Set the comparison function in v3_addr_canonize().
10075 *Rob Austein <sra@hactrn.net>*
10077 * Add support for XMPP STARTTLS in s_client.
10079 *Philip Paeps <philip@freebsd.org>*
10081 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10082 to ensure that even with this option, only ciphersuites in the
10083 server's preference list will be accepted. (Note that the option
10084 applies only when resuming a session, so the earlier behavior was
10085 just about the algorithm choice for symmetric cryptography.)
10089 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10091 * Fix NULL pointer dereference if a DTLS server received
10092 ChangeCipherSpec as first record ([CVE-2009-1386]).
10096 * Fix a state transition in s3_srvr.c and d1_srvr.c
10097 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10099 *Nagendra Modadugu*
10101 * The fix in 0.9.8c that supposedly got rid of unsafe
10102 double-checked locking was incomplete for RSA blinding,
10103 addressing just one layer of what turns out to have been
10104 doubly unsafe triple-checked locking.
10106 So now fix this for real by retiring the MONT_HELPER macro
10107 in crypto/rsa/rsa_eay.c.
10109 *Bodo Moeller; problem pointed out by Marius Schilder*
10111 * Various precautionary measures:
10113 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10115 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10116 (NB: This would require knowledge of the secret session ticket key
10117 to exploit, in which case you'd be SOL either way.)
10119 - Change bn_nist.c so that it will properly handle input BIGNUMs
10120 outside the expected range.
10122 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10125 *Neel Mehta, Bodo Moeller*
10127 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10128 the load fails. Useful for distros.
10130 *Ben Laurie and the FreeBSD team*
10132 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10136 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10140 * Expand ENGINE to support engine supplied SSL client certificate functions.
10142 This work was sponsored by Logica.
10146 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10147 keystores. Support for SSL/TLS client authentication too.
10148 Not compiled unless enable-capieng specified to Configure.
10150 This work was sponsored by Logica.
10154 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10155 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10156 attribute creation routines such as certificate requests and PKCS#12
10161 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10163 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10164 handshake which could lead to a client crash as found using the
10165 Codenomicon TLS test suite ([CVE-2008-1672])
10167 *Steve Henson, Mark Cox*
10169 * Fix double free in TLS server name extensions which could lead to
10170 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10174 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10176 Clear the error queue to ensure that error entries left from
10177 older function calls do not interfere with the correct operation.
10179 *Lutz Jaenicke, Erik de Castro Lopo*
10181 * Remove root CA certificates of commercial CAs:
10183 The OpenSSL project does not recommend any specific CA and does not
10184 have any policy with respect to including or excluding any CA.
10185 Therefore, it does not make any sense to ship an arbitrary selection
10186 of root CA certificates with the OpenSSL software.
10190 * RSA OAEP patches to fix two separate invalid memory reads.
10191 The first one involves inputs when 'lzero' is greater than
10192 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10193 before the beginning of from). The second one involves inputs where
10194 the 'db' section contains nothing but zeroes (there is a one-byte
10195 invalid read after the end of 'db').
10197 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10199 * Partial backport from 0.9.9-dev:
10201 Introduce bn_mul_mont (dedicated Montgomery multiplication
10202 procedure) as a candidate for BIGNUM assembler implementation.
10203 While 0.9.9-dev uses assembler for various architectures, only
10204 x86_64 is available by default here in the 0.9.8 branch, and
10205 32-bit x86 is available through a compile-time setting.
10207 To try the 32-bit x86 assembler implementation, use Configure
10208 option "enable-montasm" (which exists only for this backport).
10210 As "enable-montasm" for 32-bit x86 disclaims code stability
10211 anyway, in this constellation we activate additional code
10212 backported from 0.9.9-dev for further performance improvements,
10213 namely BN_from_montgomery_word. (To enable this otherwise,
10214 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10216 *Andy Polyakov (backport partially by Bodo Moeller)*
10218 * Add TLS session ticket callback. This allows an application to set
10219 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10220 values. This is useful for key rollover for example where several key
10221 sets may exist with different names.
10225 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10226 This was broken until now in 0.9.8 releases, such that the only way
10227 a registered ENGINE could be used (assuming it initialises
10228 successfully on the host) was to explicitly set it as the default
10229 for the relevant algorithms. This is in contradiction with 0.9.7
10230 behaviour and the documentation. With this fix, when an ENGINE is
10231 registered into a given algorithm's table of implementations, the
10232 'uptodate' flag is reset so that auto-discovery will be used next
10233 time a new context for that algorithm attempts to select an
10236 *Ian Lister (tweaked by Geoff Thorpe)*
10238 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10239 implementation in the following ways:
10241 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10244 Lack of BER streaming support means one pass streaming processing is
10245 only supported if data is detached: setting the streaming flag is
10246 ignored for embedded content.
10248 CMS support is disabled by default and must be explicitly enabled
10249 with the enable-cms configuration option.
10253 * Update the GMP engine glue to do direct copies between BIGNUM and
10254 mpz_t when openssl and GMP use the same limb size. Otherwise the
10255 existing "conversion via a text string export" trick is still used.
10257 *Paul Sheer <paulsheer@gmail.com>*
10259 * Zlib compression BIO. This is a filter BIO which compressed and
10260 uncompresses any data passed through it.
10264 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10265 RFC3394 compatible AES key wrapping.
10269 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10270 sets string data without copying. X509_ALGOR_set0() and
10271 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10272 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10273 from an X509_ATTRIBUTE structure optionally checking it occurs only
10274 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10279 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10280 to get the expected BN_FLG_CONSTTIME behavior.
10282 *Bodo Moeller (Google)*
10286 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10287 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10288 - added some more tests to do_tests.pl
10289 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10290 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10291 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10292 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10293 - various changes to netware.pl to enable gcc-cross builds on Win32
10295 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10296 - various changes to fix missing prototype warnings
10297 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10298 - added AES, WHIRLPOOL and CPUID assembler code to build files
10299 - added missing AES assembler make rules to mk1mf.pl
10300 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10302 *Guenter Knauf <eflash@gmx.net>*
10304 * Implement certificate status request TLS extension defined in RFC3546.
10305 A client can set the appropriate parameters and receive the encoded
10306 OCSP response via a callback. A server can query the supplied parameters
10307 and set the encoded OCSP response in the callback. Add simplified examples
10308 to s_client and s_server.
10312 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10314 * Fix various bugs:
10315 + Binary incompatibility of ssl_ctx_st structure
10316 + DTLS interoperation with non-compliant servers
10317 + Don't call get_session_cb() without proposed session
10318 + Fix ia64 assembler code
10320 *Andy Polyakov, Steve Henson*
10322 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10324 * DTLS Handshake overhaul. There were longstanding issues with
10325 OpenSSL DTLS implementation, which were making it impossible for
10326 RFC 4347 compliant client to communicate with OpenSSL server.
10327 Unfortunately just fixing these incompatibilities would "cut off"
10328 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10329 server keeps tolerating non RFC compliant syntax. The opposite is
10330 not true, 0.9.8f client can not communicate with earlier server.
10331 This update even addresses CVE-2007-4995.
10335 * Changes to avoid need for function casts in OpenSSL: some compilers
10336 (gcc 4.2 and later) reject their use.
10337 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10340 * Add RFC4507 support to OpenSSL. This includes the corrections in
10341 RFC4507bis. The encrypted ticket format is an encrypted encoded
10342 SSL_SESSION structure, that way new session features are automatically
10345 If a client application caches session in an SSL_SESSION structure
10346 support is transparent because tickets are now stored in the encoded
10349 The SSL_CTX structure automatically generates keys for ticket
10350 protection in servers so again support should be possible
10351 with no application modification.
10353 If a client or server wishes to disable RFC4507 support then the option
10354 SSL_OP_NO_TICKET can be set.
10356 Add a TLS extension debugging callback to allow the contents of any client
10357 or server extensions to be examined.
10359 This work was sponsored by Google.
10363 * Add initial support for TLS extensions, specifically for the server_name
10364 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10365 have new members for a hostname. The SSL data structure has an
10366 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10367 stored in that context to allow for session resumption, even after the
10368 SSL has been switched to a new SSL_CTX in reaction to a client's
10369 server_name extension.
10371 New functions (subject to change):
10373 SSL_get_servername()
10374 SSL_get_servername_type()
10377 New CTRL codes and macros (subject to change):
10379 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10380 - SSL_CTX_set_tlsext_servername_callback()
10381 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10382 - SSL_CTX_set_tlsext_servername_arg()
10383 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10385 openssl s_client has a new '-servername ...' option.
10387 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10388 '-key2 ...', '-servername_fatal' (subject to change). This allows
10389 testing the HostName extension for a specific single hostname ('-cert'
10390 and '-key' remain fallbacks for handshakes without HostName
10391 negotiation). If the unrecognized_name alert has to be sent, this by
10392 default is a warning; it becomes fatal with the '-servername_fatal'
10395 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10397 * Add AES and SSE2 assembly language support to VC++ build.
10401 * Mitigate attack on final subtraction in Montgomery reduction.
10405 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10406 (which previously caused an internal error).
10410 * Squeeze another 10% out of IGE mode when in != out.
10414 * AES IGE mode speedup.
10416 *Dean Gaudet (Google)*
10418 * Add the Korean symmetric 128-bit cipher SEED (see
10419 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10420 add SEED ciphersuites from RFC 4162:
10422 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10423 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10424 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10425 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10427 To minimize changes between patchlevels in the OpenSSL 0.9.8
10428 series, SEED remains excluded from compilation unless OpenSSL
10429 is configured with 'enable-seed'.
10431 *KISA, Bodo Moeller*
10433 * Mitigate branch prediction attacks, which can be practical if a
10434 single processor is shared, allowing a spy process to extract
10435 information. For detailed background information, see
10436 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10437 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10438 and Necessary Software Countermeasures"). The core of the change
10439 are new versions BN_div_no_branch() and
10440 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10441 respectively, which are slower, but avoid the security-relevant
10442 conditional branches. These are automatically called by BN_div()
10443 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10444 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10445 remove a conditional branch.
10447 BN_FLG_CONSTTIME is the new name for the previous
10448 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10449 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10450 in the exponent causes BN_mod_exp_mont() to use the alternative
10451 implementation in BN_mod_exp_mont_consttime().) The old name
10452 remains as a deprecated alias.
10454 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10455 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10456 constant-time implementations for more than just exponentiation.
10457 Here too the old name is kept as a deprecated alias.
10459 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10460 the BN_BLINDING structure gets an independent copy of the
10461 modulus. This means that the previous `BIGNUM *m` argument to
10462 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10463 essentially becomes `const BIGNUM *m`, although we can't actually
10464 change this in the header file before 0.9.9. It allows
10465 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10466 enable BN_FLG_CONSTTIME.
10468 *Matthew D Wood (Intel Corp)*
10470 * In the SSL/TLS server implementation, be strict about session ID
10471 context matching (which matters if an application uses a single
10472 external cache for different purposes). Previously,
10473 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10474 set. This did ensure strict client verification, but meant that,
10475 with applications using a single external cache for quite
10476 different requirements, clients could circumvent ciphersuite
10477 restrictions for a given session ID context by starting a session
10478 in a different context.
10482 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10483 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10484 authentication-only ciphersuites.
10488 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10489 not complete and could lead to a possible single byte overflow
10490 ([CVE-2007-5135]) [Ben Laurie]
10492 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10494 * Since AES128 and AES256 (and similarly Camellia128 and
10495 Camellia256) share a single mask bit in the logic of
10496 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10497 kludge to work properly if AES128 is available and AES256 isn't
10498 (or if Camellia128 is available and Camellia256 isn't).
10502 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10503 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10504 When a point or a seed is encoded in a BIT STRING, we need to
10505 prevent the removal of trailing zero bits to get the proper DER
10506 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10507 of a NamedBitList, for which trailing 0 bits need to be removed.)
10511 * Have SSL/TLS server implementation tolerate "mismatched" record
10512 protocol version while receiving ClientHello even if the
10513 ClientHello is fragmented. (The server can't insist on the
10514 particular protocol version it has chosen before the ServerHello
10515 message has informed the client about his choice.)
10519 * Add RFC 3779 support.
10521 *Rob Austein for ARIN, Ben Laurie*
10523 * Load error codes if they are not already present instead of using a
10524 static variable. This allows them to be cleanly unloaded and reloaded.
10525 Improve header file function name parsing.
10529 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10530 or CAPABILITY handshake as required by RFCs.
10532 *Goetz Babin-Ebell*
10534 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10536 * Introduce limits to prevent malicious keys being able to
10537 cause a denial of service. ([CVE-2006-2940])
10539 *Steve Henson, Bodo Moeller*
10541 * Fix ASN.1 parsing of certain invalid structures that can result
10542 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10544 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10545 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10547 * Fix SSL client code which could crash if connecting to a
10548 malicious SSLv2 server. ([CVE-2006-4343])
10550 *Tavis Ormandy and Will Drewry, Google Security Team*
10552 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10553 match only those. Before that, "AES256-SHA" would be interpreted
10554 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10555 the same strength classification in 0.9.7h) as we currently only
10556 have a single AES bit in the ciphersuite description bitmap.
10557 That change, however, also applied to ciphersuite strings such as
10558 "RC4-MD5" that intentionally matched multiple ciphersuites --
10559 namely, SSL 2.0 ciphersuites in addition to the more common ones
10560 from SSL 3.0/TLS 1.0.
10562 So we change the selection algorithm again: Naming an explicit
10563 ciphersuite selects this one ciphersuite, and any other similar
10564 ciphersuite (same bitmap) from *other* protocol versions.
10565 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10566 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10568 Since SSL 2.0 does not have any ciphersuites for which the
10569 128/256 bit distinction would be relevant, this works for now.
10570 The proper fix will be to use different bits for AES128 and
10571 AES256, which would have avoided the problems from the beginning;
10572 however, bits are scarce, so we can only do this in a new release
10573 (not just a patchlevel) when we can change the SSL_CIPHER
10574 definition to split the single 'unsigned long mask' bitmap into
10575 multiple values to extend the available space.
10579 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10581 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10582 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10584 * Add AES IGE and biIGE modes.
10588 * Change the Unix randomness entropy gathering to use poll() when
10589 possible instead of select(), since the latter has some
10590 undesirable limitations.
10592 *Darryl Miles via Richard Levitte and Bodo Moeller*
10594 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10595 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10596 cannot be implicitly activated as part of, e.g., the "AES" alias.
10597 However, please upgrade to OpenSSL 0.9.9[-dev] for
10598 non-experimental use of the ECC ciphersuites to get TLS extension
10599 support, which is required for curve and point format negotiation
10600 to avoid potential handshake problems.
10604 * Disable rogue ciphersuites:
10606 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10607 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10608 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10610 The latter two were purportedly from
10611 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10614 Also deactivate the remaining ciphersuites from
10615 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10616 unofficial, and the ID has long expired.
10620 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10621 dual-core machines) and other potential thread-safety issues.
10625 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10626 versions), which is now available for royalty-free use
10627 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10628 Also, add Camellia TLS ciphersuites from RFC 4132.
10630 To minimize changes between patchlevels in the OpenSSL 0.9.8
10631 series, Camellia remains excluded from compilation unless OpenSSL
10632 is configured with 'enable-camellia'.
10636 * Disable the padding bug check when compression is in use. The padding
10637 bug check assumes the first packet is of even length, this is not
10638 necessarily true if compression is enabled and can result in false
10639 positives causing handshake failure. The actual bug test is ancient
10640 code so it is hoped that implementations will either have fixed it by
10641 now or any which still have the bug do not support compression.
10645 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10647 * When applying a cipher rule check to see if string match is an explicit
10648 cipher suite and only match that one cipher suite if it is.
10652 * Link in manifests for VC++ if needed.
10654 *Austin Ziegler <halostatue@gmail.com>*
10656 * Update support for ECC-based TLS ciphersuites according to
10657 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10658 TLS extensions, which are supported starting with the 0.9.9
10659 branch, not in the OpenSSL 0.9.8 branch).
10663 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10664 opaque EVP_CIPHER_CTX handling.
10668 * Fixes and enhancements to zlib compression code. We now only use
10669 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10670 to conform with the standards mentioned here:
10671 <http://www.zlib.net/DLL_FAQ.txt>
10672 Static zlib linking now works on Windows and the new --with-zlib-include
10673 --with-zlib-lib options to Configure can be used to supply the location
10674 of the headers and library. Gracefully handle case where zlib library
10679 * Several fixes and enhancements to the OID generation code. The old code
10680 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10681 handle numbers larger than ULONG_MAX, truncated printing and had a
10682 non standard OBJ_obj2txt() behaviour.
10686 * Add support for building of engines under engine/ as shared libraries
10687 under VC++ build system.
10691 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10692 Hopefully, we will not see any false combination of paths any more.
10696 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10698 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10699 (part of SSL_OP_ALL). This option used to disable the
10700 countermeasure against man-in-the-middle protocol-version
10701 rollback in the SSL 2.0 server implementation, which is a bad
10702 idea. ([CVE-2005-2969])
10704 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10705 for Information Security, National Institute of Advanced Industrial
10706 Science and Technology [AIST], Japan)*
10708 * Add two function to clear and return the verify parameter flags.
10712 * Keep cipherlists sorted in the source instead of sorting them at
10713 runtime, thus removing the need for a lock.
10717 * Avoid some small subgroup attacks in Diffie-Hellman.
10719 *Nick Mathewson and Ben Laurie*
10721 * Add functions for well-known primes.
10725 * Extended Windows CE support.
10727 *Satoshi Nakamura and Andy Polyakov*
10729 * Initialize SSL_METHOD structures at compile time instead of during
10730 runtime, thus removing the need for a lock.
10734 * Make PKCS7_decrypt() work even if no certificate is supplied by
10735 attempting to decrypt each encrypted key in turn. Add support to
10740 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10742 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10745 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10749 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10750 key into the same file any more.
10754 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10758 * Add -utf8 command line and config file option to 'ca'.
10760 *Stefan <stf@udoma.org*
10762 * Removed the macro des_crypt(), as it seems to conflict with some
10763 libraries. Use DES_crypt().
10767 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10768 involves renaming the source and generated shared-libs for
10769 both. The engines will accept the corrected or legacy ids
10770 ('ncipher' and '4758_cca' respectively) when binding. NB,
10771 this only applies when building 'shared'.
10773 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10775 * Add attribute functions to EVP_PKEY structure. Modify
10776 PKCS12_create() to recognize a CSP name attribute and
10777 use it. Make -CSP option work again in pkcs12 utility.
10781 * Add new functionality to the bn blinding code:
10782 - automatic re-creation of the BN_BLINDING parameters after
10783 a fixed number of uses (currently 32)
10784 - add new function for parameter creation
10785 - introduce flags to control the update behaviour of the
10786 BN_BLINDING parameters
10787 - hide BN_BLINDING structure
10788 Add a second BN_BLINDING slot to the RSA structure to improve
10789 performance when a single RSA object is shared among several
10794 * Add support for DTLS.
10796 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10798 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10799 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10803 * Remove buggy and incomplete DH cert support from
10804 ssl/ssl_rsa.c and ssl/s3_both.c
10808 * Use SHA-1 instead of MD5 as the default digest algorithm for
10809 the `apps/openssl` commands.
10813 * Compile clean with "-Wall -Wmissing-prototypes
10814 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10815 DEBUG_SAFESTACK must also be set.
10819 * Change ./Configure so that certain algorithms can be disabled by default.
10820 The new counterpiece to "no-xxx" is "enable-xxx".
10822 The patented RC5 and MDC2 algorithms will now be disabled unless
10823 "enable-rc5" and "enable-mdc2", respectively, are specified.
10825 (IDEA remains enabled despite being patented. This is because IDEA
10826 is frequently required for interoperability, and there is no license
10827 fee for non-commercial use. As before, "no-idea" can be used to
10828 avoid this algorithm.)
10832 * Add processing of proxy certificates (see RFC 3820). This work was
10833 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10834 EGEE (Enabling Grids for E-science in Europe).
10838 * RC4 performance overhaul on modern architectures/implementations, such
10839 as Intel P4, IA-64 and AMD64.
10843 * New utility extract-section.pl. This can be used specify an alternative
10844 section number in a pod file instead of having to treat each file as
10845 a separate case in Makefile. This can be done by adding two lines to the
10848 =for comment openssl_section:XXX
10850 The blank line is mandatory.
10854 * New arguments -certform, -keyform and -pass for s_client and s_server
10855 to allow alternative format key and certificate files and passphrase
10860 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10861 update associated structures and add various utility functions.
10863 Add new policy related verify parameters, include policy checking in
10864 standard verify code. Enhance 'smime' application with extra parameters
10865 to support policy checking and print out.
10869 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10870 Nehemiah processors. These extensions support AES encryption in hardware
10871 as well as RNG (though RNG support is currently disabled).
10873 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10875 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10879 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10881 *Andy Polyakov and a number of other people*
10883 * Improved PowerPC platform support. Most notably BIGNUM assembler
10884 implementation contributed by IBM.
10886 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10888 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10889 exponent rather than 'unsigned long'. There is a corresponding change to
10890 the new 'rsa_keygen' element of the RSA_METHOD structure.
10892 *Jelte Jansen, Geoff Thorpe*
10894 * Functionality for creating the initial serial number file is now
10895 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10897 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10898 number file to 1, which is bound to cause problems. To avoid
10899 the problems while respecting compatibility between different 0.9.7
10900 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10901 CA.pl for serial number initialization. With the new release 0.9.8,
10902 we can fix the problem directly in the 'ca' utility.)
10906 * Reduced header interdependencies by declaring more opaque objects in
10907 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10908 give fewer recursive includes, which could break lazy source code - so
10909 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10910 developers should define this symbol when building and using openssl to
10911 ensure they track the recommended behaviour, interfaces, [etc], but
10912 backwards-compatible behaviour prevails when this isn't defined.
10916 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10920 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10921 This will generate a random key of the appropriate length based on the
10922 cipher context. The EVP_CIPHER can provide its own random key generation
10923 routine to support keys of a specific form. This is used in the des and
10924 3des routines to generate a key of the correct parity. Update S/MIME
10925 code to use new functions and hence generate correct parity DES keys.
10926 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10927 valid (weak or incorrect parity).
10931 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10932 as looking them up. This is useful when the verified structure may contain
10933 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10934 present unless the new PKCS7_NO_CRL flag is asserted.
10938 * Extend ASN1 oid configuration module. It now additionally accepts the
10941 shortName = some long name, 1.2.3.4
10945 * Reimplemented the BN_CTX implementation. There is now no more static
10946 limitation on the number of variables it can handle nor the depth of the
10947 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10948 information can now expand as required, and rather than having a single
10949 static array of bignums, BN_CTX now uses a linked-list of such arrays
10950 allowing it to expand on demand whilst maintaining the usefulness of
10951 BN_CTX's "bundling".
10955 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10956 to allow all RSA operations to function using a single BN_CTX.
10960 * Preliminary support for certificate policy evaluation and checking. This
10961 is initially intended to pass the tests outlined in "Conformance Testing
10962 of Relying Party Client Certificate Path Processing Logic" v1.07.
10966 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10967 remained unused and not that useful. A variety of other little bignum
10968 tweaks and fixes have also been made continuing on from the audit (see
10973 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10974 associated ASN1, EVP and SSL functions and old ASN1 macros.
10978 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10979 and this should never fail. So the return value from the use of
10980 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10981 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10985 * BN_CTX_get() should return zero-valued bignums, providing the same
10986 initialised value as BN_new().
10988 *Geoff Thorpe, suggested by Ulf Möller*
10990 * Support for inhibitAnyPolicy certificate extension.
10994 * An audit of the BIGNUM code is underway, for which debugging code is
10995 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10996 is considered valid when processing BIGNUMs, and causes execution to
10997 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10998 further steps are taken to deliberately pollute unused data in BIGNUM
10999 structures to try and expose faulty code further on. For now, openssl will
11000 (in its default mode of operation) continue to tolerate the inconsistent
11001 forms that it has tolerated in the past, but authors and packagers should
11002 consider trying openssl and their own applications when compiled with
11003 these debugging symbols defined. It will help highlight potential bugs in
11004 their own code, and will improve the test coverage for OpenSSL itself. At
11005 some point, these tighter rules will become openssl's default to improve
11006 maintainability, though the assert()s and other overheads will remain only
11007 in debugging configurations. See bn.h for more details.
11009 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11011 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11012 that can only be obtained through BN_CTX_new() (which implicitly
11013 initialises it). The presence of this function only made it possible
11014 to overwrite an existing structure (and cause memory leaks).
11018 * Because of the callback-based approach for implementing LHASH as a
11019 template type, lh_insert() adds opaque objects to hash-tables and
11020 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11021 to clean up those corresponding objects before destroying the hash table
11022 (and losing the object pointers). So some over-zealous constifications in
11023 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11024 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11025 prototyped to have "const" restrictions on the object pointers they are
11026 given (and so aren't required to cast them away any more).
11030 * The tmdiff.h API was so ugly and minimal that our own timing utility
11031 (speed) prefers to use its own implementation. The two implementations
11032 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11033 its object type properly exposed (MS_TM) instead of casting to/from
11034 `char *`. This may still change yet if someone realises MS_TM and
11036 aren't necessarily the greatest nomenclatures - but this is what was used
11037 internally to the implementation so I've used that for now.
11041 * Ensure that deprecated functions do not get compiled when
11042 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11043 the self-tests were still using deprecated key-generation functions so
11044 these have been updated also.
11048 * Reorganise PKCS#7 code to separate the digest location functionality
11049 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11050 New function PKCS7_set_digest() to set the digest type for PKCS#7
11051 digestedData type. Add additional code to correctly generate the
11052 digestedData type and add support for this type in PKCS7 initialization
11057 * New function PKCS7_set0_type_other() this initializes a PKCS7
11058 structure of type "other".
11062 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11063 sure the loop does correctly stop and breaking ("division by zero")
11064 modulus operations are not performed. The (pre-generated) prime
11065 table crypto/bn/bn_prime.h was already correct, but it could not be
11066 re-generated on some platforms because of the "division by zero"
11067 situation in the script.
11069 *Ralf S. Engelschall*
11071 * Update support for ECC-based TLS ciphersuites according to
11072 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11073 SHA-1 now is only used for "small" curves (where the
11074 representation of a field element takes up to 24 bytes); for
11075 larger curves, the field element resulting from ECDH is directly
11076 used as premaster secret.
11078 *Douglas Stebila (Sun Microsystems Laboratories)*
11080 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11081 curve secp160r1 to the tests.
11083 *Douglas Stebila (Sun Microsystems Laboratories)*
11085 * Add the possibility to load symbols globally with DSO.
11087 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11089 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11090 control of the error stack.
11094 * Add support for STORE in ENGINE.
11098 * Add the STORE type. The intention is to provide a common interface
11099 to certificate and key stores, be they simple file-based stores, or
11100 HSM-type store, or LDAP stores, or...
11101 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11105 * Add a generic structure called OPENSSL_ITEM. This can be used to
11106 pass a list of arguments to any function as well as provide a way
11107 for a function to pass data back to the caller.
11111 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11112 works like BUF_strdup() but can be used to duplicate a portion of
11113 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11118 * Add the function sk_find_ex() which works like sk_find(), but will
11119 return an index to an element even if an exact match couldn't be
11120 found. The index is guaranteed to point at the element where the
11121 searched-for key would be inserted to preserve sorting order.
11125 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11126 takes an extra flags argument for optional functionality. Currently,
11127 the following flags are defined:
11129 OBJ_BSEARCH_VALUE_ON_NOMATCH
11130 This one gets OBJ_bsearch_ex() to return a pointer to the first
11131 element where the comparing function returns a negative or zero
11134 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11135 This one gets OBJ_bsearch_ex() to return a pointer to the first
11136 element where the comparing function returns zero. This is useful
11137 if there are more than one element where the comparing function
11142 * Make it possible to create self-signed certificates with 'openssl ca'
11143 in such a way that the self-signed certificate becomes part of the
11144 CA database and uses the same mechanisms for serial number generation
11145 as all other certificate signing. The new flag '-selfsign' enables
11146 this functionality. Adapt CA.sh and CA.pl.in.
11150 * Add functionality to check the public key of a certificate request
11151 against a given private. This is useful to check that a certificate
11152 request can be signed by that key (self-signing).
11156 * Make it possible to have multiple active certificates with the same
11157 subject in the CA index file. This is done only if the keyword
11158 'unique_subject' is set to 'no' in the main CA section (default
11159 if 'CA_default') of the configuration file. The value is saved
11160 with the database itself in a separate index attribute file,
11161 named like the index file with '.attr' appended to the name.
11165 * Generate multi-valued AVAs using '+' notation in config files for
11170 * Support for nameConstraints certificate extension.
11174 * Support for policyConstraints certificate extension.
11178 * Support for policyMappings certificate extension.
11182 * Make sure the default DSA_METHOD implementation only uses its
11183 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11184 and change its own handlers to be NULL so as to remove unnecessary
11185 indirection. This lets alternative implementations fallback to the
11186 default implementation more easily.
11190 * Support for directoryName in GeneralName related extensions
11195 * Make it possible to link applications using Makefile.shared.
11196 Make that possible even when linking against static libraries!
11200 * Support for single pass processing for S/MIME signing. This now
11201 means that S/MIME signing can be done from a pipe, in addition
11202 cleartext signing (multipart/signed type) is effectively streaming
11203 and the signed data does not need to be all held in memory.
11205 This is done with a new flag PKCS7_STREAM. When this flag is set
11206 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11207 is done after the data is output (and digests calculated) in
11208 SMIME_write_PKCS7().
11212 * Add full support for -rpath/-R, both in shared libraries and
11213 applications, at least on the platforms where it's known how
11218 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11219 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11220 will now compute a table of multiples of the generator that
11221 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11222 faster (notably in the case of a single point multiplication,
11223 scalar * generator).
11225 *Nils Larsch, Bodo Moeller*
11227 * IPv6 support for certificate extensions. The various extensions
11228 which use the IP:a.b.c.d can now take IPv6 addresses using the
11229 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11234 * Added an ENGINE that implements RSA by performing private key
11235 exponentiations with the GMP library. The conversions to and from
11236 GMP's mpz_t format aren't optimised nor are any montgomery forms
11237 cached, and on x86 it appears OpenSSL's own performance has caught up.
11238 However there are likely to be other architectures where GMP could
11239 provide a boost. This ENGINE is not built in by default, but it can be
11240 specified at Configure time and should be accompanied by the necessary
11241 linker additions, eg;
11242 ./config -DOPENSSL_USE_GMP -lgmp
11246 * "openssl engine" will not display ENGINE/DSO load failure errors when
11247 testing availability of engines with "-t" - the old behaviour is
11248 produced by increasing the feature's verbosity with "-tt".
11252 * ECDSA routines: under certain error conditions uninitialized BN objects
11253 could be freed. Solution: make sure initialization is performed early
11254 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11259 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11260 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11261 software implementations. For DSA and DH, parameter generation can
11262 also be overridden by providing the appropriate method callbacks.
11266 * Change the "progress" mechanism used in key-generation and
11267 primality testing to functions that take a new BN_GENCB pointer in
11268 place of callback/argument pairs. The new API functions have `_ex`
11269 postfixes and the older functions are reimplemented as wrappers for
11270 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11271 declarations of the old functions to help (graceful) attempts to
11272 migrate to the new functions. Also, the new key-generation API
11273 functions operate on a caller-supplied key-structure and return
11274 success/failure rather than returning a key or NULL - this is to
11275 help make "keygen" another member function of RSA_METHOD etc.
11277 Example for using the new callback interface:
11279 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11280 void *my_arg = ...;
11283 BN_GENCB_set(&my_cb, my_callback, my_arg);
11285 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11286 /* For the meaning of a, b in calls to my_callback(), see the
11287 * documentation of the function that calls the callback.
11288 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11289 * my_callback should return 1 if it wants BN_is_prime_ex()
11290 * to continue, or 0 to stop.
11295 * Change the ZLIB compression method to be stateful, and make it
11296 available to TLS with the number defined in
11297 draft-ietf-tls-compression-04.txt.
11301 * Add the ASN.1 structures and functions for CertificatePair, which
11302 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11304 CertificatePair ::= SEQUENCE {
11305 forward [0] Certificate OPTIONAL,
11306 reverse [1] Certificate OPTIONAL,
11307 -- at least one of the pair shall be present -- }
11309 Also implement the PEM functions to read and write certificate
11310 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11312 This needed to be defined, mostly for the sake of the LDAP
11313 attribute crossCertificatePair, but may prove useful elsewhere as
11318 * Make it possible to inhibit symlinking of shared libraries in
11319 Makefile.shared, for Cygwin's sake.
11323 * Extend the BIGNUM API by creating a function
11324 void BN_set_negative(BIGNUM *a, int neg);
11325 and a macro that behave like
11326 int BN_is_negative(const BIGNUM *a);
11328 to avoid the need to access 'a->neg' directly in applications.
11332 * Implement fast modular reduction for pseudo-Mersenne primes
11333 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11334 EC_GROUP_new_curve_GFp() will now automatically use this
11337 *Nils Larsch <nla@trustcenter.de>*
11339 * Add new lock type (CRYPTO_LOCK_BN).
11343 * Change the ENGINE framework to automatically load engines
11344 dynamically from specific directories unless they could be
11345 found to already be built in or loaded. Move all the
11346 current engines except for the cryptodev one to a new
11347 directory engines/.
11348 The engines in engines/ are built as shared libraries if
11349 the "shared" options was given to ./Configure or ./config.
11350 Otherwise, they are inserted in libcrypto.a.
11351 /usr/local/ssl/engines is the default directory for dynamic
11352 engines, but that can be overridden at configure time through
11353 the usual use of --prefix and/or --openssldir, and at run
11354 time with the environment variable OPENSSL_ENGINES.
11356 *Geoff Thorpe and Richard Levitte*
11358 * Add Makefile.shared, a helper makefile to build shared
11359 libraries. Adapt Makefile.org.
11363 * Add version info to Win32 DLLs.
11365 *Peter 'Luna' Runestig" <peter@runestig.com>*
11367 * Add new 'medium level' PKCS#12 API. Certificates and keys
11368 can be added using this API to created arbitrary PKCS#12
11369 files while avoiding the low-level API.
11371 New options to PKCS12_create(), key or cert can be NULL and
11372 will then be omitted from the output file. The encryption
11373 algorithm NIDs can be set to -1 for no encryption, the mac
11374 iteration count can be set to 0 to omit the mac.
11376 Enhance pkcs12 utility by making the -nokeys and -nocerts
11377 options work when creating a PKCS#12 file. New option -nomac
11378 to omit the mac, NONE can be set for an encryption algorithm.
11379 New code is modified to use the enhanced PKCS12_create()
11380 instead of the low-level API.
11384 * Extend ASN1 encoder to support indefinite length constructed
11385 encoding. This can output sequences tags and octet strings in
11386 this form. Modify pk7_asn1.c to support indefinite length
11387 encoding. This is experimental and needs additional code to
11388 be useful, such as an ASN1 bio and some enhanced streaming
11391 Extend template encode functionality so that tagging is passed
11392 down to the template encoder.
11396 * Let 'openssl req' fail if an argument to '-newkey' is not
11397 recognized instead of using RSA as a default.
11401 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11402 As these are not official, they are not included in "ALL";
11403 the "ECCdraft" ciphersuite group alias can be used to select them.
11405 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11407 * Add ECDH engine support.
11409 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11411 * Add ECDH in new directory crypto/ecdh/.
11413 *Douglas Stebila (Sun Microsystems Laboratories)*
11415 * Let BN_rand_range() abort with an error after 100 iterations
11416 without success (which indicates a broken PRNG).
11420 * Change BN_mod_sqrt() so that it verifies that the input value
11421 is really the square of the return value. (Previously,
11422 BN_mod_sqrt would show GIGO behaviour.)
11426 * Add named elliptic curves over binary fields from X9.62, SECG,
11427 and WAP/WTLS; add OIDs that were still missing.
11429 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11431 * Extend the EC library for elliptic curves over binary fields
11432 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11435 EC_GF2m_simple_method
11439 EC_GROUP_new_curve_GF2m
11440 EC_GROUP_set_curve_GF2m
11441 EC_GROUP_get_curve_GF2m
11442 EC_POINT_set_affine_coordinates_GF2m
11443 EC_POINT_get_affine_coordinates_GF2m
11444 EC_POINT_set_compressed_coordinates_GF2m
11446 Point compression for binary fields is disabled by default for
11447 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11450 As binary polynomials are represented as BIGNUMs, various members
11451 of the EC_GROUP and EC_POINT data structures can be shared
11452 between the implementations for prime fields and binary fields;
11453 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11454 are essentially identical to their `..._GFp` counterparts.
11455 (For simplicity, the `..._GFp` prefix has been dropped from
11456 various internal method names.)
11458 An internal 'field_div' method (similar to 'field_mul' and
11459 'field_sqr') has been added; this is used only for binary fields.
11461 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11463 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11464 through methods ('mul', 'precompute_mult').
11466 The generic implementations (now internally called 'ec_wNAF_mul'
11467 and 'ec_wNAF_precomputed_mult') remain the default if these
11468 methods are undefined.
11470 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11472 * New function EC_GROUP_get_degree, which is defined through
11473 EC_METHOD. For curves over prime fields, this returns the bit
11474 length of the modulus.
11476 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11478 * New functions EC_GROUP_dup, EC_POINT_dup.
11479 (These simply call ..._new and ..._copy).
11481 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11483 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11484 Polynomials are represented as BIGNUMs (where the sign bit is not
11485 used) in the following functions [macros]:
11488 BN_GF2m_sub [= BN_GF2m_add]
11489 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11490 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11491 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11493 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11494 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11495 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11496 BN_GF2m_cmp [= BN_ucmp]
11498 (Note that only the 'mod' functions are actually for fields GF(2^m).
11499 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11501 For some functions, an the irreducible polynomial defining a
11502 field can be given as an 'unsigned int[]' with strictly
11503 decreasing elements giving the indices of those bits that are set;
11504 i.e., p[] represents the polynomial
11505 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11507 p[0] > p[1] > ... > p[k] = 0.
11508 This applies to the following functions:
11511 BN_GF2m_mod_mul_arr
11512 BN_GF2m_mod_sqr_arr
11513 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11514 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11515 BN_GF2m_mod_exp_arr
11516 BN_GF2m_mod_sqrt_arr
11517 BN_GF2m_mod_solve_quad_arr
11521 Conversion can be performed by the following functions:
11526 bntest.c has additional tests for binary polynomial arithmetic.
11528 Two implementations for BN_GF2m_mod_div() are available.
11529 The default algorithm simply uses BN_GF2m_mod_inv() and
11530 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11531 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11532 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11534 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11536 * Add new error code 'ERR_R_DISABLED' that can be used when some
11537 functionality is disabled at compile-time.
11539 *Douglas Stebila <douglas.stebila@sun.com>*
11541 * Change default behaviour of 'openssl asn1parse' so that more
11542 information is visible when viewing, e.g., a certificate:
11544 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11545 mode the content of non-printable OCTET STRINGs is output in a
11546 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11547 avoid the appearance of a printable string.
11549 *Nils Larsch <nla@trustcenter.de>*
11551 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11553 EC_GROUP_set_asn1_flag()
11554 EC_GROUP_get_asn1_flag()
11555 EC_GROUP_set_point_conversion_form()
11556 EC_GROUP_get_point_conversion_form()
11557 These control ASN1 encoding details:
11558 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11559 has been set to OPENSSL_EC_NAMED_CURVE.
11560 - Points are encoded in uncompressed form by default; options for
11561 asn1_for are as for point2oct, namely
11562 POINT_CONVERSION_COMPRESSED
11563 POINT_CONVERSION_UNCOMPRESSED
11564 POINT_CONVERSION_HYBRID
11566 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11568 EC_GROUP_set_seed()
11569 EC_GROUP_get0_seed()
11570 EC_GROUP_get_seed_len()
11571 This is used only for ASN1 purposes (so far).
11573 *Nils Larsch <nla@trustcenter.de>*
11575 * Add 'field_type' member to EC_METHOD, which holds the NID
11576 of the appropriate field type OID. The new function
11577 EC_METHOD_get_field_type() returns this value.
11579 *Nils Larsch <nla@trustcenter.de>*
11582 EC_POINT_point2bn()
11583 EC_POINT_bn2point()
11584 EC_POINT_point2hex()
11585 EC_POINT_hex2point()
11586 providing useful interfaces to EC_POINT_point2oct() and
11587 EC_POINT_oct2point().
11589 *Nils Larsch <nla@trustcenter.de>*
11591 * Change internals of the EC library so that the functions
11592 EC_GROUP_set_generator()
11593 EC_GROUP_get_generator()
11594 EC_GROUP_get_order()
11595 EC_GROUP_get_cofactor()
11596 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11597 to methods, which would lead to unnecessary code duplication when
11598 adding different types of curves.
11600 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11602 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11603 arithmetic, and such that modified wNAFs are generated
11604 (which avoid length expansion in many cases).
11608 * Add a function EC_GROUP_check_discriminant() (defined via
11609 EC_METHOD) that verifies that the curve discriminant is non-zero.
11611 Add a function EC_GROUP_check() that makes some sanity tests
11612 on a EC_GROUP, its generator and order. This includes
11613 EC_GROUP_check_discriminant().
11615 *Nils Larsch <nla@trustcenter.de>*
11617 * Add ECDSA in new directory crypto/ecdsa/.
11619 Add applications 'openssl ecparam' and 'openssl ecdsa'
11620 (these are based on 'openssl dsaparam' and 'openssl dsa').
11622 ECDSA support is also included in various other files across the
11623 library. Most notably,
11624 - 'openssl req' now has a '-newkey ecdsa:file' option;
11625 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11626 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11627 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11628 them suitable for ECDSA where domain parameters must be
11629 extracted before the specific public key;
11630 - ECDSA engine support has been added.
11632 *Nils Larsch <nla@trustcenter.de>*
11634 * Include some named elliptic curves, and add OIDs from X9.62,
11635 SECG, and WAP/WTLS. Each curve can be obtained from the new
11637 EC_GROUP_new_by_curve_name(),
11638 and the list of available named curves can be obtained with
11639 EC_get_builtin_curves().
11640 Also add a 'curve_name' member to EC_GROUP objects, which can be
11642 EC_GROUP_set_curve_name()
11643 EC_GROUP_get_curve_name()
11645 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11647 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11648 was actually never needed) and in BN_mul(). The removal in BN_mul()
11649 required a small change in bn_mul_part_recursive() and the addition
11650 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11651 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11652 bn_sub_words() and bn_add_words() except they take arrays with
11657 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11659 * Cleanse PEM buffers before freeing them since they may contain
11662 *Benjamin Bennett <ben@psc.edu>*
11664 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11665 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11666 authentication-only ciphersuites.
11670 * Since AES128 and AES256 share a single mask bit in the logic of
11671 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11672 kludge to work properly if AES128 is available and AES256 isn't.
11676 * Expand security boundary to match 1.1.1 module.
11680 * Remove redundant features: hash file source, editing of test vectors
11681 modify fipsld to use external fips_premain.c signature.
11685 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11686 run algorithm test programs.
11690 * Make algorithm test programs more tolerant of whitespace.
11694 * Have SSL/TLS server implementation tolerate "mismatched" record
11695 protocol version while receiving ClientHello even if the
11696 ClientHello is fragmented. (The server can't insist on the
11697 particular protocol version it has chosen before the ServerHello
11698 message has informed the client about his choice.)
11702 * Load error codes if they are not already present instead of using a
11703 static variable. This allows them to be cleanly unloaded and reloaded.
11707 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11709 * Introduce limits to prevent malicious keys being able to
11710 cause a denial of service. ([CVE-2006-2940])
11712 *Steve Henson, Bodo Moeller*
11714 * Fix ASN.1 parsing of certain invalid structures that can result
11715 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11717 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11718 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11720 * Fix SSL client code which could crash if connecting to a
11721 malicious SSLv2 server. ([CVE-2006-4343])
11723 *Tavis Ormandy and Will Drewry, Google Security Team*
11725 * Change ciphersuite string processing so that an explicit
11726 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11727 will no longer include "AES128-SHA"), and any other similar
11728 ciphersuite (same bitmap) from *other* protocol versions (so that
11729 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11730 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11731 changes from 0.9.8b and 0.9.8d.
11735 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11737 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11738 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11740 * Change the Unix randomness entropy gathering to use poll() when
11741 possible instead of select(), since the latter has some
11742 undesirable limitations.
11744 *Darryl Miles via Richard Levitte and Bodo Moeller*
11746 * Disable rogue ciphersuites:
11748 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11749 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11750 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11752 The latter two were purportedly from
11753 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11756 Also deactivate the remaining ciphersuites from
11757 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11758 unofficial, and the ID has long expired.
11762 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11763 dual-core machines) and other potential thread-safety issues.
11767 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11769 * Adapt fipsld and the build system to link against the validated FIPS
11770 module in FIPS mode.
11774 * Fixes for VC++ 2005 build under Windows.
11778 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11779 from a Windows bash shell such as MSYS. It is autodetected from the
11780 "config" script when run from a VC++ environment. Modify standard VC++
11781 build to use fipscanister.o from the GNU make build.
11785 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11787 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11788 The value now differs depending on if you build for FIPS or not.
11789 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11790 safely run with a non-FIPSed libcrypto, as it may crash because of
11791 the difference induced by this change.
11795 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11797 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11798 (part of SSL_OP_ALL). This option used to disable the
11799 countermeasure against man-in-the-middle protocol-version
11800 rollback in the SSL 2.0 server implementation, which is a bad
11801 idea. ([CVE-2005-2969])
11803 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11804 for Information Security, National Institute of Advanced Industrial
11805 Science and Technology [AIST, Japan)]*
11807 * Minimal support for X9.31 signatures and PSS padding modes. This is
11808 mainly for FIPS compliance and not fully integrated at this stage.
11812 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11813 the exponentiation using a fixed-length exponent. (Otherwise,
11814 the information leaked through timing could expose the secret key
11815 after many signatures; cf. Bleichenbacher's attack on DSA with
11820 * Make a new fixed-window mod_exp implementation the default for
11821 RSA, DSA, and DH private-key operations so that the sequence of
11822 squares and multiplies and the memory access pattern are
11823 independent of the particular secret key. This will mitigate
11824 cache-timing and potential related attacks.
11826 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11827 and this is automatically used by BN_mod_exp_mont() if the new flag
11828 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11829 will use this BN flag for private exponents unless the flag
11830 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11831 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11833 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11835 * Change the client implementation for SSLv23_method() and
11836 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11837 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11838 (Previously, the SSL 2.0 backwards compatible Client Hello
11839 message format would be used even with SSL_OP_NO_SSLv2.)
11843 * Add support for smime-type MIME parameter in S/MIME messages which some
11848 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11849 a threadsafe manner. Modify rsa code to use new function and add calls
11850 to dsa and dh code (which had race conditions before).
11854 * Include the fixed error library code in the C error file definitions
11855 instead of fixing them up at runtime. This keeps the error code
11856 structures constant.
11860 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11862 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11865 * Fixes for newer kerberos headers. NB: the casts are needed because
11866 the 'length' field is signed on one version and unsigned on another
11867 with no (?) obvious way to tell the difference, without these VC++
11868 complains. Also the "definition" of FAR (blank) is no longer included
11869 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11870 some needed definitions.
11874 * Undo Cygwin change.
11878 * Added support for proxy certificates according to RFC 3820.
11879 Because they may be a security thread to unaware applications,
11880 they must be explicitly allowed in run-time. See
11881 docs/HOWTO/proxy_certificates.txt for further information.
11885 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11887 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11888 server and client random values. Previously
11889 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11890 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11892 This change has negligible security impact because:
11894 1. Server and client random values still have 24 bytes of pseudo random
11897 2. Server and client random values are sent in the clear in the initial
11900 3. The master secret is derived using the premaster secret (48 bytes in
11901 size for static RSA ciphersuites) as well as client server and random
11904 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11907 *Stephen Henson, reported by UK NISCC*
11909 * Use Windows randomness collection on Cygwin.
11913 * Fix hang in EGD/PRNGD query when communication socket is closed
11914 prematurely by EGD/PRNGD.
11916 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11918 * Prompt for pass phrases when appropriate for PKCS12 input format.
11922 * Back-port of selected performance improvements from development
11923 branch, as well as improved support for PowerPC platforms.
11927 * Add lots of checks for memory allocation failure, error codes to indicate
11928 failure and freeing up memory if a failure occurs.
11930 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11932 * Add new -passin argument to dgst.
11936 * Perform some character comparisons of different types in X509_NAME_cmp:
11937 this is needed for some certificates that re-encode DNs into UTF8Strings
11938 (in violation of RFC3280) and can't or won't issue name rollover
11943 * Make an explicit check during certificate validation to see that
11944 the CA setting in each certificate on the chain is correct. As a
11945 side effect always do the following basic checks on extensions,
11946 not just when there's an associated purpose to the check:
11948 - if there is an unhandled critical extension (unless the user
11949 has chosen to ignore this fault)
11950 - if the path length has been exceeded (if one is set at all)
11951 - that certain extensions fit the associated purpose (if one has
11956 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11958 * Avoid a race condition when CRLs are checked in a multi threaded
11959 environment. This would happen due to the reordering of the revoked
11960 entries during signature checking and serial number lookup. Now the
11961 encoding is cached and the serial number sort performed under a lock.
11962 Add new STACK function sk_is_sorted().
11966 * Add Delta CRL to the extension code.
11970 * Various fixes to s3_pkt.c so alerts are sent properly.
11972 *David Holmes <d.holmes@f5.com>*
11974 * Reduce the chances of duplicate issuer name and serial numbers (in
11975 violation of RFC3280) using the OpenSSL certificate creation utilities.
11976 This is done by creating a random 64 bit value for the initial serial
11977 number when a serial number file is created or when a self signed
11978 certificate is created using 'openssl req -x509'. The initial serial
11979 number file is created using 'openssl x509 -next_serial' in CA.pl
11980 rather than being initialized to 1.
11984 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11986 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11987 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11989 *Joe Orton, Steve Henson*
11991 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11994 *Joe Orton, Steve Henson*
11996 * Make it possible to have multiple active certificates with the same
11997 subject in the CA index file. This is done only if the keyword
11998 'unique_subject' is set to 'no' in the main CA section (default
11999 if 'CA_default') of the configuration file. The value is saved
12000 with the database itself in a separate index attribute file,
12001 named like the index file with '.attr' appended to the name.
12005 * X509 verify fixes. Disable broken certificate workarounds when
12006 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12007 keyUsage extension present. Don't accept CRLs with unhandled critical
12008 extensions: since verify currently doesn't process CRL extensions this
12009 rejects a CRL with *any* critical extensions. Add new verify error codes
12014 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12015 A clarification of RFC2560 will require the use of OCTET STRINGs and
12016 some implementations cannot handle the current raw format. Since OpenSSL
12017 copies and compares OCSP nonces as opaque blobs without any attempt at
12018 parsing them this should not create any compatibility issues.
12022 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12023 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12024 this HMAC (and other) operations are several times slower than OpenSSL
12029 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12031 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12033 * Use the correct content when signing type "other".
12037 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12039 * Fix various bugs revealed by running the NISCC test suite:
12041 Stop out of bounds reads in the ASN1 code when presented with
12042 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12044 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12046 If verify callback ignores invalid public key errors don't try to check
12047 certificate signature with the NULL public key.
12051 * New -ignore_err option in ocsp application to stop the server
12052 exiting on the first error in a request.
12056 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12057 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12062 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12063 extra data after the compression methods not only for TLS 1.0
12064 but also for SSL 3.0 (as required by the specification).
12066 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12068 * Change X509_certificate_type() to mark the key as exported/exportable
12069 when it's 512 *bits* long, not 512 bytes.
12073 * Change AES_cbc_encrypt() so it outputs exact multiple of
12074 blocks during encryption.
12078 * Various fixes to base64 BIO and non blocking I/O. On write
12079 flushes were not handled properly if the BIO retried. On read
12080 data was not being buffered properly and had various logic bugs.
12081 This also affects blocking I/O when the data being decoded is a
12086 * Various S/MIME bugfixes and compatibility changes:
12087 output correct application/pkcs7 MIME type if
12088 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12089 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12090 of files as .eml work). Correctly handle very long lines in MIME
12095 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12097 * Countermeasure against the Klima-Pokorny-Rosa extension of
12098 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12099 a protocol version number mismatch like a decryption error
12100 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12104 * Turn on RSA blinding by default in the default implementation
12105 to avoid a timing attack. Applications that don't want it can call
12106 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12107 They would be ill-advised to do so in most cases.
12109 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12111 * Change RSA blinding code so that it works when the PRNG is not
12112 seeded (in this case, the secret RSA exponent is abused as
12113 an unpredictable seed -- if it is not unpredictable, there
12114 is no point in blinding anyway). Make RSA blinding thread-safe
12115 by remembering the creator's thread ID in rsa->blinding and
12116 having all other threads use local one-time blinding factors
12117 (this requires more computation than sharing rsa->blinding, but
12118 avoids excessive locking; and if an RSA object is not shared
12119 between threads, blinding will still be very fast).
12123 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12124 ENGINE as defaults for all supported algorithms irrespective of
12125 the 'flags' parameter. 'flags' is now honoured, so applications
12126 should make sure they are passing it correctly.
12130 * Target "mingw" now allows native Windows code to be generated in
12131 the Cygwin environment as well as with the MinGW compiler.
12135 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12137 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12138 via timing by performing a MAC computation even if incorrect
12139 block cipher padding has been found. This is a countermeasure
12140 against active attacks where the attacker has to distinguish
12141 between bad padding and a MAC verification error. ([CVE-2003-0078])
12143 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12144 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12145 Martin Vuagnoux (EPFL, Ilion)*
12147 * Make the no-err option work as intended. The intention with no-err
12148 is not to have the whole error stack handling routines removed from
12149 libcrypto, it's only intended to remove all the function name and
12150 reason texts, thereby removing some of the footprint that may not
12151 be interesting if those errors aren't displayed anyway.
12153 NOTE: it's still possible for any application or module to have its
12154 own set of error texts inserted. The routines are there, just not
12155 used by default when no-err is given.
12159 * Add support for FreeBSD on IA64.
12161 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12163 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12164 Kerberos function mit_des_cbc_cksum(). Before this change,
12165 the value returned by DES_cbc_cksum() was like the one from
12166 mit_des_cbc_cksum(), except the bytes were swapped.
12168 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12170 * Allow an application to disable the automatic SSL chain building.
12171 Before this a rather primitive chain build was always performed in
12172 ssl3_output_cert_chain(): an application had no way to send the
12173 correct chain if the automatic operation produced an incorrect result.
12175 Now the chain builder is disabled if either:
12177 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12179 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12181 The reasoning behind this is that an application would not want the
12182 auto chain building to take place if extra chain certificates are
12183 present and it might also want a means of sending no additional
12184 certificates (for example the chain has two certificates and the
12189 * Add the possibility to build without the ENGINE framework.
12191 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12193 * Under Win32 gmtime() can return NULL: check return value in
12194 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12198 * DSA routines: under certain error conditions uninitialized BN objects
12199 could be freed. Solution: make sure initialization is performed early
12200 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12201 Nils Larsch <nla@trustcenter.de> via PR#459)
12205 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12206 checked on reconnect on the client side, therefore session resumption
12207 could still fail with a "ssl session id is different" error. This
12208 behaviour is masked when SSL_OP_ALL is used due to
12209 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12210 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12211 followup to PR #377.
12215 * IA-32 assembler support enhancements: unified ELF targets, support
12216 for SCO/Caldera platforms, fix for Cygwin shared build.
12220 * Add support for FreeBSD on sparc64. As a consequence, support for
12221 FreeBSD on non-x86 processors is separate from x86 processors on
12222 the config script, much like the NetBSD support.
12224 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12226 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12228 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12231 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12232 code (06) was taken as the first octet of the session ID and the last
12233 octet was ignored consequently. As a result SSLv2 client side session
12234 caching could not have worked due to the session ID mismatch between
12236 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12241 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12242 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12247 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12248 seems that in spite of existing for more than a year, many application
12249 author have done nothing to provide the necessary callbacks, which
12250 means that this particular engine will not work properly anywhere.
12251 This is a very unfortunate situation which forces us, in the name
12252 of usability, to give the hw_ncipher.c a static lock, which is part
12254 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12255 appear in 0.9.8 or later. We EXPECT application authors to have
12256 dealt properly with this when 0.9.8 is released (unless we actually
12257 make such changes in the libcrypto locking code that changes will
12258 have to be made anyway).
12262 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12263 octets have been read, EOF or an error occurs. Without this change
12264 some truncated ASN1 structures will not produce an error.
12268 * Disable Heimdal support, since it hasn't been fully implemented.
12269 Still give the possibility to force the use of Heimdal, but with
12270 warnings and a request that patches get sent to openssl-dev.
12274 * Add the VC-CE target, introduce the WINCE sysname, and add
12275 INSTALL.WCE and appropriate conditionals to make it build.
12277 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12279 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12280 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12281 edit numbers of the version.
12283 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12285 * Introduce safe string copy and catenation functions
12286 (BUF_strlcpy() and BUF_strlcat()).
12288 *Ben Laurie (CHATS) and Richard Levitte*
12290 * Avoid using fixed-size buffers for one-line DNs.
12292 *Ben Laurie (CHATS)*
12294 * Add BUF_MEM_grow_clean() to avoid information leakage when
12295 resizing buffers containing secrets, and use where appropriate.
12297 *Ben Laurie (CHATS)*
12299 * Avoid using fixed size buffers for configuration file location.
12301 *Ben Laurie (CHATS)*
12303 * Avoid filename truncation for various CA files.
12305 *Ben Laurie (CHATS)*
12307 * Use sizeof in preference to magic numbers.
12309 *Ben Laurie (CHATS)*
12311 * Avoid filename truncation in cert requests.
12313 *Ben Laurie (CHATS)*
12315 * Add assertions to check for (supposedly impossible) buffer
12318 *Ben Laurie (CHATS)*
12320 * Don't cache truncated DNS entries in the local cache (this could
12321 potentially lead to a spoofing attack).
12323 *Ben Laurie (CHATS)*
12325 * Fix various buffers to be large enough for hex/decimal
12326 representations in a platform independent manner.
12328 *Ben Laurie (CHATS)*
12330 * Add CRYPTO_realloc_clean() to avoid information leakage when
12331 resizing buffers containing secrets, and use where appropriate.
12333 *Ben Laurie (CHATS)*
12335 * Add BIO_indent() to avoid much slightly worrying code to do
12338 *Ben Laurie (CHATS)*
12340 * Convert sprintf()/BIO_puts() to BIO_printf().
12342 *Ben Laurie (CHATS)*
12344 * buffer_gets() could terminate with the buffer only half
12347 *Ben Laurie (CHATS)*
12349 * Add assertions to prevent user-supplied crypto functions from
12350 overflowing internal buffers by having large block sizes, etc.
12352 *Ben Laurie (CHATS)*
12354 * New OPENSSL_assert() macro (similar to assert(), but enabled
12357 *Ben Laurie (CHATS)*
12359 * Eliminate unused copy of key in RC4.
12361 *Ben Laurie (CHATS)*
12363 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12365 *Ben Laurie (CHATS)*
12367 * Fix off-by-one error in EGD path.
12369 *Ben Laurie (CHATS)*
12371 * If RANDFILE path is too long, ignore instead of truncating.
12373 *Ben Laurie (CHATS)*
12375 * Eliminate unused and incorrectly sized X.509 structure
12378 *Ben Laurie (CHATS)*
12380 * Eliminate unused and dangerous function knumber().
12382 *Ben Laurie (CHATS)*
12384 * Eliminate unused and dangerous structure, KSSL_ERR.
12386 *Ben Laurie (CHATS)*
12388 * Protect against overlong session ID context length in an encoded
12389 session object. Since these are local, this does not appear to be
12392 *Ben Laurie (CHATS)*
12394 * Change from security patch (see 0.9.6e below) that did not affect
12395 the 0.9.6 release series:
12397 Remote buffer overflow in SSL3 protocol - an attacker could
12398 supply an oversized master key in Kerberos-enabled versions.
12401 *Ben Laurie (CHATS)*
12403 * Change the SSL kerb5 codes to match RFC 2712.
12407 * Make -nameopt work fully for req and add -reqopt switch.
12409 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12411 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12413 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12415 * Make sure tests can be performed even if the corresponding algorithms
12416 have been removed entirely. This was also the last step to make
12417 OpenSSL compilable with DJGPP under all reasonable conditions.
12419 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12421 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12422 to allow version independent disabling of normally unselected ciphers,
12423 which may be activated as a side-effect of selecting a single cipher.
12425 (E.g., cipher list string "RSA" enables ciphersuites that are left
12426 out of "ALL" because they do not provide symmetric encryption.
12427 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12429 *Lutz Jaenicke, Bodo Moeller*
12431 * Add appropriate support for separate platform-dependent build
12432 directories. The recommended way to make a platform-dependent
12433 build directory is the following (tested on Linux), maybe with
12436 # Place yourself outside of the OpenSSL source tree. In
12437 # this example, the environment variable OPENSSL_SOURCE
12438 # is assumed to contain the absolute OpenSSL source directory.
12439 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12440 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12441 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12442 mkdir -p `dirname $F`
12443 ln -s $OPENSSL_SOURCE/$F $F
12446 To be absolutely sure not to disturb the source tree, a "make clean"
12447 is a good thing. If it isn't successful, don't worry about it,
12448 it probably means the source directory is very clean.
12452 * Make sure any ENGINE control commands make local copies of string
12453 pointers passed to them whenever necessary. Otherwise it is possible
12454 the caller may have overwritten (or deallocated) the original string
12455 data when a later ENGINE operation tries to use the stored values.
12457 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12459 * Improve diagnostics in file reading and command-line digests.
12461 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12463 * Add AES modes CFB and OFB to the object database. Correct an
12464 error in AES-CFB decryption.
12468 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12469 allows existing EVP_CIPHER_CTX structures to be reused after
12470 calling `EVP_*Final()`. This behaviour is used by encryption
12471 BIOs and some applications. This has the side effect that
12472 applications must explicitly clean up cipher contexts with
12473 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12477 * Check the values of dna and dnb in bn_mul_recursive before calling
12478 bn_mul_comba (a non zero value means the a or b arrays do not contain
12479 n2 elements) and fallback to bn_mul_normal if either is not zero.
12483 * Fix escaping of non-ASCII characters when using the -subj option
12484 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12488 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12489 form for "surname", serialNumber has no short form.
12490 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12491 therefore remove "mail" short name for "internet 7".
12492 The OID for unique identifiers in X509 certificates is
12493 x500UniqueIdentifier, not uniqueIdentifier.
12494 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12498 * Add an "init" command to the ENGINE config module and auto initialize
12499 ENGINEs. Without any "init" command the ENGINE will be initialized
12500 after all ctrl commands have been executed on it. If init=1 the
12501 ENGINE is initialized at that point (ctrls before that point are run
12502 on the uninitialized ENGINE and after on the initialized one). If
12503 init=0 then the ENGINE will not be initialized at all.
12507 * Fix the 'app_verify_callback' interface so that the user-defined
12508 argument is actually passed to the callback: In the
12509 SSL_CTX_set_cert_verify_callback() prototype, the callback
12510 declaration has been changed from
12513 int (*cb)(X509_STORE_CTX *,void *);
12514 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12515 i=s->ctx->app_verify_callback(&ctx)
12516 has been changed into
12517 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12519 To update applications using SSL_CTX_set_cert_verify_callback(),
12520 a dummy argument can be added to their callback functions.
12522 *D. K. Smetters <smetters@parc.xerox.com>*
12524 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12526 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12528 * Add and OPENSSL_LOAD_CONF define which will cause
12529 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12530 This allows older applications to transparently support certain
12531 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12532 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12533 load the config file and OPENSSL_add_all_algorithms_conf() which will
12534 always load it have also been added.
12538 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12539 Adjust NIDs and EVP layer.
12541 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12543 * Config modules support in openssl utility.
12545 Most commands now load modules from the config file,
12546 though in a few (such as version) this isn't done
12547 because it couldn't be used for anything.
12549 In the case of ca and req the config file used is
12550 the same as the utility itself: that is the -config
12551 command line option can be used to specify an
12556 * Move default behaviour from OPENSSL_config(). If appname is NULL
12557 use "openssl_conf" if filename is NULL use default openssl config file.
12561 * Add an argument to OPENSSL_config() to allow the use of an alternative
12562 config section name. Add a new flag to tolerate a missing config file
12563 and move code to CONF_modules_load_file().
12567 * Support for crypto accelerator cards from Accelerated Encryption
12568 Processing, www.aep.ie. (Use engine 'aep')
12569 The support was copied from 0.9.6c [engine] and adapted/corrected
12570 to work with the new engine framework.
12572 *AEP Inc. and Richard Levitte*
12574 * Support for SureWare crypto accelerator cards from Baltimore
12575 Technologies. (Use engine 'sureware')
12576 The support was copied from 0.9.6c [engine] and adapted
12577 to work with the new engine framework.
12581 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12582 make the newer ENGINE framework commands for the CHIL engine work.
12584 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12586 * Make it possible to produce shared libraries on ReliantUNIX.
12588 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12590 * Add the configuration target debug-linux-ppro.
12591 Make 'openssl rsa' use the general key loading routines
12592 implemented in `apps.c`, and make those routines able to
12593 handle the key format FORMAT_NETSCAPE and the variant
12596 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12598 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12600 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12602 * Add -keyform to rsautl, and document -engine.
12604 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12606 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12607 BIO_R_NO_SUCH_FILE error code rather than the generic
12608 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12612 * Add new functions
12613 ERR_peek_last_error
12614 ERR_peek_last_error_line
12615 ERR_peek_last_error_line_data.
12616 These are similar to
12618 ERR_peek_error_line
12619 ERR_peek_error_line_data,
12620 but report on the latest error recorded rather than the first one
12621 still in the error queue.
12623 *Ben Laurie, Bodo Moeller*
12625 * default_algorithms option in ENGINE config module. This allows things
12627 default_algorithms = ALL
12628 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12632 * Preliminary ENGINE config module.
12636 * New experimental application configuration code.
12640 * Change the AES code to follow the same name structure as all other
12641 symmetric ciphers, and behave the same way. Move everything to
12642 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12644 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12646 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12648 *Ben Laurie and Theo de Raadt*
12650 * Add option to output public keys in req command.
12652 *Massimiliano Pala madwolf@openca.org*
12654 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12655 (up to about 10% better than before for P-192 and P-224).
12659 * New functions/macros
12661 SSL_CTX_set_msg_callback(ctx, cb)
12662 SSL_CTX_set_msg_callback_arg(ctx, arg)
12663 SSL_set_msg_callback(ssl, cb)
12664 SSL_set_msg_callback_arg(ssl, arg)
12666 to request calling a callback function
12668 void cb(int write_p, int version, int content_type,
12669 const void *buf, size_t len, SSL *ssl, void *arg)
12671 whenever a protocol message has been completely received
12672 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12673 protocol version according to which the SSL library interprets
12674 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12675 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12676 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12677 specification (change_cipher_spec(20), alert(21), handshake(22)).
12678 'buf' and 'len' point to the actual message, 'ssl' to the
12679 SSL object, and 'arg' is the application-defined value set by
12680 SSL[_CTX]_set_msg_callback_arg().
12682 'openssl s_client' and 'openssl s_server' have new '-msg' options
12683 to enable a callback that displays all protocol messages.
12687 * Change the shared library support so shared libraries are built as
12688 soon as the corresponding static library is finished, and thereby get
12689 openssl and the test programs linked against the shared library.
12690 This still only happens when the keyword "shard" has been given to
12691 the configuration scripts.
12693 NOTE: shared library support is still an experimental thing, and
12694 backward binary compatibility is still not guaranteed.
12696 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12698 * Add support for Subject Information Access extension.
12700 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12702 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12703 additional bytes when new memory had to be allocated, not just
12704 when reusing an existing buffer.
12708 * New command line and configuration option 'utf8' for the req command.
12709 This allows field values to be specified as UTF8 strings.
12713 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12714 runs for the former and machine-readable output for the latter.
12718 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12719 of the e-mail address in the DN (i.e., it will go into a certificate
12720 extension only). The new configuration file option 'email_in_dn = no'
12721 has the same effect.
12723 *Massimiliano Pala madwolf@openca.org*
12725 * Change all functions with names starting with `des_` to be starting
12726 with `DES_` instead. Add wrappers that are compatible with libdes,
12727 but are named `_ossl_old_des_*`. Finally, add macros that map the
12728 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12729 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12730 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12733 Since we provide two compatibility mappings, the user needs to
12734 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12735 compatibility is desired. The default (i.e., when that macro
12736 isn't defined) is OpenSSL 0.9.6c compatibility.
12738 There are also macros that enable and disable the support of old
12739 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12740 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12741 are defined, the default will apply: to support the old des routines.
12743 In either case, one must include openssl/des.h to get the correct
12744 definitions. Do not try to just include openssl/des_old.h, that
12747 NOTE: This is a major break of an old API into a new one. Software
12748 authors are encouraged to switch to the `DES_` style functions. Some
12749 time in the future, des_old.h and the libdes compatibility functions
12750 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12751 default), and then completely removed.
12755 * Test for certificates which contain unsupported critical extensions.
12756 If such a certificate is found during a verify operation it is
12757 rejected by default: this behaviour can be overridden by either
12758 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12759 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12760 X509_supported_extension() has also been added which returns 1 if a
12761 particular extension is supported.
12765 * Modify the behaviour of EVP cipher functions in similar way to digests
12766 to retain compatibility with existing code.
12770 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12771 compatibility with existing code. In particular the 'ctx' parameter does
12772 not have to be to be initialized before the call to EVP_DigestInit() and
12773 it is tidied up after a call to EVP_DigestFinal(). New function
12774 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12775 EVP_MD_CTX_copy() changed to not require the destination to be
12776 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12777 requires the destination to be valid.
12779 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12780 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12784 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12785 so that complete 'Handshake' protocol structures are kept in memory
12786 instead of overwriting 'msg_type' and 'length' with 'body' data.
12790 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12792 *Massimo Santin via Richard Levitte*
12794 * Major restructuring to the underlying ENGINE code. This includes
12795 reduction of linker bloat, separation of pure "ENGINE" manipulation
12796 (initialisation, etc) from functionality dealing with implementations
12797 of specific crypto interfaces. This change also introduces integrated
12798 support for symmetric ciphers and digest implementations - so ENGINEs
12799 can now accelerate these by providing EVP_CIPHER and EVP_MD
12800 implementations of their own. This is detailed in
12801 [crypto/engine/README.md](crypto/engine/README.md)
12802 as it couldn't be adequately described here. However, there are a few
12803 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12804 were changed in the original introduction of ENGINE code have now
12805 reverted back - the hooking from this code to ENGINE is now a good
12806 deal more passive and at run-time, operations deal directly with
12807 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12808 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12809 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12810 they were not being used by the framework as there is no concept of a
12811 BIGNUM_METHOD and they could not be generalised to the new
12812 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12813 ENGINE_cpy() has been removed as it cannot be consistently defined in
12818 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12822 * Change mkdef.pl to sort symbols that get the same entry number,
12823 and make sure the automatically generated functions `ERR_load_*`
12824 become part of libeay.num as well.
12828 * New function SSL_renegotiate_pending(). This returns true once
12829 renegotiation has been requested (either SSL_renegotiate() call
12830 or HelloRequest/ClientHello received from the peer) and becomes
12831 false once a handshake has been completed.
12832 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12833 sends a HelloRequest, but does not ensure that a handshake takes
12834 place. SSL_renegotiate_pending() is useful for checking if the
12835 client has followed the request.)
12839 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12840 By default, clients may request session resumption even during
12841 renegotiation (if session ID contexts permit); with this option,
12842 session resumption is possible only in the first handshake.
12844 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12845 more bits available for options that should not be part of
12846 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12850 * Add some demos for certificate and certificate request creation.
12854 * Make maximum certificate chain size accepted from the peer application
12855 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12856 "Douglas E. Engert" <deengert@anl.gov>.
12860 * Add support for shared libraries for Unixware-7
12861 (Boyd Lynn Gerber <gerberb@zenez.com>).
12865 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12866 be done prior to destruction. Use this to unload error strings from
12867 ENGINEs that load their own error strings. NB: This adds two new API
12868 functions to "get" and "set" this destroy handler in an ENGINE.
12872 * Alter all existing ENGINE implementations (except "openssl" and
12873 "openbsd") to dynamically instantiate their own error strings. This
12874 makes them more flexible to be built both as statically-linked ENGINEs
12875 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12876 Also, add stub code to each that makes building them as self-contained
12877 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12881 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12882 implementations into applications that are completely implemented in
12883 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12884 commands that can be used to configure what shared-library to load and
12885 to control aspects of the way it is handled. Also, made an update to
12886 the [README-Engine.md](README-Engine.md) file
12887 that brings its information up-to-date and
12888 provides some information and instructions on the "dynamic" ENGINE
12889 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12893 * Make it possible to unload ranges of ERR strings with a new
12894 "ERR_unload_strings" function.
12898 * Add a copy() function to EVP_MD.
12902 * Make EVP_MD routines take a context pointer instead of just the
12903 md_data void pointer.
12907 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12908 that the digest can only process a single chunk of data
12909 (typically because it is provided by a piece of
12910 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12911 is only going to provide a single chunk of data, and hence the
12912 framework needn't accumulate the data for oneshot drivers.
12916 * As with "ERR", make it possible to replace the underlying "ex_data"
12917 functions. This change also alters the storage and management of global
12918 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12919 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12920 index counters. The API functions that use this state have been changed
12921 to take a "class_index" rather than pointers to the class's local STACK
12922 and counter, and there is now an API function to dynamically create new
12923 classes. This centralisation allows us to (a) plug a lot of the
12924 thread-safety problems that existed, and (b) makes it possible to clean
12925 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12926 such data would previously have always leaked in application code and
12927 workarounds were in place to make the memory debugging turn a blind eye
12928 to it. Application code that doesn't use this new function will still
12929 leak as before, but their memory debugging output will announce it now
12930 rather than letting it slide.
12932 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12933 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12934 has a return value to indicate success or failure.
12938 * Make it possible to replace the underlying "ERR" functions such that the
12939 global state (2 LHASH tables and 2 locks) is only used by the "default"
12940 implementation. This change also adds two functions to "get" and "set"
12941 the implementation prior to it being automatically set the first time
12942 any other ERR function takes place. Ie. an application can call "get",
12943 pass the return value to a module it has just loaded, and that module
12944 can call its own "set" function using that value. This means the
12945 module's "ERR" operations will use (and modify) the error state in the
12946 application and not in its own statically linked copy of OpenSSL code.
12950 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12951 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12952 the operation, and provides a more encapsulated way for external code
12953 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12954 to use these functions rather than manually incrementing the counts.
12956 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12960 * Add EVP test program.
12964 * Add symmetric cipher support to ENGINE. Expect the API to change!
12968 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12969 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12970 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12971 These allow a CRL to be built without having to access X509_CRL fields
12972 directly. Modify 'ca' application to use new functions.
12976 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12977 bug workarounds. Rollback attack detection is a security feature.
12978 The problem will only arise on OpenSSL servers when TLSv1 is not
12979 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12980 Software authors not wanting to support TLSv1 will have special reasons
12981 for their choice and can explicitly enable this option.
12983 *Bodo Moeller, Lutz Jaenicke*
12985 * Rationalise EVP so it can be extended: don't include a union of
12986 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12987 (similar to those existing for EVP_CIPHER_CTX).
12992 EVP_MD_CTX_init(&md); /* new function call */
12993 EVP_DigestInit(&md, EVP_sha1());
12994 EVP_DigestUpdate(&md, in, len);
12995 EVP_DigestFinal(&md, out, NULL);
12996 EVP_MD_CTX_cleanup(&md); /* new function call */
13000 * Make DES key schedule conform to the usual scheme, as well as
13001 correcting its structure. This means that calls to DES functions
13002 now have to pass a pointer to a des_key_schedule instead of a
13003 plain des_key_schedule (which was actually always a pointer
13006 des_key_schedule ks;
13008 des_set_key_checked(..., &ks);
13009 des_ncbc_encrypt(..., &ks, ...);
13011 (Note that a later change renames 'des_...' into 'DES_...'.)
13015 * Initial reduction of linker bloat: the use of some functions, such as
13016 PEM causes large amounts of unused functions to be linked in due to
13017 poor organisation. For example pem_all.c contains every PEM function
13018 which has a knock on effect of linking in large amounts of (unused)
13019 ASN1 code. Grouping together similar functions and splitting unrelated
13020 functions prevents this.
13024 * Cleanup of EVP macros.
13028 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13029 correct `_ecb suffix`.
13033 * Add initial OCSP responder support to ocsp application. The
13034 revocation information is handled using the text based index
13035 use by the ca application. The responder can either handle
13036 requests generated internally, supplied in files (for example
13037 via a CGI script) or using an internal minimal server.
13041 * Add configuration choices to get zlib compression for TLS.
13045 * Changes to Kerberos SSL for RFC 2712 compliance:
13046 1. Implemented real KerberosWrapper, instead of just using
13047 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13048 2. Implemented optional authenticator field of KerberosWrapper.
13050 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13051 and authenticator structs; see crypto/krb5/.
13053 Generalized Kerberos calls to support multiple Kerberos libraries.
13054 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13055 via Richard Levitte*
13057 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13058 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13059 values for each of the key sizes rather than having just
13060 parameters (and 'speed' generating keys each time).
13064 * Speed up EVP routines.
13067 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13068 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13069 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13070 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13072 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13073 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13074 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13077 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13079 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13083 * Added the OS2-EMX target.
13085 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13087 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13088 New functions to support `NCONF` routines in extension code.
13089 New function `CONF_set_nconf()`
13090 to allow functions which take an `NCONF` to also handle the old `LHASH`
13091 structure: this means that the old `CONF` compatible routines can be
13092 retained (in particular w.rt. extensions) without having to duplicate the
13093 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13097 * Enhance the general user interface with mechanisms for inner control
13098 and with possibilities to have yes/no kind of prompts.
13102 * Change all calls to low-level digest routines in the library and
13103 applications to use EVP. Add missing calls to HMAC_cleanup() and
13104 don't assume HMAC_CTX can be copied using memcpy().
13106 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13108 * Add the possibility to control engines through control names but with
13109 arbitrary arguments instead of just a string.
13110 Change the key loaders to take a UI_METHOD instead of a callback
13111 function pointer. NOTE: this breaks binary compatibility with earlier
13112 versions of OpenSSL [engine].
13113 Adapt the nCipher code for these new conditions and add a card insertion
13118 * Enhance the general user interface with mechanisms to better support
13119 dialog box interfaces, application-defined prompts, the possibility
13120 to use defaults (for example default passwords from somewhere else)
13121 and interrupts/cancellations.
13125 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13126 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13130 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13131 tidy up some unnecessarily weird code in 'sk_new()').
13133 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13135 * Change the key loading routines for ENGINEs to use the same kind
13136 callback (pem_password_cb) as all other routines that need this
13141 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13142 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13143 than this minimum value is recommended.
13147 * New random seeder for OpenVMS, using the system process statistics
13148 that are easily reachable.
13152 * Windows apparently can't transparently handle global
13153 variables defined in DLLs. Initialisations such as:
13155 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13157 won't compile. This is used by the any applications that need to
13158 declare their own ASN1 modules. This was fixed by adding the option
13159 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13160 needed for static libraries under Win32.
13164 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13165 setting of purpose and trust fields. New X509_STORE trust and
13166 purpose functions and tidy up setting in other SSL functions.
13170 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13171 structure. These are inherited by X509_STORE_CTX when it is
13172 initialised. This allows various defaults to be set in the
13173 X509_STORE structure (such as flags for CRL checking and custom
13174 purpose or trust settings) for functions which only use X509_STORE_CTX
13175 internally such as S/MIME.
13177 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13178 trust settings if they are not set in X509_STORE. This allows X509_STORE
13179 purposes and trust (in S/MIME for example) to override any set by default.
13181 Add command line options for CRL checking to smime, s_client and s_server
13186 * Initial CRL based revocation checking. If the CRL checking flag(s)
13187 are set then the CRL is looked up in the X509_STORE structure and
13188 its validity and signature checked, then if the certificate is found
13189 in the CRL the verify fails with a revoked error.
13191 Various new CRL related callbacks added to X509_STORE_CTX structure.
13193 Command line options added to 'verify' application to support this.
13195 This needs some additional work, such as being able to handle multiple
13196 CRLs with different times, extension based lookup (rather than just
13197 by subject name) and ultimately more complete V2 CRL extension
13202 * Add a general user interface API (crypto/ui/). This is designed
13203 to replace things like des_read_password and friends (backward
13204 compatibility functions using this new API are provided).
13205 The purpose is to remove prompting functions from the DES code
13206 section as well as provide for prompting through dialog boxes in
13207 a window system and the like.
13211 * Add "ex_data" support to ENGINE so implementations can add state at a
13212 per-structure level rather than having to store it globally.
13216 * Make it possible for ENGINE structures to be copied when retrieved by
13217 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13218 This causes the "original" ENGINE structure to act like a template,
13219 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13220 operational state can be localised to each ENGINE structure, despite the
13221 fact they all share the same "methods". New ENGINE structures returned in
13222 this case have no functional references and the return value is the single
13223 structural reference. This matches the single structural reference returned
13224 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13229 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13230 needs to match any other type at all we need to manually clear the
13235 * Changes to the "openssl engine" utility to include;
13236 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13237 about an ENGINE's available control commands.
13238 - executing control commands from command line arguments using the
13239 '-pre' and '-post' switches. '-post' is only used if '-t' is
13240 specified and the ENGINE is successfully initialised. The syntax for
13241 the individual commands are colon-separated, for example;
13242 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13246 * New dynamic control command support for ENGINEs. ENGINEs can now
13247 declare their own commands (numbers), names (strings), descriptions,
13248 and input types for run-time discovery by calling applications. A
13249 subset of these commands are implicitly classed as "executable"
13250 depending on their input type, and only these can be invoked through
13251 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13252 can be based on user input, config files, etc). The distinction is
13253 that "executable" commands cannot return anything other than a boolean
13254 result and can only support numeric or string input, whereas some
13255 discoverable commands may only be for direct use through
13256 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13257 pointers, or other custom uses. The "executable" commands are to
13258 support parameterisations of ENGINE behaviour that can be
13259 unambiguously defined by ENGINEs and used consistently across any
13260 OpenSSL-based application. Commands have been added to all the
13261 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13262 control over shared-library paths without source code alterations.
13266 * Changed all ENGINE implementations to dynamically allocate their
13267 ENGINEs rather than declaring them statically. Apart from this being
13268 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13269 this also allows the implementations to compile without using the
13270 internal engine_int.h header.
13274 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13275 'const' value. Any code that should be able to modify a RAND_METHOD
13276 should already have non-const pointers to it (ie. they should only
13277 modify their own ones).
13281 * Made a variety of little tweaks to the ENGINE code.
13282 - "atalla" and "ubsec" string definitions were moved from header files
13283 to C code. "nuron" string definitions were placed in variables
13284 rather than hard-coded - allowing parameterisation of these values
13285 later on via ctrl() commands.
13286 - Removed unused "#if 0"'d code.
13287 - Fixed engine list iteration code so it uses ENGINE_free() to release
13288 structural references.
13289 - Constified the RAND_METHOD element of ENGINE structures.
13290 - Constified various get/set functions as appropriate and added
13291 missing functions (including a catch-all ENGINE_cpy that duplicates
13292 all ENGINE values onto a new ENGINE except reference counts/state).
13293 - Removed NULL parameter checks in get/set functions. Setting a method
13294 or function to NULL is a way of cancelling out a previously set
13295 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13296 and doesn't justify the extra error symbols and code.
13297 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13298 flags from engine_int.h to engine.h.
13299 - Changed prototypes for ENGINE handler functions (init(), finish(),
13300 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13304 * Implement binary inversion algorithm for BN_mod_inverse in addition
13305 to the algorithm using long division. The binary algorithm can be
13306 used only if the modulus is odd. On 32-bit systems, it is faster
13307 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13308 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13309 up to 450 bits. In 64-bit environments, the binary algorithm
13310 appears to be advantageous for much longer moduli; here we use it
13311 for moduli up to 2048 bits.
13315 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13316 could not support the combine flag in choice fields.
13320 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13321 extensions from a certificate request to the certificate.
13325 * Allow multiple 'certopt' and 'nameopt' options to be separated
13326 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13327 file: this allows the display of the certificate about to be
13328 signed to be customised, to allow certain fields to be included
13329 or excluded and extension details. The old system didn't display
13330 multicharacter strings properly, omitted fields not in the policy
13331 and couldn't display additional details such as extensions.
13335 * Function EC_POINTs_mul for multiple scalar multiplication
13336 of an arbitrary number of elliptic curve points
13337 \sum scalars[i]*points[i],
13338 optionally including the generator defined for the EC_GROUP:
13339 scalar*generator + \sum scalars[i]*points[i].
13341 EC_POINT_mul is a simple wrapper function for the typical case
13342 that the point list has just one item (besides the optional
13347 * First EC_METHODs for curves over GF(p):
13349 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13350 operations and provides various method functions that can also
13351 operate with faster implementations of modular arithmetic.
13353 EC_GFp_mont_method() reuses most functions that are part of
13354 EC_GFp_simple_method, but uses Montgomery arithmetic.
13356 *Bodo Moeller; point addition and point doubling
13357 implementation directly derived from source code provided by
13358 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13360 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13361 crypto/ec/ec_lib.c):
13363 Curves are EC_GROUP objects (with an optional group generator)
13364 based on EC_METHODs that are built into the library.
13366 Points are EC_POINT objects based on EC_GROUP objects.
13368 Most of the framework would be able to handle curves over arbitrary
13369 finite fields, but as there are no obvious types for fields other
13370 than GF(p), some functions are limited to that for now.
13374 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13375 that the file contains a complete HTTP response.
13379 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13380 change the def and num file printf format specifier from "%-40sXXX"
13381 to "%-39s XXX". The latter will always guarantee a space after the
13382 field while the former will cause them to run together if the field
13383 is 40 of more characters long.
13387 * Constify the cipher and digest 'method' functions and structures
13388 and modify related functions to take constant EVP_MD and EVP_CIPHER
13393 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13394 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13398 * Modify `EVP_Digest*()` routines so they now return values. Although the
13399 internal software routines can never fail additional hardware versions
13404 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13406 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13407 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13410 ERR_R_NESTED_ASN1_ERROR
13412 ERR_R_MISSING_ASN1_EOS
13413 were 4 .. 9, conflicting with
13414 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13416 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13417 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13419 Add new error code 'ERR_R_INTERNAL_ERROR'.
13423 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13428 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13429 sets the subject name for a new request or supersedes the
13430 subject name in a given request. Formats that can be parsed are
13431 'CN=Some Name, OU=myOU, C=IT'
13433 'CN=Some Name/OU=myOU/C=IT'.
13435 Add options '-batch' and '-verbose' to 'openssl req'.
13437 *Massimiliano Pala <madwolf@hackmasters.net>*
13439 * Introduce the possibility to access global variables through
13440 functions on platform were that's the best way to handle exporting
13441 global variables in shared libraries. To enable this functionality,
13442 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13443 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13444 is normally done by Configure or something similar).
13446 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13447 in the source file (foo.c) like this:
13449 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13450 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13452 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13453 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13455 OPENSSL_DECLARE_GLOBAL(int,foo);
13456 #define foo OPENSSL_GLOBAL_REF(foo)
13457 OPENSSL_DECLARE_GLOBAL(double,bar);
13458 #define bar OPENSSL_GLOBAL_REF(bar)
13460 The #defines are very important, and therefore so is including the
13461 header file everywhere where the defined globals are used.
13463 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13464 of ASN.1 items, but that structure is a bit different.
13466 The largest change is in util/mkdef.pl which has been enhanced with
13467 better and easier to understand logic to choose which symbols should
13468 go into the Windows .def files as well as a number of fixes and code
13469 cleanup (among others, algorithm keywords are now sorted
13470 lexicographically to avoid constant rewrites).
13474 * In BN_div() keep a copy of the sign of 'num' before writing the
13475 result to 'rm' because if rm==num the value will be overwritten
13476 and produce the wrong result if 'num' is negative: this caused
13477 problems with BN_mod() and BN_nnmod().
13481 * Function OCSP_request_verify(). This checks the signature on an
13482 OCSP request and verifies the signer certificate. The signer
13483 certificate is just checked for a generic purpose and OCSP request
13488 * Add OCSP_check_validity() function to check the validity of OCSP
13489 responses. OCSP responses are prepared in real time and may only
13490 be a few seconds old. Simply checking that the current time lies
13491 between thisUpdate and nextUpdate max reject otherwise valid responses
13492 caused by either OCSP responder or client clock inaccuracy. Instead
13493 we allow thisUpdate and nextUpdate to fall within a certain period of
13494 the current time. The age of the response can also optionally be
13495 checked. Two new options -validity_period and -status_age added to
13500 * If signature or public key algorithm is unrecognized print out its
13501 OID rather that just UNKNOWN.
13505 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13506 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13507 ID to be generated from the issuer certificate alone which can then be
13508 passed to OCSP_id_issuer_cmp().
13512 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13513 ASN1 modules to export functions returning ASN1_ITEM pointers
13514 instead of the ASN1_ITEM structures themselves. This adds several
13515 new macros which allow the underlying ASN1 function/structure to
13516 be accessed transparently. As a result code should not use ASN1_ITEM
13517 references directly (such as &X509_it) but instead use the relevant
13518 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13519 use of the new ASN1 code on platforms where exporting structures
13520 is problematical (for example in shared libraries) but exporting
13521 functions returning pointers to structures is not.
13525 * Add support for overriding the generation of SSL/TLS session IDs.
13526 These callbacks can be registered either in an SSL_CTX or per SSL.
13527 The purpose of this is to allow applications to control, if they wish,
13528 the arbitrary values chosen for use as session IDs, particularly as it
13529 can be useful for session caching in multiple-server environments. A
13530 command-line switch for testing this (and any client code that wishes
13531 to use such a feature) has been added to "s_server".
13533 *Geoff Thorpe, Lutz Jaenicke*
13535 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13536 of the form `#if defined(...) || defined(...) || ...` and
13537 `#if !defined(...) && !defined(...) && ...`. This also avoids
13538 the growing number of special cases it was previously handling.
13542 * Make all configuration macros available for application by making
13543 sure they are available in opensslconf.h, by giving them names starting
13544 with `OPENSSL_` to avoid conflicts with other packages and by making
13545 sure e_os2.h will cover all platform-specific cases together with
13547 Additionally, it is now possible to define configuration/platform-
13548 specific names (called "system identities"). In the C code, these
13549 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13550 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13551 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13556 * New option -set_serial to 'req' and 'x509' this allows the serial
13557 number to use to be specified on the command line. Previously self
13558 signed certificates were hard coded with serial number 0 and the
13559 CA options of 'x509' had to use a serial number in a file which was
13564 * New options to 'ca' utility to support V2 CRL entry extensions.
13565 Currently CRL reason, invalidity date and hold instruction are
13566 supported. Add new CRL extensions to V3 code and some new objects.
13570 * New function EVP_CIPHER_CTX_set_padding() this is used to
13571 disable standard block padding (aka PKCS#5 padding) in the EVP
13572 API, which was previously mandatory. This means that the data is
13573 not padded in any way and so the total length much be a multiple
13574 of the block size, otherwise an error occurs.
13578 * Initial (incomplete) OCSP SSL support.
13582 * New function OCSP_parse_url(). This splits up a URL into its host,
13583 port and path components: primarily to parse OCSP URLs. New -url
13584 option to ocsp utility.
13588 * New nonce behavior. The return value of OCSP_check_nonce() now
13589 reflects the various checks performed. Applications can decide
13590 whether to tolerate certain situations such as an absent nonce
13591 in a response when one was present in a request: the ocsp application
13592 just prints out a warning. New function OCSP_add1_basic_nonce()
13593 this is to allow responders to include a nonce in a response even if
13594 the request is nonce-less.
13598 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13599 skipped when using openssl x509 multiple times on a single input file,
13600 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13604 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13605 set string type: to handle setting ASN1_TIME structures. Fix ca
13606 utility to correctly initialize revocation date of CRLs.
13610 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13611 the clients preferred ciphersuites and rather use its own preferences.
13612 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13613 Internet Explorer by ensuring unchanged hash method during stepup.
13614 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13618 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13619 to aes and add a new 'exist' option to print out symbols that don't
13624 * Additional options to ocsp utility to allow flags to be set and
13625 additional certificates supplied.
13629 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13630 OCSP client a number of certificate to only verify the response
13635 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13636 handle the new API. Currently only ECB, CBC modes supported. Add new
13639 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13640 Encryption Standard (AES) Ciphersuites for Transport Layer
13641 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13642 not enabled by default and were not part of the "ALL" ciphersuite
13643 alias because they were not yet official; they could be
13644 explicitly requested by specifying the "AESdraft" ciphersuite
13645 group alias. In the final release of OpenSSL 0.9.7, the group
13646 alias is called "AES" and is part of "ALL".)
13648 *Ben Laurie, Steve Henson, Bodo Moeller*
13650 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13651 request to response.
13655 * Functions for OCSP responders. OCSP_request_onereq_count(),
13656 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13657 extract information from a certificate request. OCSP_response_create()
13658 creates a response and optionally adds a basic response structure.
13659 OCSP_basic_add1_status() adds a complete single response to a basic
13660 response and returns the OCSP_SINGLERESP structure just added (to allow
13661 extensions to be included for example). OCSP_basic_add1_cert() adds a
13662 certificate to a basic response and OCSP_basic_sign() signs a basic
13663 response with various flags. New helper functions ASN1_TIME_check()
13664 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13665 (converts ASN1_TIME to GeneralizedTime).
13669 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13670 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13671 structure from a certificate. X509_pubkey_digest() digests the public_key
13672 contents: this is used in various key identifiers.
13676 * Make sk_sort() tolerate a NULL argument.
13678 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13680 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13681 passed by the function are trusted implicitly. If any of them signed the
13682 response then it is assumed to be valid and is not verified.
13686 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13687 to data. This was previously part of the PKCS7 ASN1 code. This
13688 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13689 *Steve Henson, reported by Kenneth R. Robinette
13690 <support@securenetterm.com>*
13692 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13693 routines: without these tracing memory leaks is very painful.
13694 Fix leaks in PKCS12 and PKCS7 routines.
13698 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13699 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13700 effectively meant GeneralizedTime would never be used. Now it
13701 is initialised to -1 but X509_time_adj() now has to check the value
13702 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13703 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13704 *Steve Henson, reported by Kenneth R. Robinette
13705 <support@securenetterm.com>*
13707 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13708 result in a zero length in the ASN1_INTEGER structure which was
13709 not consistent with the structure when d2i_ASN1_INTEGER() was used
13710 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13711 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13712 where it did not print out a minus for negative ASN1_INTEGER.
13716 * Add summary printout to ocsp utility. The various functions which
13717 convert status values to strings have been renamed to:
13718 OCSP_response_status_str(), OCSP_cert_status_str() and
13719 OCSP_crl_reason_str() and are no longer static. New options
13720 to verify nonce values and to disable verification. OCSP response
13721 printout format cleaned up.
13725 * Add additional OCSP certificate checks. These are those specified
13726 in RFC2560. This consists of two separate checks: the CA of the
13727 certificate being checked must either be the OCSP signer certificate
13728 or the issuer of the OCSP signer certificate. In the latter case the
13729 OCSP signer certificate must contain the OCSP signing extended key
13730 usage. This check is performed by attempting to match the OCSP
13731 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13732 in the OCSP_CERTID structures of the response.
13736 * Initial OCSP certificate verification added to OCSP_basic_verify()
13737 and related routines. This uses the standard OpenSSL certificate
13738 verify routines to perform initial checks (just CA validity) and
13739 to obtain the certificate chain. Then additional checks will be
13740 performed on the chain. Currently the root CA is checked to see
13741 if it is explicitly trusted for OCSP signing. This is used to set
13742 a root CA as a global signing root: that is any certificate that
13743 chains to that CA is an acceptable OCSP signing certificate.
13747 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13748 extensions from a separate configuration file.
13749 As when reading extensions from the main configuration file,
13750 the '-extensions ...' option may be used for specifying the
13753 *Massimiliano Pala <madwolf@comune.modena.it>*
13755 * New OCSP utility. Allows OCSP requests to be generated or
13756 read. The request can be sent to a responder and the output
13757 parsed, outputted or printed in text form. Not complete yet:
13758 still needs to check the OCSP response validity.
13762 * New subcommands for 'openssl ca':
13763 `openssl ca -status <serial>` prints the status of the cert with
13764 the given serial number (according to the index file).
13765 `openssl ca -updatedb` updates the expiry status of certificates
13768 *Massimiliano Pala <madwolf@comune.modena.it>*
13770 * New '-newreq-nodes' command option to CA.pl. This is like
13771 '-newreq', but calls 'openssl req' with the '-nodes' option
13772 so that the resulting key is not encrypted.
13774 *Damien Miller <djm@mindrot.org>*
13776 * New configuration for the GNU Hurd.
13778 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13780 * Initial code to implement OCSP basic response verify. This
13781 is currently incomplete. Currently just finds the signer's
13782 certificate and verifies the signature on the response.
13786 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13787 value of OPENSSLDIR. This is available via the new '-d' option
13788 to 'openssl version', and is also included in 'openssl version -a'.
13792 * Allowing defining memory allocation callbacks that will be given
13793 file name and line number information in additional arguments
13794 (a `const char*` and an int). The basic functionality remains, as
13795 well as the original possibility to just replace malloc(),
13796 realloc() and free() by functions that do not know about these
13797 additional arguments. To register and find out the current
13798 settings for extended allocation functions, the following
13799 functions are provided:
13801 CRYPTO_set_mem_ex_functions
13802 CRYPTO_set_locked_mem_ex_functions
13803 CRYPTO_get_mem_ex_functions
13804 CRYPTO_get_locked_mem_ex_functions
13806 These work the same way as CRYPTO_set_mem_functions and friends.
13807 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13808 extended allocation function is enabled.
13809 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13810 a conventional allocation function is enabled.
13812 *Richard Levitte, Bodo Moeller*
13814 * Finish off removing the remaining LHASH function pointer casts.
13815 There should no longer be any prototype-casting required when using
13816 the LHASH abstraction, and any casts that remain are "bugs". See
13817 the callback types and macros at the head of lhash.h for details
13818 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13822 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13823 If /dev/[u]random devices are not available or do not return enough
13824 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13826 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13827 /etc/entropy will be queried once each in this sequence, querying stops
13828 when enough entropy was collected without querying more sockets.
13832 * Change the Unix RAND_poll() variant to be able to poll several
13833 random devices, as specified by DEVRANDOM, until a sufficient amount
13834 of data has been collected. We spend at most 10 ms on each file
13835 (select timeout) and read in non-blocking mode. DEVRANDOM now
13836 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13837 (previously it was just the string "/dev/urandom"), so on typical
13838 platforms the 10 ms delay will never occur.
13839 Also separate out the Unix variant to its own file, rand_unix.c.
13840 For VMS, there's a currently-empty rand_vms.c.
13844 * Move OCSP client related routines to ocsp_cl.c. These
13845 provide utility functions which an application needing
13846 to issue a request to an OCSP responder and analyse the
13847 response will typically need: as opposed to those which an
13848 OCSP responder itself would need which will be added later.
13850 OCSP_request_sign() signs an OCSP request with an API similar
13851 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13852 response. OCSP_response_get1_basic() extracts basic response
13853 from response. OCSP_resp_find_status(): finds and extracts status
13854 information from an OCSP_CERTID structure (which will be created
13855 when the request structure is built). These are built from lower
13856 level functions which work on OCSP_SINGLERESP structures but
13857 won't normally be used unless the application wishes to examine
13858 extensions in the OCSP response for example.
13860 Replace nonce routines with a pair of functions.
13861 OCSP_request_add1_nonce() adds a nonce value and optionally
13862 generates a random value. OCSP_check_nonce() checks the
13863 validity of the nonce in an OCSP response.
13867 * Change function OCSP_request_add() to OCSP_request_add0_id().
13868 This doesn't copy the supplied OCSP_CERTID and avoids the
13869 need to free up the newly created id. Change return type
13870 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13871 This can then be used to add extensions to the request.
13872 Deleted OCSP_request_new(), since most of its functionality
13873 is now in OCSP_REQUEST_new() (and the case insensitive name
13874 clash) apart from the ability to set the request name which
13875 will be added elsewhere.
13879 * Update OCSP API. Remove obsolete extensions argument from
13880 various functions. Extensions are now handled using the new
13881 OCSP extension code. New simple OCSP HTTP function which
13882 can be used to send requests and parse the response.
13886 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13887 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13888 uses the special reorder version of SET OF to sort the attributes
13889 and reorder them to match the encoded order. This resolves a long
13890 standing problem: a verify on a PKCS7 structure just after signing
13891 it used to fail because the attribute order did not match the
13892 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13893 it uses the received order. This is necessary to tolerate some broken
13894 software that does not order SET OF. This is handled by encoding
13895 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13896 to produce the required SET OF.
13900 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13901 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13902 files to get correct declarations of the ASN.1 item variables.
13906 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13907 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13908 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13909 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13910 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13911 ASN1_ITEM and no wrapper functions.
13915 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13916 replace the old function pointer based I/O routines. Change most of
13917 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13921 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13922 lines, recognize more "algorithms" that can be deselected, and make
13923 it complain about algorithm deselection that isn't recognised.
13927 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13928 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13929 to use new functions. Add NO_ASN1_OLD which can be set to remove
13930 some old style ASN1 functions: this can be used to determine if old
13931 code will still work when these eventually go away.
13935 * New extension functions for OCSP structures, these follow the
13936 same conventions as certificates and CRLs.
13940 * New function X509V3_add1_i2d(). This automatically encodes and
13941 adds an extension. Its behaviour can be customised with various
13942 flags to append, replace or delete. Various wrappers added for
13943 certificates and CRLs.
13947 * Fix to avoid calling the underlying ASN1 print routine when
13948 an extension cannot be parsed. Correct a typo in the
13949 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13953 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13954 entries for variables.
13958 * Add functionality to `apps/openssl.c` for detecting locking
13959 problems: As the program is single-threaded, all we have
13960 to do is register a locking callback using an array for
13961 storing which locks are currently held by the program.
13965 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13966 SSL_get_ex_data_X509_STORE_idx(), which is used in
13967 ssl_verify_cert_chain() and thus can be called at any time
13968 during TLS/SSL handshakes so that thread-safety is essential.
13969 Unfortunately, the ex_data design is not at all suited
13970 for multi-threaded use, so it probably should be abolished.
13974 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13976 *Broadcom, tweaked and integrated by Geoff Thorpe*
13978 * Move common extension printing code to new function
13979 X509V3_print_extensions(). Reorganise OCSP print routines and
13980 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13984 * New function X509_signature_print() to remove duplication in some
13989 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13990 set (this was treated exactly the same as SET OF previously). This
13991 is used to reorder the STACK representing the structure to match the
13992 encoding. This will be used to get round a problem where a PKCS7
13993 structure which was signed could not be verified because the STACK
13994 order did not reflect the encoded order.
13998 * Reimplement the OCSP ASN1 module using the new code.
14002 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14003 for its ASN1 operations. The old style function pointers still exist
14004 for now but they will eventually go away.
14008 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14009 completely replaces the old ASN1 functionality with a table driven
14010 encoder and decoder which interprets an ASN1_ITEM structure describing
14011 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14012 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14013 has also been converted to the new form.
14017 * Change BN_mod_exp_recp so that negative moduli are tolerated
14018 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14019 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14020 for negative moduli.
14024 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14025 of not touching the result's sign bit.
14029 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14034 * Changed the LHASH code to use prototypes for callbacks, and created
14035 macros to declare and implement thin (optionally static) functions
14036 that provide type-safety and avoid function pointer casting for the
14037 type-specific callbacks.
14041 * Added Kerberos Cipher Suites to be used with TLS, as written in
14043 *Veers Staats <staatsvr@asc.hpc.mil>,
14044 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14046 * Reformat the FAQ so the different questions and answers can be divided
14047 in sections depending on the subject.
14051 * Have the zlib compression code load ZLIB.DLL dynamically under
14056 * New function BN_mod_sqrt for computing square roots modulo a prime
14057 (using the probabilistic Tonelli-Shanks algorithm unless
14058 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14059 be handled deterministically).
14061 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14063 * Make BN_mod_inverse faster by explicitly handling small quotients
14064 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14065 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14069 * New function BN_kronecker.
14073 * Fix BN_gcd so that it works on negative inputs; the result is
14074 positive unless both parameters are zero.
14075 Previously something reasonably close to an infinite loop was
14076 possible because numbers could be growing instead of shrinking
14077 in the implementation of Euclid's algorithm.
14081 * Fix BN_is_word() and BN_is_one() macros to take into account the
14082 sign of the number in question.
14084 Fix BN_is_word(a,w) to work correctly for w == 0.
14086 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14087 because its test if the absolute value of 'a' equals 'w'.
14088 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14089 it exists mostly for use in the implementations of BN_is_zero(),
14090 BN_is_one(), and BN_is_word().
14094 * New function BN_swap.
14098 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14099 the exponentiation functions are more likely to produce reasonable
14100 results on negative inputs.
14104 * Change BN_mod_mul so that the result is always non-negative.
14105 Previously, it could be negative if one of the factors was negative;
14106 I don't think anyone really wanted that behaviour.
14110 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14111 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14112 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14113 and add new functions:
14122 BN_mod_lshift1_quick
14124 BN_mod_lshift_quick
14126 These functions always generate non-negative results.
14128 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14129 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14131 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14132 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14133 be reduced modulo `m`.
14135 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14138 The following entry accidentally appeared in the CHANGES file
14139 distributed with OpenSSL 0.9.7. The modifications described in
14140 it do *not* apply to OpenSSL 0.9.7.
14142 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14143 was actually never needed) and in BN_mul(). The removal in BN_mul()
14144 required a small change in bn_mul_part_recursive() and the addition
14145 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14146 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14147 bn_sub_words() and bn_add_words() except they take arrays with
14153 * In 'openssl passwd', verify passwords read from the terminal
14154 unless the '-salt' option is used (which usually means that
14155 verification would just waste user's time since the resulting
14156 hash is going to be compared with some given password hash)
14157 or the new '-noverify' option is used.
14159 This is an incompatible change, but it does not affect
14160 non-interactive use of 'openssl passwd' (passwords on the command
14161 line, '-stdin' option, '-in ...' option) and thus should not
14162 cause any problems.
14166 * Remove all references to RSAref, since there's no more need for it.
14170 * Make DSO load along a path given through an environment variable
14171 (SHLIB_PATH) with shl_load().
14175 * Constify the ENGINE code as a result of BIGNUM constification.
14176 Also constify the RSA code and most things related to it. In a
14177 few places, most notable in the depth of the ASN.1 code, ugly
14178 casts back to non-const were required (to be solved at a later
14183 * Make it so the openssl application has all engines loaded by default.
14187 * Constify the BIGNUM routines a little more.
14191 * Add the following functions:
14193 ENGINE_load_cswift()
14195 ENGINE_load_atalla()
14196 ENGINE_load_nuron()
14197 ENGINE_load_builtin_engines()
14199 That way, an application can itself choose if external engines that
14200 are built-in in OpenSSL shall ever be used or not. The benefit is
14201 that applications won't have to be linked with libdl or other dso
14202 libraries unless it's really needed.
14204 Changed 'openssl engine' to load all engines on demand.
14205 Changed the engine header files to avoid the duplication of some
14206 declarations (they differed!).
14210 * 'openssl engine' can now list capabilities.
14214 * Better error reporting in 'openssl engine'.
14218 * Never call load_dh_param(NULL) in s_server.
14222 * Add engine application. It can currently list engines by name and
14223 identity, and test if they are actually available.
14227 * Improve RPM specification file by forcing symbolic linking and making
14228 sure the installed documentation is also owned by root.root.
14230 *Damien Miller <djm@mindrot.org>*
14232 * Give the OpenSSL applications more possibilities to make use of
14233 keys (public as well as private) handled by engines.
14237 * Add OCSP code that comes from CertCo.
14241 * Add VMS support for the Rijndael code.
14245 * Added untested support for Nuron crypto accelerator.
14249 * Add support for external cryptographic devices. This code was
14250 previously distributed separately as the "engine" branch.
14252 *Geoff Thorpe, Richard Levitte*
14254 * Rework the filename-translation in the DSO code. It is now possible to
14255 have far greater control over how a "name" is turned into a filename
14256 depending on the operating environment and any oddities about the
14257 different shared library filenames on each system.
14261 * Support threads on FreeBSD-elf in Configure.
14265 * Fix for SHA1 assembly problem with MASM: it produces
14266 warnings about corrupt line number information when assembling
14267 with debugging information. This is caused by the overlapping
14270 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14273 NCONF_get_number() has no error checking at all. As a replacement,
14274 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14275 promoted strongly. The old NCONF_get_number is kept around for
14276 binary backward compatibility.
14277 Make it possible for methods to load from something other than a BIO,
14278 by providing a function pointer that is given a name instead of a BIO.
14279 For example, this could be used to load configuration data from an
14284 * Fix for non blocking accept BIOs. Added new I/O special reason
14285 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14286 with non blocking I/O was not possible because no retry code was
14287 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14292 * Added the beginnings of Rijndael support.
14296 * Fix for bug in DirectoryString mask setting. Add support for
14297 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14298 to allow certificate printing to more controllable, additional
14299 'certopt' option to 'x509' to allow new printing options to be
14304 * Clean old EAY MD5 hack from e_os.h.
14308 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14310 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14311 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14313 *Joe Orton, Steve Henson*
14315 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14317 * Fix additional bug revealed by the NISCC test suite:
14319 Stop bug triggering large recursion when presented with
14320 certain ASN.1 tags ([CVE-2003-0851])
14324 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14326 * Fix various bugs revealed by running the NISCC test suite:
14328 Stop out of bounds reads in the ASN1 code when presented with
14329 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14331 If verify callback ignores invalid public key errors don't try to check
14332 certificate signature with the NULL public key.
14336 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14337 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14342 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14343 extra data after the compression methods not only for TLS 1.0
14344 but also for SSL 3.0 (as required by the specification).
14346 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14348 * Change X509_certificate_type() to mark the key as exported/exportable
14349 when it's 512 *bits* long, not 512 bytes.
14353 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14355 * Countermeasure against the Klima-Pokorny-Rosa extension of
14356 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14357 a protocol version number mismatch like a decryption error
14358 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14362 * Turn on RSA blinding by default in the default implementation
14363 to avoid a timing attack. Applications that don't want it can call
14364 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14365 They would be ill-advised to do so in most cases.
14367 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14369 * Change RSA blinding code so that it works when the PRNG is not
14370 seeded (in this case, the secret RSA exponent is abused as
14371 an unpredictable seed -- if it is not unpredictable, there
14372 is no point in blinding anyway). Make RSA blinding thread-safe
14373 by remembering the creator's thread ID in rsa->blinding and
14374 having all other threads use local one-time blinding factors
14375 (this requires more computation than sharing rsa->blinding, but
14376 avoids excessive locking; and if an RSA object is not shared
14377 between threads, blinding will still be very fast).
14381 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14383 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14384 via timing by performing a MAC computation even if incorrect
14385 block cipher padding has been found. This is a countermeasure
14386 against active attacks where the attacker has to distinguish
14387 between bad padding and a MAC verification error. ([CVE-2003-0078])
14389 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14390 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14391 Martin Vuagnoux (EPFL, Ilion)*
14393 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14395 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14396 memory from its contents. This is done with a counter that will
14397 place alternating values in each byte. This can be used to solve
14398 two issues: 1) the removal of calls to memset() by highly optimizing
14399 compilers, and 2) cleansing with other values than 0, since those can
14400 be read through on certain media, for example a swap space on disk.
14404 * Bugfix: client side session caching did not work with external caching,
14405 because the session->cipher setting was not restored when reloading
14406 from the external cache. This problem was masked, when
14407 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14408 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14412 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14413 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14415 *Zeev Lieber <zeev-l@yahoo.com>*
14417 * Undo an undocumented change introduced in 0.9.6e which caused
14418 repeated calls to OpenSSL_add_all_ciphers() and
14419 OpenSSL_add_all_digests() to be ignored, even after calling
14424 * Change the default configuration reader to deal with last line not
14425 being properly terminated.
14429 * Change X509_NAME_cmp() so it applies the special rules on handling
14430 DN values that are of type PrintableString, as well as RDNs of type
14431 emailAddress where the value has the type ia5String.
14433 *stefank@valicert.com via Richard Levitte*
14435 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14436 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14437 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14438 the bitwise-OR of the two for use by the majority of applications
14439 wanting this behaviour, and update the docs. The documented
14440 behaviour and actual behaviour were inconsistent and had been
14441 changing anyway, so this is more a bug-fix than a behavioural
14444 *Geoff Thorpe, diagnosed by Nadav Har'El*
14446 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14447 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14451 * Fix initialization code race conditions in
14452 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14453 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14454 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14455 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14456 ssl2_get_cipher_by_char(),
14457 ssl3_get_cipher_by_char().
14459 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14461 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14462 the cached sessions are flushed, as the remove_cb() might use ex_data
14463 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14464 (see [openssl.org #212]).
14466 *Geoff Thorpe, Lutz Jaenicke*
14468 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14469 length, instead of the encoding length to d2i_ASN1_OBJECT.
14473 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14475 * [In 0.9.6g-engine release:]
14476 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14478 *Lynn Gazis <lgazis@rainbow.com>*
14480 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14482 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14483 and get fix the header length calculation.
14484 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14485 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14487 * Use proper error handling instead of 'assertions' in buffer
14488 overflow checks added in 0.9.6e. This prevents DoS (the
14489 assertions could call abort()).
14491 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14493 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14495 * Add various sanity checks to asn1_get_length() to reject
14496 the ASN1 length bytes if they exceed sizeof(long), will appear
14497 negative or the content length exceeds the length of the
14500 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14502 * Fix cipher selection routines: ciphers without encryption had no flags
14503 for the cipher strength set and where therefore not handled correctly
14504 by the selection routines (PR #130).
14508 * Fix EVP_dsa_sha macro.
14513 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14514 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14515 that was added in OpenSSL 0.9.6d.
14517 As the countermeasure turned out to be incompatible with some
14518 broken SSL implementations, the new option is part of SSL_OP_ALL.
14519 SSL_OP_ALL is usually employed when compatibility with weird SSL
14520 implementations is desired (e.g. '-bugs' option to 's_client' and
14521 's_server'), so the new option is automatically set in many
14526 * Changes in security patch:
14528 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14529 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14530 Air Force Materiel Command, USAF, under agreement number
14533 * Add various sanity checks to asn1_get_length() to reject
14534 the ASN1 length bytes if they exceed sizeof(long), will appear
14535 negative or the content length exceeds the length of the
14536 supplied buffer. ([CVE-2002-0659])
14538 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14540 * Assertions for various potential buffer overflows, not known to
14541 happen in practice.
14543 *Ben Laurie (CHATS)*
14545 * Various temporary buffers to hold ASCII versions of integers were
14546 too small for 64 bit platforms. ([CVE-2002-0655])
14547 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14549 * Remote buffer overflow in SSL3 protocol - an attacker could
14550 supply an oversized session ID to a client. ([CVE-2002-0656])
14552 *Ben Laurie (CHATS)*
14554 * Remote buffer overflow in SSL2 protocol - an attacker could
14555 supply an oversized client master key. ([CVE-2002-0656])
14557 *Ben Laurie (CHATS)*
14559 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14561 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14562 encoded as NULL) with id-dsa-with-sha1.
14564 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14566 * Check various `X509_...()` return values in `apps/req.c`.
14568 *Nils Larsch <nla@trustcenter.de>*
14570 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14571 an end-of-file condition would erroneously be flagged, when the CRLF
14572 was just at the end of a processed block. The bug was discovered when
14573 processing data through a buffering memory BIO handing the data to a
14574 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14575 <ptsekov@syntrex.com> and Nedelcho Stanev.
14579 * Implement a countermeasure against a vulnerability recently found
14580 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14581 before application data chunks to avoid the use of known IVs
14582 with data potentially chosen by the attacker.
14586 * Fix length checks in ssl3_get_client_hello().
14590 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14591 to prevent ssl3_read_internal() from incorrectly assuming that
14592 ssl3_read_bytes() found application data while handshake
14593 processing was enabled when in fact s->s3->in_read_app_data was
14594 merely automatically cleared during the initial handshake.
14596 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14598 * Fix object definitions for Private and Enterprise: they were not
14599 recognized in their shortname (=lowercase) representation. Extend
14600 obj_dat.pl to issue an error when using undefined keywords instead
14601 of silently ignoring the problem (Svenning Sorensen
14602 <sss@sss.dnsalias.net>).
14606 * Fix DH_generate_parameters() so that it works for 'non-standard'
14607 generators, i.e. generators other than 2 and 5. (Previously, the
14608 code did not properly initialise the 'add' and 'rem' values to
14609 BN_generate_prime().)
14611 In the new general case, we do not insist that 'generator' is
14612 actually a primitive root: This requirement is rather pointless;
14613 a generator of the order-q subgroup is just as good, if not
14618 * Map new X509 verification errors to alerts. Discovered and submitted by
14619 Tom Wu <tom@arcot.com>.
14623 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14624 returning non-zero before the data has been completely received
14625 when using non-blocking I/O.
14627 *Bodo Moeller; problem pointed out by John Hughes*
14629 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14631 *Ben Laurie, Lutz Jaenicke*
14633 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14634 Yoram Zahavi <YoramZ@gilian.com>).
14638 * Add information about CygWin 1.3 and on, and preserve proper
14639 configuration for the versions before that.
14641 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14643 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14644 check whether we deal with a copy of a session and do not delete from
14645 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14646 <izhar@checkpoint.com>.
14650 * Do not store session data into the internal session cache, if it
14651 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14652 flag is set). Proposed by Aslam <aslam@funk.com>.
14656 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14661 * [In 0.9.6d-engine release:]
14662 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14664 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14666 * Add the configuration target linux-s390x.
14668 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14670 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14671 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14672 variable as an indication that a ClientHello message has been
14673 received. As the flag value will be lost between multiple
14674 invocations of ssl3_accept when using non-blocking I/O, the
14675 function may not be aware that a handshake has actually taken
14676 place, thus preventing a new session from being added to the
14679 To avoid this problem, we now set s->new_session to 2 instead of
14680 using a local variable.
14682 *Lutz Jaenicke, Bodo Moeller*
14684 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14685 if the SSL_R_LENGTH_MISMATCH error is detected.
14687 *Geoff Thorpe, Bodo Moeller*
14689 * New 'shared_ldflag' column in Configure platform table.
14693 * Fix EVP_CIPHER_mode macro.
14695 *"Dan S. Camper" <dan@bti.net>*
14697 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14698 type, we must throw them away by setting rr->length to 0.
14700 *D P Chang <dpc@qualys.com>*
14702 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14704 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14705 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14706 worked incorrectly for those cases where range = `10..._2` and
14707 `3*range` is two bits longer than range.)
14711 * Only add signing time to PKCS7 structures if it is not already
14716 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14717 OBJ_ld_ce should be OBJ_id_ce.
14718 Also some ip-pda OIDs in crypto/objects/objects.txt were
14719 incorrect (cf. RFC 3039).
14721 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14723 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14724 returns early because it has nothing to do.
14726 *Andy Schneider <andy.schneider@bjss.co.uk>*
14728 * [In 0.9.6c-engine release:]
14729 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14731 *Andy Schneider <andy.schneider@bjss.co.uk>*
14733 * [In 0.9.6c-engine release:]
14734 Add support for Cryptographic Appliance's keyserver technology.
14735 (Use engine 'keyclient')
14737 *Cryptographic Appliances and Geoff Thorpe*
14739 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14740 is called via tools/c89.sh because arguments have to be
14741 rearranged (all '-L' options must appear before the first object
14744 *Richard Shapiro <rshapiro@abinitio.com>*
14746 * [In 0.9.6c-engine release:]
14747 Add support for Broadcom crypto accelerator cards, backported
14750 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14752 * [In 0.9.6c-engine release:]
14753 Add support for SureWare crypto accelerator cards from
14754 Baltimore Technologies. (Use engine 'sureware')
14756 *Baltimore Technologies and Mark Cox*
14758 * [In 0.9.6c-engine release:]
14759 Add support for crypto accelerator cards from Accelerated
14760 Encryption Processing, www.aep.ie. (Use engine 'aep')
14762 *AEP Inc. and Mark Cox*
14764 * Add a configuration entry for gcc on UnixWare.
14766 *Gary Benson <gbenson@redhat.com>*
14768 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14769 messages are stored in a single piece (fixed-length part and
14770 variable-length part combined) and fix various bugs found on the way.
14774 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14775 instead. BIO_gethostbyname() does not know what timeouts are
14776 appropriate, so entries would stay in cache even when they have
14778 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14780 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14781 faced with a pathologically small ClientHello fragment that does
14782 not contain client_version: Instead of aborting with an error,
14783 simply choose the highest available protocol version (i.e.,
14784 TLS 1.0 unless it is disabled). In practice, ClientHello
14785 messages are never sent like this, but this change gives us
14786 strictly correct behaviour at least for TLS.
14790 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14791 never resets s->method to s->ctx->method when called from within
14792 one of the SSL handshake functions.
14794 *Bodo Moeller; problem pointed out by Niko Baric*
14796 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14797 (sent using the client's version number) if client_version is
14798 smaller than the protocol version in use. Also change
14799 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14800 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14801 the client will at least see that alert.
14805 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14810 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14811 client receives HelloRequest while in a handshake.
14813 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14815 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14816 should end in 'break', not 'goto end' which circumvents various
14817 cleanups done in state SSL_ST_OK. But session related stuff
14818 must be disabled for SSL_ST_OK in the case that we just sent a
14821 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14822 before just sending a HelloRequest.
14824 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14826 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14827 reveal whether illegal block cipher padding was found or a MAC
14828 verification error occurred. (Neither SSLerr() codes nor alerts
14829 are directly visible to potential attackers, but the information
14830 may leak via logfiles.)
14832 Similar changes are not required for the SSL 2.0 implementation
14833 because the number of padding bytes is sent in clear for SSL 2.0,
14834 and the extra bytes are just ignored. However ssl/s2_pkt.c
14835 failed to verify that the purported number of padding bytes is in
14840 * Add OpenUNIX-8 support including shared libraries
14841 (Boyd Lynn Gerber <gerberb@zenez.com>).
14845 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14846 'wristwatch attack' using huge encoding parameters (cf.
14847 James H. Manger's CRYPTO 2001 paper). Note that the
14848 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14849 encoding parameters and hence was not vulnerable.
14853 * BN_sqr() bug fix.
14855 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14857 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14858 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14859 followed by modular reduction.
14861 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14863 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14864 equivalent based on BN_pseudo_rand() instead of BN_rand().
14868 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14869 This function was broken, as the check for a new client hello message
14870 to handle SGC did not allow these large messages.
14871 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14875 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14879 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14880 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14884 * Rework the configuration and shared library support for Tru64 Unix.
14885 The configuration part makes use of modern compiler features and
14886 still retains old compiler behavior for those that run older versions
14887 of the OS. The shared library support part includes a variant that
14888 uses the RPATH feature, and is available through the special
14889 configuration target "alpha-cc-rpath", which will never be selected
14892 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14894 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14895 with the same message size as in ssl3_get_certificate_request().
14896 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14897 messages might inadvertently be reject as too long.
14899 *Petr Lampa <lampa@fee.vutbr.cz>*
14901 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14905 * Modified SSL library such that the verify_callback that has been set
14906 specifically for an SSL object with SSL_set_verify() is actually being
14907 used. Before the change, a verify_callback set with this function was
14908 ignored and the verify_callback() set in the SSL_CTX at the time of
14909 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14910 to allow the necessary settings.
14914 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14915 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14916 done automatically (in contradiction to the requirements of the C
14917 standard). This made problems when used from OpenSSH.
14921 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14922 dh->length and always used
14924 BN_rand_range(priv_key, dh->p).
14926 BN_rand_range() is not necessary for Diffie-Hellman, and this
14927 specific range makes Diffie-Hellman unnecessarily inefficient if
14928 dh->length (recommended exponent length) is much smaller than the
14929 length of dh->p. We could use BN_rand_range() if the order of
14930 the subgroup was stored in the DH structure, but we only have
14935 BN_rand(priv_key, l, ...)
14937 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14944 RSA_eay_public_encrypt
14945 RSA_eay_private_decrypt
14946 RSA_eay_private_encrypt (signing)
14947 RSA_eay_public_decrypt (signature verification)
14949 (default implementations for RSA_public_encrypt,
14950 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14951 always reject numbers >= n.
14955 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14956 to synchronize access to 'locking_thread'. This is necessary on
14957 systems where access to 'locking_thread' (an 'unsigned long'
14958 variable) is not atomic.
14962 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14963 *before* setting the 'crypto_lock_rand' flag. The previous code had
14964 a race condition if 0 is a valid thread ID.
14966 *Travis Vitek <vitek@roguewave.com>*
14968 * Add support for shared libraries under Irix.
14970 *Albert Chin-A-Young <china@thewrittenword.com>*
14972 * Add configuration option to build on Linux on both big-endian and
14973 little-endian MIPS.
14975 *Ralf Baechle <ralf@uni-koblenz.de>*
14977 * Add the possibility to create shared libraries on HP-UX.
14981 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14983 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14984 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14985 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14986 PRNG state recovery was possible based on the output of
14987 one PRNG request appropriately sized to gain knowledge on
14988 'md' followed by enough consecutive 1-byte PRNG requests
14989 to traverse all of 'state'.
14991 1. When updating 'md_local' (the current thread's copy of 'md')
14992 during PRNG output generation, hash all of the previous
14993 'md_local' value, not just the half used for PRNG output.
14995 2. Make the number of bytes from 'state' included into the hash
14996 independent from the number of PRNG bytes requested.
14998 The first measure alone would be sufficient to avoid
14999 Markku-Juhani's attack. (Actually it had never occurred
15000 to me that the half of 'md_local' used for chaining was the
15001 half from which PRNG output bytes were taken -- I had always
15002 assumed that the secret half would be used.) The second
15003 measure makes sure that additional data from 'state' is never
15004 mixed into 'md_local' in small portions; this heuristically
15005 further strengthens the PRNG.
15009 * Fix crypto/bn/asm/mips3.s.
15013 * When only the key is given to "enc", the IV is undefined. Print out
15014 an error message in this case.
15018 * Handle special case when X509_NAME is empty in X509 printing routines.
15022 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15023 positive and less than q.
15027 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15028 used: it isn't thread safe and the add_lock_callback should handle
15031 *Paul Rose <Paul.Rose@bridge.com>*
15033 * Verify that incoming data obeys the block size in
15034 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15040 *Ulf Möller, Bodo Möller*
15042 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15043 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15044 when fixing the server behaviour for backwards-compatible 'client
15045 hello' messages. (Note that the attack is impractical against
15046 SSL 3.0 and TLS 1.0 anyway because length and version checking
15047 means that the probability of guessing a valid ciphertext is
15048 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15051 Before 0.9.5, the countermeasure (hide the error by generating a
15052 random 'decryption result') did not work properly because
15053 ERR_clear_error() was missing, meaning that SSL_get_error() would
15054 detect the supposedly ignored error.
15056 Both problems are now fixed.
15060 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15061 (previously it was 1024).
15065 * Fix for compatibility mode trust settings: ignore trust settings
15066 unless some valid trust or reject settings are present.
15070 * Fix for blowfish EVP: its a variable length cipher.
15074 * Fix various bugs related to DSA S/MIME verification. Handle missing
15075 parameters in DSA public key structures and return an error in the
15076 DSA routines if parameters are absent.
15080 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15081 in the current directory if neither $RANDFILE nor $HOME was set.
15082 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15083 caused some confusion to Windows users who haven't defined $HOME.
15084 Thus RAND_file_name() is changed again: e_os.h can define a
15085 DEFAULT_HOME, which will be used if $HOME is not set.
15086 For Windows, we use "C:"; on other platforms, we still require
15087 environment variables.
15089 * Move 'if (!initialized) RAND_poll()' into regions protected by
15090 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15091 having multiple threads call RAND_poll() concurrently.
15095 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15096 combination of a flag and a thread ID variable.
15097 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15098 flag), *other* threads can enter ssleay_add_bytes without obeying
15099 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15100 that they do not hold after the first thread unsets add_do_not_lock).
15104 * Change bctest again: '-x' expressions are not available in all
15105 versions of 'test'.
15109 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15111 * Fix a couple of memory leaks in PKCS7_dataDecode()
15113 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15115 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15116 the default extension for executables, if any. Also, make the perl
15117 scripts that use symlink() to test if it really exists and use "cp"
15118 if it doesn't. All this made OpenSSL compilable and installable in
15123 * Fix for asn1_GetSequence() for indefinite length constructed data.
15124 If SEQUENCE is length is indefinite just set c->slen to the total
15125 amount of data available.
15127 *Steve Henson, reported by shige@FreeBSD.org*
15129 *This change does not apply to 0.9.7.*
15131 * Change bctest to avoid here-documents inside command substitution
15132 (workaround for FreeBSD /bin/sh bug).
15133 For compatibility with Ultrix, avoid shell functions (introduced
15134 in the bctest version that searches along $PATH).
15138 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15139 with des_encrypt() defined on some operating systems, like Solaris
15144 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15145 On the Importance of Eliminating Errors in Cryptographic
15146 Computations, J. Cryptology 14 (2001) 2, 101-119,
15147 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15151 * MIPS assembler BIGNUM division bug fix.
15155 * Disabled incorrect Alpha assembler code.
15159 * Fix PKCS#7 decode routines so they correctly update the length
15160 after reading an EOC for the EXPLICIT tag.
15164 *This change does not apply to 0.9.7.*
15166 * Fix bug in PKCS#12 key generation routines. This was triggered
15167 if a 3DES key was generated with a 0 initial byte. Include
15168 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15169 (but broken) behaviour.
15173 * Enhance bctest to search for a working bc along $PATH and print
15176 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15178 * Fix memory leaks in err.c: free err_data string if necessary;
15179 don't write to the wrong index in ERR_set_error_data.
15183 * Implement ssl23_peek (analogous to ssl23_read), which previously
15188 * Replace rdtsc with `_emit` statements for VC++ version 5.
15190 *Jeremy Cooper <jeremy@baymoo.org>*
15192 * Make it possible to reuse SSLv2 sessions.
15196 * In copy_email() check for >= 0 as a return value for
15197 X509_NAME_get_index_by_NID() since 0 is a valid index.
15199 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15201 * Avoid coredump with unsupported or invalid public keys by checking if
15202 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15203 PKCS7_verify() fails with non detached data.
15207 * Don't use getenv in library functions when run as setuid/setgid.
15208 New function OPENSSL_issetugid().
15212 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15213 due to incorrect handling of multi-threading:
15215 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15217 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15219 3. Count how many times MemCheck_off() has been called so that
15220 nested use can be treated correctly. This also avoids
15221 inband-signalling in the previous code (which relied on the
15222 assumption that thread ID 0 is impossible).
15226 * Add "-rand" option also to s_client and s_server.
15230 * Fix CPU detection on Irix 6.x.
15231 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15232 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15234 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15239 *This change does not apply to 0.9.7.*
15241 * Use the cached encoding of an X509_NAME structure rather than
15242 copying it. This is apparently the reason for the libsafe "errors"
15243 but the code is actually correct.
15247 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15248 Bleichenbacher's DSA attack.
15249 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15250 to be set and top=0 forces the highest bit to be set; top=-1 is new
15251 and leaves the highest bit random.
15253 *Ulf Moeller, Bodo Moeller*
15255 * In the `NCONF_...`-based implementations for `CONF_...` queries
15256 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15257 a temporary CONF structure with the data component set to NULL
15258 (which gives segmentation faults in lh_retrieve).
15259 Instead, use NULL for the CONF pointer in CONF_get_string and
15260 CONF_get_number (which may use environment variables) and directly
15261 return NULL from CONF_get_section.
15265 * Fix potential buffer overrun for EBCDIC.
15269 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15270 keyUsage if basicConstraints absent for a CA.
15274 * Make SMIME_write_PKCS7() write mail header values with a format that
15275 is more generally accepted (no spaces before the semicolon), since
15276 some programs can't parse those values properly otherwise. Also make
15277 sure BIO's that break lines after each write do not create invalid
15282 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15283 macros previously used would not encode an empty SEQUENCE OF
15284 and break the signature.
15288 *This change does not apply to 0.9.7.*
15290 * Zero the premaster secret after deriving the master secret in
15295 * Add some EVP_add_digest_alias registrations (as found in
15296 OpenSSL_add_all_digests()) to SSL_library_init()
15297 aka OpenSSL_add_ssl_algorithms(). This provides improved
15298 compatibility with peers using X.509 certificates
15299 with unconventional AlgorithmIdentifier OIDs.
15303 * Fix for Irix with NO_ASM.
15305 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15307 * ./config script fixes.
15309 *Ulf Moeller, Richard Levitte*
15311 * Fix 'openssl passwd -1'.
15315 * Change PKCS12_key_gen_asc() so it can cope with non null
15316 terminated strings whose length is passed in the passlen
15317 parameter, for example from PEM callbacks. This was done
15318 by adding an extra length parameter to asc2uni().
15320 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15322 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15323 call failed, free the DSA structure.
15327 * Fix to uni2asc() to cope with zero length Unicode strings.
15328 These are present in some PKCS#12 files.
15332 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15333 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15334 when writing a 32767 byte record.
15336 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15338 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15339 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15341 (RSA objects have a reference count access to which is protected
15342 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15343 so they are meant to be shared between threads.)
15344 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15345 "Reddie, Steven" <Steven.Reddie@ca.com>*
15347 * Fix a deadlock in CRYPTO_mem_leaks().
15351 * Use better test patterns in bntest.
15355 * rand_win.c fix for Borland C.
15359 * BN_rshift bugfix for n == 0.
15363 * Add a 'bctest' script that checks for some known 'bc' bugs
15364 so that 'make test' does not abort just because 'bc' is broken.
15368 * Store verify_result within SSL_SESSION also for client side to
15369 avoid potential security hole. (Re-used sessions on the client side
15370 always resulted in verify_result==X509_V_OK, not using the original
15371 result of the server certificate verification.)
15375 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15376 SSL3_RT_APPLICATION_DATA, return 0.
15377 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15382 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15383 releases, have been re-implemented by renaming the previous
15384 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15385 and ssl3_read_internal, respectively, and adding 'peek' parameters
15386 to them. The new ssl[23]_{read,peek} functions are calls to
15387 ssl[23]_read_internal with the 'peek' flag set appropriately.
15388 A 'peek' parameter has also been added to ssl3_read_bytes, which
15389 does the actual work for ssl3_read_internal.
15393 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15394 the method-specific "init()" handler. Also clean up ex_data after
15395 calling the method-specific "finish()" handler. Previously, this was
15396 happening the other way round.
15400 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15401 The previous value, 12, was not always sufficient for BN_mod_exp().
15405 * Make sure that shared libraries get the internal name engine with
15406 the full version number and not just 0. This should mark the
15407 shared libraries as not backward compatible. Of course, this should
15408 be changed again when we can guarantee backward binary compatibility.
15412 * Fix typo in get_cert_by_subject() in by_dir.c
15414 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15416 * Rework the system to generate shared libraries:
15418 - Make note of the expected extension for the shared libraries and
15419 if there is a need for symbolic links from for example libcrypto.so.0
15420 to libcrypto.so.0.9.7. There is extended info in Configure for
15423 - Make as few rebuilds of the shared libraries as possible.
15425 - Still avoid linking the OpenSSL programs with the shared libraries.
15427 - When installing, install the shared libraries separately from the
15432 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15434 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15435 and not in SSL_clear because the latter is also used by the
15436 accept/connect functions; previously, the settings made by
15437 SSL_set_read_ahead would be lost during the handshake.
15439 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15441 * Correct util/mkdef.pl to be selective about disabled algorithms.
15442 Previously, it would create entries for disabled algorithms no
15447 * Added several new manual pages for SSL_* function.
15451 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15453 * In ssl23_get_client_hello, generate an error message when faced
15454 with an initial SSL 3.0/TLS record that is too small to contain the
15455 first two bytes of the ClientHello message, i.e. client_version.
15456 (Note that this is a pathologic case that probably has never happened
15457 in real life.) The previous approach was to use the version number
15458 from the record header as a substitute; but our protocol choice
15459 should not depend on that one because it is not authenticated
15460 by the Finished messages.
15464 * More robust randomness gathering functions for Windows.
15466 *Jeffrey Altman <jaltman@columbia.edu>*
15468 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15469 not set then we don't setup the error code for issuer check errors
15470 to avoid possibly overwriting other errors which the callback does
15471 handle. If an application does set the flag then we assume it knows
15472 what it is doing and can handle the new informational codes
15477 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15478 a general "ANY" type, as such it should be able to decode anything
15479 including tagged types. However it didn't check the class so it would
15480 wrongly interpret tagged types in the same way as their universal
15481 counterpart and unknown types were just rejected. Changed so that the
15482 tagged and unknown types are handled in the same way as a SEQUENCE:
15483 that is the encoding is stored intact. There is also a new type
15484 "V_ASN1_OTHER" which is used when the class is not universal, in this
15485 case we have no idea what the actual type is so we just lump them all
15490 * On VMS, stdout may very well lead to a file that is written to
15491 in a record-oriented fashion. That means that every write() will
15492 write a separate record, which will be read separately by the
15493 programs trying to read from it. This can be very confusing.
15495 The solution is to put a BIO filter in the way that will buffer
15496 text until a linefeed is reached, and then write everything a
15497 line at a time, so every record written will be an actual line,
15498 not chunks of lines and not (usually doesn't happen, but I've
15499 seen it once) several lines in one record. BIO_f_linebuffer() is
15502 Currently, it's a VMS-only method, because that's where it has
15503 been tested well enough.
15507 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15508 it can return incorrect results.
15509 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15510 but it was in 0.9.6-beta[12].)
15514 * Disable the check for content being present when verifying detached
15515 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15516 include zero length content when signing messages.
15520 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15521 BIO_ctrl (for BIO pairs).
15525 * Add DSO method for VMS.
15529 * Bug fix: Montgomery multiplication could produce results with the
15534 * Add RPM specification openssl.spec and modify it to build three
15535 packages. The default package contains applications, application
15536 documentation and run-time libraries. The devel package contains
15537 include files, static libraries and function documentation. The
15538 doc package contains the contents of the doc directory. The original
15539 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15543 * Add a large number of documentation files for many SSL routines.
15545 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15547 * Add a configuration entry for Sony News 4.
15549 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15551 * Don't set the two most significant bits to one when generating a
15552 random number < q in the DSA library.
15556 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15557 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15558 the underlying transport is blocking) if a handshake took place.
15559 (The default behaviour is needed by applications such as s_client
15560 and s_server that use select() to determine when to use SSL_read;
15561 but for applications that know in advance when to expect data, it
15562 just makes things more complicated.)
15566 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15571 * Add a few more EBCDIC conditionals that make `req` and `x509`
15572 work better on such systems.
15574 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15576 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15577 Update PKCS12_parse() so it copies the friendlyName and the
15578 keyid to the certificates aux info.
15582 * Fix bug in PKCS7_verify() which caused an infinite loop
15583 if there was more than one signature.
15585 *Sven Uszpelkat <su@celocom.de>*
15587 * Major change in util/mkdef.pl to include extra information
15588 about each symbol, as well as presenting variables as well
15589 as functions. This change means that there's n more need
15590 to rebuild the .num files when some algorithms are excluded.
15594 * Allow the verify time to be set by an application,
15595 rather than always using the current time.
15599 * Phase 2 verify code reorganisation. The certificate
15600 verify code now looks up an issuer certificate by a
15601 number of criteria: subject name, authority key id
15602 and key usage. It also verifies self signed certificates
15603 by the same criteria. The main comparison function is
15604 X509_check_issued() which performs these checks.
15606 Lot of changes were necessary in order to support this
15607 without completely rewriting the lookup code.
15609 Authority and subject key identifier are now cached.
15611 The LHASH 'certs' is X509_STORE has now been replaced
15612 by a STACK_OF(X509_OBJECT). This is mainly because an
15613 LHASH can't store or retrieve multiple objects with
15614 the same hash value.
15616 As a result various functions (which were all internal
15617 use only) have changed to handle the new X509_STORE
15618 structure. This will break anything that messed round
15619 with X509_STORE internally.
15621 The functions X509_STORE_add_cert() now checks for an
15622 exact match, rather than just subject name.
15624 The X509_STORE API doesn't directly support the retrieval
15625 of multiple certificates matching a given criteria, however
15626 this can be worked round by performing a lookup first
15627 (which will fill the cache with candidate certificates)
15628 and then examining the cache for matches. This is probably
15629 the best we can do without throwing out X509_LOOKUP
15630 entirely (maybe later...).
15632 The X509_VERIFY_CTX structure has been enhanced considerably.
15634 All certificate lookup operations now go via a get_issuer()
15635 callback. Although this currently uses an X509_STORE it
15636 can be replaced by custom lookups. This is a simple way
15637 to bypass the X509_STORE hackery necessary to make this
15638 work and makes it possible to use more efficient techniques
15639 in future. A very simple version which uses a simple
15640 STACK for its trusted certificate store is also provided
15641 using X509_STORE_CTX_trusted_stack().
15643 The verify_cb() and verify() callbacks now have equivalents
15644 in the X509_STORE_CTX structure.
15646 X509_STORE_CTX also has a 'flags' field which can be used
15647 to customise the verify behaviour.
15651 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15652 excludes S/MIME capabilities.
15656 * When a certificate request is read in keep a copy of the
15657 original encoding of the signed data and use it when outputting
15658 again. Signatures then use the original encoding rather than
15659 a decoded, encoded version which may cause problems if the
15660 request is improperly encoded.
15664 * For consistency with other BIO_puts implementations, call
15665 buffer_write(b, ...) directly in buffer_puts instead of calling
15668 In BIO_puts, increment b->num_write as in BIO_write.
15670 *Peter.Sylvester@EdelWeb.fr*
15672 * Fix BN_mul_word for the case where the word is 0. (We have to use
15673 BN_zero, we may not return a BIGNUM with an array consisting of
15674 words set to zero.)
15678 * Avoid calling abort() from within the library when problems are
15679 detected, except if preprocessor symbols have been defined
15680 (such as REF_CHECK, BN_DEBUG etc.).
15684 * New openssl application 'rsautl'. This utility can be
15685 used for low-level RSA operations. DER public key
15686 BIO/fp routines also added.
15690 * New Configure entry and patches for compiling on QNX 4.
15692 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15694 * A demo state-machine implementation was sponsored by
15695 Nuron (<http://www.nuron.com/>) and is now available in
15696 demos/state_machine.
15700 * New options added to the 'dgst' utility for signature
15701 generation and verification.
15705 * Unrecognized PKCS#7 content types are now handled via a
15706 catch all ASN1_TYPE structure. This allows unsupported
15707 types to be stored as a "blob" and an application can
15708 encode and decode it manually.
15712 * Fix various signed/unsigned issues to make a_strex.c
15713 compile under VC++.
15715 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15717 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15718 length if passed a buffer. ASN1_INTEGER_to_BN failed
15719 if passed a NULL BN and its argument was negative.
15721 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15723 * Modification to PKCS#7 encoding routines to output definite
15724 length encoding. Since currently the whole structures are in
15725 memory there's not real point in using indefinite length
15726 constructed encoding. However if OpenSSL is compiled with
15727 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15731 * Added BIO_vprintf() and BIO_vsnprintf().
15735 * Added more prefixes to parse for in the strings written
15736 through a logging bio, to cover all the levels that are available
15737 through syslog. The prefixes are now:
15739 PANIC, EMERG, EMR => LOG_EMERG
15740 ALERT, ALR => LOG_ALERT
15741 CRIT, CRI => LOG_CRIT
15742 ERROR, ERR => LOG_ERR
15743 WARNING, WARN, WAR => LOG_WARNING
15744 NOTICE, NOTE, NOT => LOG_NOTICE
15745 INFO, INF => LOG_INFO
15746 DEBUG, DBG => LOG_DEBUG
15748 and as before, if none of those prefixes are present at the
15749 beginning of the string, LOG_ERR is chosen.
15751 On Win32, the `LOG_*` levels are mapped according to this:
15753 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15754 LOG_WARNING => EVENTLOG_WARNING_TYPE
15755 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15759 * Made it possible to reconfigure with just the configuration
15760 argument "reconf" or "reconfigure". The command line arguments
15761 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15762 and are retrieved from there when reconfiguring.
15768 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15770 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15774 * The obj_dat.pl script was messing up the sorting of object
15775 names. The reason was that it compared the quoted version
15776 of strings as a result "OCSP" > "OCSP Signing" because
15777 " > SPACE. Changed script to store unquoted versions of
15778 names and add quotes on output. It was also omitting some
15779 names from the lookup table if they were given a default
15780 value (that is if SN is missing it is given the same
15781 value as LN and vice versa), these are now added on the
15782 grounds that if an object has a name we should be able to
15783 look it up. Finally added warning output when duplicate
15784 short or long names are found.
15788 * Changes needed for Tandem NSK.
15790 *Scott Uroff <scott@xypro.com>*
15792 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15793 RSA_padding_check_SSLv23(), special padding was never detected
15794 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15795 version rollback attacks was not effective.
15797 In s23_clnt.c, don't use special rollback-attack detection padding
15798 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15799 client; similarly, in s23_srvr.c, don't do the rollback check if
15800 SSL 2.0 is the only protocol enabled in the server.
15804 * Make it possible to get hexdumps of unprintable data with 'openssl
15805 asn1parse'. By implication, the functions ASN1_parse_dump() and
15806 BIO_dump_indent() are added.
15810 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15811 these print out strings and name structures based on various
15812 flags including RFC2253 support and proper handling of
15813 multibyte characters. Added options to the 'x509' utility
15814 to allow the various flags to be set.
15818 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15819 Also change the functions X509_cmp_current_time() and
15820 X509_gmtime_adj() work with an ASN1_TIME structure,
15821 this will enable certificates using GeneralizedTime in validity
15822 dates to be checked.
15826 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15827 negative public key encodings) on by default,
15828 NO_NEG_PUBKEY_BUG can be set to disable it.
15832 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15833 content octets. An i2c_ASN1_OBJECT is unnecessary because
15834 the encoding can be trivially obtained from the structure.
15838 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15839 not read locks (`CRYPTO_r_[un]lock`).
15843 * A first attempt at creating official support for shared
15844 libraries through configuration. I've kept it so the
15845 default is static libraries only, and the OpenSSL programs
15846 are always statically linked for now, but there are
15847 preparations for dynamic linking in place.
15848 This has been tested on Linux and Tru64.
15852 * Randomness polling function for Win9x, as described in:
15853 Peter Gutmann, Software Generation of Practically Strong
15858 * Fix so PRNG is seeded in req if using an already existing
15863 * New options to smime application. -inform and -outform
15864 allow alternative formats for the S/MIME message including
15865 PEM and DER. The -content option allows the content to be
15866 specified separately. This should allow things like Netscape
15867 form signing output easier to verify.
15871 * Fix the ASN1 encoding of tags using the 'long form'.
15875 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15876 STRING types. These convert content octets to and from the
15877 underlying type. The actual tag and length octets are
15878 already assumed to have been read in and checked. These
15879 are needed because all other string types have virtually
15880 identical handling apart from the tag. By having versions
15881 of the ASN1 functions that just operate on content octets
15882 IMPLICIT tagging can be handled properly. It also allows
15883 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15884 and ASN1_INTEGER are identical apart from the tag.
15888 * Change the handling of OID objects as follows:
15890 - New object identifiers are inserted in objects.txt, following
15891 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15892 - objects.pl is used to process obj_mac.num and create a new
15894 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15897 This is currently kind of a hack, and the perl code in objects.pl
15898 isn't very elegant, but it works as I intended. The simplest way
15899 to check that it worked correctly is to look in obj_dat.h and
15900 check the array nid_objs and make sure the objects haven't moved
15901 around (this is important!). Additions are OK, as well as
15902 consistent name changes.
15906 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15910 * Addition of the command line parameter '-rand file' to 'openssl req'.
15911 The given file adds to whatever has already been seeded into the
15912 random pool through the RANDFILE configuration file option or
15913 environment variable, or the default random state file.
15917 * mkstack.pl now sorts each macro group into lexical order.
15918 Previously the output order depended on the order the files
15919 appeared in the directory, resulting in needless rewriting
15924 * Patches to make OpenSSL compile under Win32 again. Mostly
15925 work arounds for the VC++ problem that it treats func() as
15926 func(void). Also stripped out the parts of mkdef.pl that
15927 added extra typesafe functions: these no longer exist.
15931 * Reorganisation of the stack code. The macros are now all
15932 collected in safestack.h . Each macro is defined in terms of
15933 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15934 DEBUG_SAFESTACK is now handled in terms of function casts,
15935 this has the advantage of retaining type safety without the
15936 use of additional functions. If DEBUG_SAFESTACK is not defined
15937 then the non typesafe macros are used instead. Also modified the
15938 mkstack.pl script to handle the new form. Needs testing to see
15939 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15940 the default if no major problems. Similar behaviour for ASN1_SET_OF
15941 and PKCS12_STACK_OF.
15945 * When some versions of IIS use the 'NET' form of private key the
15946 key derivation algorithm is different. Normally MD5(password) is
15947 used as a 128 bit RC4 key. In the modified case
15948 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15949 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15950 as the old Netscape_RSA functions except they have an additional
15951 'sgckey' parameter which uses the modified algorithm. Also added
15952 an -sgckey command line option to the rsa utility. Thanks to
15953 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15954 algorithm to openssl-dev.
15958 * The evp_local.h macros were using 'c.##kname' which resulted in
15959 invalid expansion on some systems (SCO 5.0.5 for example).
15960 Corrected to 'c.kname'.
15962 *Phillip Porch <root@theporch.com>*
15964 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15965 a STACK of email addresses from a certificate or request, these look
15966 in the subject name and the subject alternative name extensions and
15967 omit any duplicate addresses.
15971 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15972 This makes DSA verification about 2 % faster.
15976 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15977 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15978 plus overhead for 1024 bit moduli).
15979 This makes exponentiations about 0.5 % faster for 1024 bit
15980 exponents (as measured by "openssl speed rsa2048").
15984 * Rename memory handling macros to avoid conflicts with other
15986 Malloc => OPENSSL_malloc
15987 Malloc_locked => OPENSSL_malloc_locked
15988 Realloc => OPENSSL_realloc
15989 Free => OPENSSL_free
15993 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15994 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15998 * CygWin32 support.
16000 *John Jarvie <jjarvie@newsguy.com>*
16002 * The type-safe stack code has been rejigged. It is now only compiled
16003 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16004 by default all type-specific stack functions are "#define"d back to
16005 standard stack functions. This results in more streamlined output
16006 but retains the type-safety checking possibilities of the original
16011 * The STACK code has been cleaned up, and certain type declarations
16012 that didn't make a lot of sense have been brought in line. This has
16013 also involved a cleanup of sorts in safestack.h to more correctly
16014 map type-safe stack functions onto their plain stack counterparts.
16015 This work has also resulted in a variety of "const"ifications of
16016 lots of the code, especially `_cmp` operations which should normally
16017 be prototyped with "const" parameters anyway.
16021 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16022 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16023 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16024 where all of 'md' is used each time the PRNG is used, but 'state'
16025 is used only indexed by a cyclic counter. As entropy may not be
16026 well distributed from the beginning, 'md' is important as a
16027 chaining variable. However, the output function chains only half
16028 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16029 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16030 in all of 'state' being rewritten, with the new values depending
16031 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16035 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16036 the handshake is continued after ssl_verify_cert_chain();
16037 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16038 can lead to 'unexplainable' connection aborts later.
16040 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16042 * Major EVP API cipher revision.
16043 Add hooks for extra EVP features. This allows various cipher
16044 parameters to be set in the EVP interface. Support added for variable
16045 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16046 setting of RC2 and RC5 parameters.
16048 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16051 Remove lots of duplicated code from the EVP library. For example *every*
16052 cipher init() function handles the 'iv' in the same way according to the
16053 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16054 for CFB and OFB modes they zero ctx->num.
16056 New functionality allows removal of S/MIME code RC2 hack.
16058 Most of the routines have the same form and so can be declared in terms
16061 By shifting this to the top level EVP_CipherInit() it can be removed from
16062 all individual ciphers. If the cipher wants to handle IVs or keys
16063 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16066 Change lots of functions like EVP_EncryptUpdate() to now return a
16067 value: although software versions of the algorithms cannot fail
16068 any installed hardware versions can.
16072 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16073 this option is set, tolerate broken clients that send the negotiated
16074 protocol version number instead of the requested protocol version
16079 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16080 i.e. non-zero for export ciphersuites, zero otherwise.
16081 Previous versions had this flag inverted, inconsistent with
16082 rsa_tmp_cb (..._TMP_RSA_CB).
16084 *Bodo Moeller; problem reported by Amit Chopra*
16086 * Add missing DSA library text string. Work around for some IIS
16087 key files with invalid SEQUENCE encoding.
16091 * Add a document (doc/standards.txt) that list all kinds of standards
16092 and so on that are implemented in OpenSSL.
16096 * Enhance c_rehash script. Old version would mishandle certificates
16097 with the same subject name hash and wouldn't handle CRLs at all.
16098 Added -fingerprint option to crl utility, to support new c_rehash
16103 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16107 * Fix for SSL server purpose checking. Server checking was
16108 rejecting certificates which had extended key usage present
16109 but no ssl client purpose.
16111 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16113 * Make PKCS#12 code work with no password. The PKCS#12 spec
16114 is a little unclear about how a blank password is handled.
16115 Since the password in encoded as a BMPString with terminating
16116 double NULL a zero length password would end up as just the
16117 double NULL. However no password at all is different and is
16118 handled differently in the PKCS#12 key generation code. NS
16119 treats a blank password as zero length. MSIE treats it as no
16120 password on export: but it will try both on import. We now do
16121 the same: PKCS12_parse() tries zero length and no password if
16122 the password is set to "" or NULL (NULL is now a valid password:
16123 it wasn't before) as does the pkcs12 application.
16127 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16128 perror when PEM_read_bio_X509_REQ fails, the error message must
16129 be obtained from the error queue.
16133 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16134 it in ERR_remove_state if appropriate, and change ERR_get_state
16135 accordingly to avoid race conditions (this is necessary because
16136 thread_hash is no longer constant once set).
16140 * Bugfix for linux-elf makefile.one.
16144 * RSA_get_default_method() will now cause a default
16145 RSA_METHOD to be chosen if one doesn't exist already.
16146 Previously this was only set during a call to RSA_new()
16147 or RSA_new_method(NULL) meaning it was possible for
16148 RSA_get_default_method() to return NULL.
16152 * Added native name translation to the existing DSO code
16153 that will convert (if the flag to do so is set) filenames
16154 that are sufficiently small and have no path information
16155 into a canonical native form. Eg. "blah" converted to
16156 "libblah.so" or "blah.dll" etc.
16160 * New function ERR_error_string_n(e, buf, len) which is like
16161 ERR_error_string(e, buf), but writes at most 'len' bytes
16162 including the 0 terminator. For ERR_error_string_n, 'buf'
16165 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16167 * CONF library reworked to become more general. A new CONF
16168 configuration file reader "class" is implemented as well as a
16169 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16170 old `CONF_*` functions are still there, but are reimplemented to
16171 work in terms of the new functions. Also, a set of functions
16172 to handle the internal storage of the configuration data is
16173 provided to make it easier to write new configuration file
16174 reader "classes" (I can definitely see something reading a
16175 configuration file in XML format, for example), called `_CONF_*`,
16176 or "the configuration storage API"...
16178 The new configuration file reading functions are:
16180 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16181 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16183 NCONF_default, NCONF_WIN32
16185 NCONF_dump_fp, NCONF_dump_bio
16187 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16188 NCONF_new creates a new CONF object. This works in the same way
16189 as other interfaces in OpenSSL, like the BIO interface.
16190 `NCONF_dump_*` dump the internal storage of the configuration file,
16191 which is useful for debugging. All other functions take the same
16192 arguments as the old `CONF_*` functions with the exception of the
16193 first that must be a `CONF *` instead of a `LHASH *`.
16195 To make it easier to use the new classes with the old `CONF_*` functions,
16196 the function CONF_set_default_method is provided.
16200 * Add '-tls1' option to 'openssl ciphers', which was already
16201 mentioned in the documentation but had not been implemented.
16202 (This option is not yet really useful because even the additional
16203 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16207 * Initial DSO code added into libcrypto for letting OpenSSL (and
16208 OpenSSL-based applications) load shared libraries and bind to
16209 them in a portable way.
16211 *Geoff Thorpe, with contributions from Richard Levitte*
16213 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16215 * Make sure _lrotl and _lrotr are only used with MSVC.
16217 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16218 (the default implementation of RAND_status).
16220 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16221 to '-clrext' (= clear extensions), as intended and documented.
16222 *Bodo Moeller; inconsistency pointed out by Michael Attili
16223 <attili@amaxo.com>*
16225 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16226 was larger than the MD block size.
16228 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16230 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16231 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16232 using the passed key: if the passed key was a private key the result
16233 of X509_print(), for example, would be to print out all the private key
16238 * des_quad_cksum() byte order bug fix.
16239 *Ulf Möller, using the problem description in krb4-0.9.7, where
16240 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16242 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16245 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16247 * For easily testing in shell scripts whether some command
16248 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16249 returns with exit code 0 iff no command of the given name is available.
16250 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16251 the output goes to stdout and nothing is printed to stderr.
16252 Additional arguments are always ignored.
16254 Since for each cipher there is a command of the same name,
16255 the 'no-cipher' compilation switches can be tested this way.
16257 ('openssl no-XXX' is not able to detect pseudo-commands such
16258 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16262 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16266 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16267 is set; it will be thrown away anyway because each handshake creates
16269 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16270 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16271 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16272 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16276 * New s_client option -ign_eof: EOF at stdin is ignored, and
16277 'Q' and 'R' lose their special meanings (quit/renegotiate).
16278 This is part of what -quiet does; unlike -quiet, -ign_eof
16279 does not suppress any output.
16283 * Add compatibility options to the purpose and trust code. The
16284 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16285 accepts a certificate or CA, this was the previous behaviour,
16286 with all the associated security issues.
16288 X509_TRUST_COMPAT is the old trust behaviour: only and
16289 automatically trust self signed roots in certificate store. A
16290 new trust setting X509_TRUST_DEFAULT is used to specify that
16291 a purpose has no associated trust setting and it should instead
16292 use the value in the default purpose.
16296 * Fix the PKCS#8 DSA private key code so it decodes keys again
16297 and fix a memory leak.
16301 * In util/mkerr.pl (which implements 'make errors'), preserve
16302 reason strings from the previous version of the .c file, as
16303 the default to have only downcase letters (and digits) in
16304 automatically generated reasons codes is not always appropriate.
16308 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16309 using strerror. Previously, ERR_reason_error_string() returned
16310 library names as reason strings for SYSerr; but SYSerr is a special
16311 case where small numbers are errno values, not library numbers.
16315 * Add '-dsaparam' option to 'openssl dhparam' application. This
16316 converts DSA parameters into DH parameters. (When creating parameters,
16317 DSA_generate_parameters is used.)
16321 * Include 'length' (recommended exponent length) in C code generated
16322 by 'openssl dhparam -C'.
16326 * The second argument to set_label in perlasm was already being used
16327 so couldn't be used as a "file scope" flag. Moved to third argument
16332 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16333 instead of RAND_bytes for encryption IVs and salts.
16337 * Include RAND_status() into RAND_METHOD instead of implementing
16338 it only for md_rand.c Otherwise replacing the PRNG by calling
16339 RAND_set_rand_method would be impossible.
16343 * Don't let DSA_generate_key() enter an infinite loop if the random
16344 number generation fails.
16348 * New 'rand' application for creating pseudo-random output.
16352 * Added configuration support for Linux/IA64
16354 *Rolf Haberrecker <rolf@suse.de>*
16356 * Assembler module support for Mingw32.
16360 * Shared library support for HPUX (in shlib/).
16362 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16364 * Shared library support for Solaris gcc.
16366 *Lutz Behnke <behnke@trustcenter.de>*
16368 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16370 * PKCS7_encrypt() was adding text MIME headers twice because they
16371 were added manually and by SMIME_crlf_copy().
16375 * In bntest.c don't call BN_rand with zero bits argument.
16377 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16379 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16380 case was implemented. This caused BN_div_recp() to fail occasionally.
16384 * Add an optional second argument to the set_label() in the perl
16385 assembly language builder. If this argument exists and is set
16386 to 1 it signals that the assembler should use a symbol whose
16387 scope is the entire file, not just the current function. This
16388 is needed with MASM which uses the format label:: for this scope.
16390 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16392 * Change the ASN1 types so they are typedefs by default. Before
16393 almost all types were #define'd to ASN1_STRING which was causing
16394 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16399 * Change names of new functions to the new get1/get0 naming
16400 convention: After 'get1', the caller owns a reference count
16401 and has to call `..._free`; 'get0' returns a pointer to some
16402 data structure without incrementing reference counters.
16403 (Some of the existing 'get' functions increment a reference
16404 counter, some don't.)
16405 Similarly, 'set1' and 'add1' functions increase reference
16406 counters or duplicate objects.
16410 * Allow for the possibility of temp RSA key generation failure:
16411 the code used to assume it always worked and crashed on failure.
16415 * Fix potential buffer overrun problem in BIO_printf().
16416 *Ulf Möller, using public domain code by Patrick Powell; problem
16417 pointed out by David Sacerdote <das33@cornell.edu>*
16419 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16420 RAND_egd() and RAND_status(). In the command line application,
16421 the EGD socket can be specified like a seed file using RANDFILE
16426 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16427 Some CAs (e.g. Verisign) distribute certificates in this form.
16431 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16432 list to exclude them. This means that no special compilation option
16433 is needed to use anonymous DH: it just needs to be included in the
16438 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16439 EVP_MD_type. The old functionality is available in a new macro called
16440 EVP_MD_md(). Change code that uses it and update docs.
16444 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16445 where the `void *` argument is replaced by a function pointer argument.
16446 Previously `void *` was abused to point to functions, which works on
16447 many platforms, but is not correct. As these functions are usually
16448 called by macros defined in OpenSSL header files, most source code
16449 should work without changes.
16453 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16454 sections with information on -D... compiler switches used for
16455 compiling the library so that applications can see them. To enable
16456 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16457 must be defined. E.g.,
16458 #define OPENSSL_ALGORITHM_DEFINES
16459 #include <openssl/opensslconf.h>
16460 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16462 *Richard Levitte, Ulf and Bodo Möller*
16464 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16469 * Change the 'other' type in certificate aux info to a STACK_OF
16470 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16471 the required ASN1 format: arbitrary types determined by an OID.
16475 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16476 argument to 'req'. This is not because the function is newer or
16477 better than others it just uses the work 'NEW' in the certificate
16478 request header lines. Some software needs this.
16482 * Reorganise password command line arguments: now passwords can be
16483 obtained from various sources. Delete the PEM_cb function and make
16484 it the default behaviour: i.e. if the callback is NULL and the
16485 usrdata argument is not NULL interpret it as a null terminated pass
16486 phrase. If usrdata and the callback are NULL then the pass phrase
16487 is prompted for as usual.
16491 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16492 the support is automatically enabled. The resulting binaries will
16493 autodetect the card and use it if present.
16495 *Ben Laurie and Compaq Inc.*
16497 * Work around for Netscape hang bug. This sends certificate request
16498 and server done in one record. Since this is perfectly legal in the
16499 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16500 the bugs/SSLv3 entry for more info.
16504 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16508 * Add -rand argument to smime and pkcs12 applications and read/write
16513 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16517 * Add command line password options to the remaining applications.
16521 * Bug fix for BN_div_recp() for numerators with an even number of
16526 * More tests in bntest.c, and changed test_bn output.
16530 * ./config recognizes MacOS X now.
16534 * Bug fix for BN_div() when the first words of num and divisor are
16535 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16539 * Add support for various broken PKCS#8 formats, and command line
16540 options to produce them.
16544 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16545 get temporary BIGNUMs from a BN_CTX.
16549 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16554 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16555 include a #define from the old name to the new. The original intent
16556 was that statically linked binaries could for example just call
16557 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16558 link with digests. This never worked because SSLeay_add_all_digests()
16559 and SSLeay_add_all_ciphers() were in the same source file so calling
16560 one would link with the other. They are now in separate source files.
16564 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16568 * Use a less unusual form of the Miller-Rabin primality test (it used
16569 a binary algorithm for exponentiation integrated into the Miller-Rabin
16570 loop, our standard modexp algorithms are faster).
16574 * Support for the EBCDIC character set completed.
16576 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16578 * Source code cleanups: use const where appropriate, eliminate casts,
16579 use `void *` instead of `char *` in lhash.
16583 * Bugfix: ssl3_send_server_key_exchange was not restartable
16584 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16585 this the server could overwrite ephemeral keys that the client
16590 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16591 using 50 iterations of the Rabin-Miller test.
16593 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16594 iterations of the Rabin-Miller test as required by the appendix
16595 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16596 As BN_is_prime_fasttest includes trial division, DSA parameter
16597 generation becomes much faster.
16599 This implies a change for the callback functions in DSA_is_prime
16600 and DSA_generate_parameters: The callback function is called once
16601 for each positive witness in the Rabin-Miller test, not just
16602 occasionally in the inner loop; and the parameters to the
16603 callback function now provide an iteration count for the outer
16604 loop rather than for the current invocation of the inner loop.
16605 DSA_generate_parameters additionally can call the callback
16606 function with an 'iteration count' of -1, meaning that a
16607 candidate has passed the trial division test (when q is generated
16608 from an application-provided seed, trial division is skipped).
16612 * New function BN_is_prime_fasttest that optionally does trial
16613 division before starting the Rabin-Miller test and has
16614 an additional BN_CTX * argument (whereas BN_is_prime always
16615 has to allocate at least one BN_CTX).
16616 'callback(1, -1, cb_arg)' is called when a number has passed the
16617 trial division stage.
16621 * Fix for bug in CRL encoding. The validity dates weren't being handled
16626 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16630 * New function BN_pseudo_rand().
16634 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16635 bignum version of BN_from_montgomery() with the working code from
16636 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16641 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16642 made it impossible to use the same SSL_SESSION data structure in
16643 SSL2 clients in multiple threads.
16647 * The return value of RAND_load_file() no longer counts bytes obtained
16648 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16649 to seed the PRNG (previously an explicit byte count was required).
16651 *Ulf Möller, Bodo Möller*
16653 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16654 used `char *` instead of `void *` and had casts all over the place.
16658 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16662 * Retain source code compatibility for BN_prime_checks macro:
16663 BN_is_prime(..., BN_prime_checks, ...) now uses
16664 BN_prime_checks_for_size to determine the appropriate number of
16665 Rabin-Miller iterations.
16669 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16670 DH_CHECK_P_NOT_SAFE_PRIME.
16671 (Check if this is true? OpenPGP calls them "strong".)
16675 * Merge the functionality of "dh" and "gendh" programs into a new program
16676 "dhparam". The old programs are retained for now but will handle DH keys
16677 (instead of parameters) in future.
16681 * Make the ciphers, s_server and s_client programs check the return values
16682 when a new cipher list is set.
16686 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16687 ciphers. Before when the 56bit ciphers were enabled the sorting was
16690 The syntax for the cipher sorting has been extended to support sorting by
16691 cipher-strength (using the strength_bits hard coded in the tables).
16692 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16694 Fix a bug in the cipher-command parser: when supplying a cipher command
16695 string with an "undefined" symbol (neither command nor alphanumeric
16696 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16697 an error is flagged.
16699 Due to the strength-sorting extension, the code of the
16700 ssl_create_cipher_list() function was completely rearranged. I hope that
16701 the readability was also increased :-)
16703 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16705 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16706 for the first serial number and places 2 in the serial number file. This
16707 avoids problems when the root CA is created with serial number zero and
16708 the first user certificate has the same issuer name and serial number
16713 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16714 the new code. Add documentation for this stuff.
16718 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16719 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16720 structures and behave in an analogous way to the X509v3 functions:
16721 they shouldn't be called directly but wrapper functions should be used
16724 So we also now have some wrapper functions that call the X509at functions
16725 when passed certificate requests. (TO DO: similar things can be done with
16726 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16727 things. Some of these need some d2i or i2d and print functionality
16728 because they handle more complex structures.)
16732 * Add missing #ifndefs that caused missing symbols when building libssl
16733 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16734 NO_RSA in `ssl/s2*.c`.
16736 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16738 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16739 has a return value which indicates the quality of the random data
16740 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16741 error queue. New function RAND_pseudo_bytes() generates output that is
16742 guaranteed to be unique but not unpredictable. RAND_add is like
16743 RAND_seed, but takes an extra argument for an entropy estimate
16744 (RAND_seed always assumes full entropy).
16748 * Do more iterations of Rabin-Miller probable prime test (specifically,
16749 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16750 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16751 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16752 false-positive rate of at most 2^-80 for random input.
16756 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16760 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16761 in the 0.9.5 release), this returns the chain
16762 from an X509_CTX structure with a dup of the stack and all
16763 the X509 reference counts upped: so the stack will exist
16764 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16767 Also make SSL_SESSION_print() print out the verify return
16772 * Add manpage for the pkcs12 command. Also change the default
16773 behaviour so MAC iteration counts are used unless the new
16774 -nomaciter option is used. This improves file security and
16775 only older versions of MSIE (4.0 for example) need it.
16779 * Honor the no-xxx Configure options when creating .DEF files.
16783 * Add PKCS#10 attributes to field table: challengePassword,
16784 unstructuredName and unstructuredAddress. These are taken from
16785 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16786 international characters are used.
16788 More changes to X509_ATTRIBUTE code: allow the setting of types
16789 based on strings. Remove the 'loc' parameter when adding
16790 attributes because these will be a SET OF encoding which is sorted
16795 * Initial changes to the 'req' utility to allow request generation
16796 automation. This will allow an application to just generate a template
16797 file containing all the field values and have req construct the
16800 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16801 used all over the place including certificate requests and PKCS#7
16802 structures. They are currently handled manually where necessary with
16803 some primitive wrappers for PKCS#7. The new functions behave in a
16804 manner analogous to the X509 extension functions: they allow
16805 attributes to be looked up by NID and added.
16807 Later something similar to the X509V3 code would be desirable to
16808 automatically handle the encoding, decoding and printing of the
16809 more complex types. The string types like challengePassword can
16810 be handled by the string table functions.
16812 Also modified the multi byte string table handling. Now there is
16813 a 'global mask' which masks out certain types. The table itself
16814 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16815 is useful when for example there is only one permissible type
16816 (as in countryName) and using the mask might result in no valid
16821 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16822 SSL_get_peer_finished to allow applications to obtain the latest
16823 Finished messages sent to the peer or expected from the peer,
16824 respectively. (SSL_get_peer_finished is usually the Finished message
16825 actually received from the peer, otherwise the protocol will be aborted.)
16827 As the Finished message are message digests of the complete handshake
16828 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16829 be used for external authentication procedures when the authentication
16830 provided by SSL/TLS is not desired or is not enough.
16834 * Enhanced support for Alpha Linux is added. Now ./config checks if
16835 the host supports BWX extension and if Compaq C is present on the
16836 $PATH. Just exploiting of the BWX extension results in 20-30%
16837 performance kick for some algorithms, e.g. DES and RC4 to mention
16838 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16843 * Add support for MS "fast SGC". This is arguably a violation of the
16844 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16845 weak crypto and after checking the certificate is SGC a second one
16846 with strong crypto. MS SGC stops the first handshake after receiving
16847 the server certificate message and sends a second client hello. Since
16848 a server will typically do all the time consuming operations before
16849 expecting any further messages from the client (server key exchange
16850 is the most expensive) there is little difference between the two.
16852 To get OpenSSL to support MS SGC we have to permit a second client
16853 hello message after we have sent server done. In addition we have to
16854 reset the MAC if we do get this second client hello.
16858 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16859 if a DER encoded private key is RSA or DSA traditional format. Changed
16860 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16861 format DER encoded private key. Newer code should use PKCS#8 format which
16862 has the key type encoded in the ASN1 structure. Added DER private key
16863 support to pkcs8 application.
16867 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16868 ciphersuites has been selected (as required by the SSL 3/TLS 1
16869 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16870 is set, we interpret this as a request to violate the specification
16871 (the worst that can happen is a handshake failure, and 'correct'
16872 behaviour would result in a handshake failure anyway).
16876 * In SSL_CTX_add_session, take into account that there might be multiple
16877 SSL_SESSION structures with the same session ID (e.g. when two threads
16878 concurrently obtain them from an external cache).
16879 The internal cache can handle only one SSL_SESSION with a given ID,
16880 so if there's a conflict, we now throw out the old one to achieve
16885 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16886 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16887 some routines that use cipher OIDs: some ciphers do not have OIDs
16888 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16893 * Simplify the trust setting structure and code. Now we just have
16894 two sequences of OIDs for trusted and rejected settings. These will
16895 typically have values the same as the extended key usage extension
16896 and any application specific purposes.
16898 The trust checking code now has a default behaviour: it will just
16899 check for an object with the same NID as the passed id. Functions can
16900 be provided to override either the default behaviour or the behaviour
16901 for a given id. SSL client, server and email already have functions
16902 in place for compatibility: they check the NID and also return "trusted"
16903 if the certificate is self signed.
16907 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16908 traditional format into an EVP_PKEY structure.
16912 * Add a password callback function PEM_cb() which either prompts for
16913 a password if usr_data is NULL or otherwise assumes it is a null
16914 terminated password. Allow passwords to be passed on command line
16915 environment or config files in a few more utilities.
16919 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16920 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16921 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16922 Update documentation.
16926 * Support for ASN1 "NULL" type. This could be handled before by using
16927 ASN1_TYPE but there wasn't any function that would try to read a NULL
16928 and produce an error if it couldn't. For compatibility we also have
16929 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16930 don't allocate anything because they don't need to.
16934 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16937 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16939 * Rebuild of the memory allocation routines used by OpenSSL code and
16940 possibly others as well. The purpose is to make an interface that
16941 provide hooks so anyone can build a separate set of allocation and
16942 deallocation routines to be used by OpenSSL, for example memory
16943 pool implementations, or something else, which was previously hard
16944 since Malloc(), Realloc() and Free() were defined as macros having
16945 the values malloc, realloc and free, respectively (except for Win32
16946 compilations). The same is provided for memory debugging code.
16947 OpenSSL already comes with functionality to find memory leaks, but
16948 this gives people a chance to debug other memory problems.
16950 With these changes, a new set of functions and macros have appeared:
16952 CRYPTO_set_mem_debug_functions() [F]
16953 CRYPTO_get_mem_debug_functions() [F]
16954 CRYPTO_dbg_set_options() [F]
16955 CRYPTO_dbg_get_options() [F]
16956 CRYPTO_malloc_debug_init() [M]
16958 The memory debug functions are NULL by default, unless the library
16959 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16960 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16961 gives the standard debugging functions that come with OpenSSL) or
16962 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16963 provided by the library user) must be used. When the standard
16964 debugging functions are used, CRYPTO_dbg_set_options can be used to
16965 request additional information:
16966 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16967 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16969 Also, things like CRYPTO_set_mem_functions will always give the
16970 expected result (the new set of functions is used for allocation
16971 and deallocation) at all times, regardless of platform and compiler
16974 To finish it up, some functions that were never use in any other
16975 way than through macros have a new API and new semantic:
16977 CRYPTO_dbg_malloc()
16978 CRYPTO_dbg_realloc()
16981 All macros of value have retained their old syntax.
16983 *Richard Levitte and Bodo Moeller*
16985 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16986 ordering of SMIMECapabilities wasn't in "strength order" and there
16987 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16992 * Some ASN1 types with illegal zero length encoding (INTEGER,
16993 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16995 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16997 * Merge in my S/MIME library for OpenSSL. This provides a simple
16998 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16999 functionality to handle multipart/signed properly) and a utility
17000 called 'smime' to call all this stuff. This is based on code I
17001 originally wrote for Celo who have kindly allowed it to be
17002 included in OpenSSL.
17006 * Add variants des_set_key_checked and des_set_key_unchecked of
17007 des_set_key (aka des_key_sched). Global variable des_check_key
17008 decides which of these is called by des_set_key; this way
17009 des_check_key behaves as it always did, but applications and
17010 the library itself, which was buggy for des_check_key == 1,
17011 have a cleaner way to pick the version they need.
17015 * New function PKCS12_newpass() which changes the password of a
17020 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17021 dynamic mix. In both cases the ids can be used as an index into the
17022 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17023 functions so they accept a list of the field values and the
17024 application doesn't need to directly manipulate the X509_TRUST
17029 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17034 * Modify the way the V3 extension code looks up extensions. This now
17035 works in a similar way to the object code: we have some "standard"
17036 extensions in a static table which is searched with OBJ_bsearch()
17037 and the application can add dynamic ones if needed. The file
17038 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17039 updated whenever a new extension is added to the core code and kept
17040 in ext_nid order. There is a simple program 'tabtest.c' which checks
17041 this. New extensions are not added too often so this file can readily
17042 be maintained manually.
17044 There are two big advantages in doing things this way. The extensions
17045 can be looked up immediately and no longer need to be "added" using
17046 X509V3_add_standard_extensions(): this function now does nothing.
17047 Side note: I get *lots* of email saying the extension code doesn't
17048 work because people forget to call this function.
17049 Also no dynamic allocation is done unless new extensions are added:
17050 so if we don't add custom extensions there is no need to call
17051 X509V3_EXT_cleanup().
17055 * Modify enc utility's salting as follows: make salting the default. Add a
17056 magic header, so unsalted files fail gracefully instead of just decrypting
17057 to garbage. This is because not salting is a big security hole, so people
17058 should be discouraged from doing it.
17062 * Fixes and enhancements to the 'x509' utility. It allowed a message
17063 digest to be passed on the command line but it only used this
17064 parameter when signing a certificate. Modified so all relevant
17065 operations are affected by the digest parameter including the
17066 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17067 DSA key was used because it didn't fix the digest.
17071 * Initial certificate chain verify code. Currently tests the untrusted
17072 certificates for consistency with the verify purpose (which is set
17073 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17075 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17076 this is because it will reject chains with invalid extensions whereas
17077 every previous version of OpenSSL and SSLeay made no checks at all.
17079 Trust code: checks the root CA for the relevant trust settings. Trust
17080 settings have an initial value consistent with the verify purpose: e.g.
17081 if the verify purpose is for SSL client use it expects the CA to be
17082 trusted for SSL client use. However the default value can be changed to
17083 permit custom trust settings: one example of this would be to only trust
17084 certificates from a specific "secure" set of CAs.
17086 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17087 which should be used for version portability: especially since the
17088 verify structure is likely to change more often now.
17090 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17091 to set them. If not set then assume SSL clients will verify SSL servers
17094 Two new options to the verify program: -untrusted allows a set of
17095 untrusted certificates to be passed in and -purpose which sets the
17096 intended purpose of the certificate. If a purpose is set then the
17097 new chain verify code is used to check extension consistency.
17101 * Support for the authority information access extension.
17105 * Modify RSA and DSA PEM read routines to transparently handle
17106 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17107 public keys in a format compatible with certificate
17108 SubjectPublicKeyInfo structures. Unfortunately there were already
17109 functions called *_PublicKey_* which used various odd formats so
17110 these are retained for compatibility: however the DSA variants were
17111 never in a public release so they have been deleted. Changed dsa/rsa
17112 utilities to handle the new format: note no releases ever handled public
17113 keys so we should be OK.
17115 The primary motivation for this change is to avoid the same fiasco
17116 that dogs private keys: there are several incompatible private key
17117 formats some of which are standard and some OpenSSL specific and
17118 require various evil hacks to allow partial transparent handling and
17119 even then it doesn't work with DER formats. Given the option anything
17120 other than PKCS#8 should be dumped: but the other formats have to
17121 stay in the name of compatibility.
17123 With public keys and the benefit of hindsight one standard format
17124 is used which works with EVP_PKEY, RSA or DSA structures: though
17125 it clearly returns an error if you try to read the wrong kind of key.
17127 Added a -pubkey option to the 'x509' utility to output the public key.
17128 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17129 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17130 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17131 that do the same as the `EVP_PKEY_assign_*()` except they up the
17132 reference count of the added key (they don't "swallow" the
17137 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17138 CRLs would fail if the file contained no certificates or no CRLs:
17139 added a new function to read in both types and return the number
17140 read: this means that if none are read it will be an error. The
17141 DER versions of the certificate and CRL reader would always fail
17142 because it isn't possible to mix certificates and CRLs in DER format
17143 without choking one or the other routine. Changed this to just read
17144 a certificate: this is the best we can do. Also modified the code
17145 in `apps/verify.c` to take notice of return codes: it was previously
17146 attempting to read in certificates from NULL pointers and ignoring
17147 any errors: this is one reason why the cert and CRL reader seemed
17148 to work. It doesn't check return codes from the default certificate
17149 routines: these may well fail if the certificates aren't installed.
17153 * Code to support otherName option in GeneralName.
17157 * First update to verify code. Change the verify utility
17158 so it warns if it is passed a self signed certificate:
17159 for consistency with the normal behaviour. X509_verify
17160 has been modified to it will now verify a self signed
17161 certificate if *exactly* the same certificate appears
17162 in the store: it was previously impossible to trust a
17163 single self signed certificate. This means that:
17164 openssl verify ss.pem
17165 now gives a warning about a self signed certificate but
17166 openssl verify -CAfile ss.pem ss.pem
17171 * For servers, store verify_result in SSL_SESSION data structure
17172 (and add it to external session representation).
17173 This is needed when client certificate verifications fails,
17174 but an application-provided verification callback (set by
17175 SSL_CTX_set_cert_verify_callback) allows accepting the session
17176 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17177 but returns 1): When the session is reused, we have to set
17178 ssl->verify_result to the appropriate error code to avoid
17181 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17183 * Fix a bug in the new PKCS#7 code: it didn't consider the
17184 case in PKCS7_dataInit() where the signed PKCS7 structure
17185 didn't contain any existing data because it was being created.
17187 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17189 * Add a salt to the key derivation routines in enc.c. This
17190 forms the first 8 bytes of the encrypted file. Also add a
17191 -S option to allow a salt to be input on the command line.
17195 * New function X509_cmp(). Oddly enough there wasn't a function
17196 to compare two certificates. We do this by working out the SHA1
17197 hash and comparing that. X509_cmp() will be needed by the trust
17202 * SSL_get1_session() is like SSL_get_session(), but increments
17203 the reference count in the SSL_SESSION returned.
17205 *Geoff Thorpe <geoff@eu.c2.net>*
17207 * Fix for 'req': it was adding a null to request attributes.
17208 Also change the X509_LOOKUP and X509_INFO code to handle
17209 certificate auxiliary information.
17213 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17218 * Add the possibility to add extra information to the memory leak
17219 detecting output, to form tracebacks, showing from where each
17220 allocation was originated: CRYPTO_push_info("constant string") adds
17221 the string plus current file name and line number to a per-thread
17222 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17223 is like calling CYRPTO_pop_info() until the stack is empty.
17224 Also updated memory leak detection code to be multi-thread-safe.
17228 * Add options -text and -noout to pkcs7 utility and delete the
17229 encryption options which never did anything. Update docs.
17233 * Add options to some of the utilities to allow the pass phrase
17234 to be included on either the command line (not recommended on
17235 OSes like Unix) or read from the environment. Update the
17236 manpages and fix a few bugs.
17240 * Add a few manpages for some of the openssl commands.
17244 * Fix the -revoke option in ca. It was freeing up memory twice,
17245 leaking and not finding already revoked certificates.
17249 * Extensive changes to support certificate auxiliary information.
17250 This involves the use of X509_CERT_AUX structure and X509_AUX
17251 functions. An X509_AUX function such as PEM_read_X509_AUX()
17252 can still read in a certificate file in the usual way but it
17253 will also read in any additional "auxiliary information". By
17254 doing things this way a fair degree of compatibility can be
17255 retained: existing certificates can have this information added
17256 using the new 'x509' options.
17258 Current auxiliary information includes an "alias" and some trust
17259 settings. The trust settings will ultimately be used in enhanced
17260 certificate chain verification routines: currently a certificate
17261 can only be trusted if it is self signed and then it is trusted
17266 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17267 The problem was that one of the replacement routines had not been working
17268 since SSLeay releases. For now the offending routine has been replaced
17269 with non-optimised assembler. Even so, this now gives around 95%
17270 performance improvement for 1024 bit RSA signs.
17274 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17275 handling. Most clients have the effective key size in bits equal to
17276 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17277 A few however don't do this and instead use the size of the decrypted key
17278 to determine the RC2 key length and the AlgorithmIdentifier to determine
17279 the effective key length. In this case the effective key length can still
17280 be 40 bits but the key length can be 168 bits for example. This is fixed
17281 by manually forcing an RC2 key into the EVP_PKEY structure because the
17282 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17283 the key length and effective key length are equal.
17287 * Add a bunch of functions that should simplify the creation of
17288 X509_NAME structures. Now you should be able to do:
17289 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17290 and have it automatically work out the correct field type and fill in
17291 the structures. The more adventurous can try:
17292 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17293 and it will (hopefully) work out the correct multibyte encoding.
17297 * Change the 'req' utility to use the new field handling and multibyte
17298 copy routines. Before the DN field creation was handled in an ad hoc
17299 way in req, ca, and x509 which was rather broken and didn't support
17300 BMPStrings or UTF8Strings. Since some software doesn't implement
17301 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17302 using the dirstring_type option. See the new comment in the default
17303 openssl.cnf for more info.
17307 * Make crypto/rand/md_rand.c more robust:
17308 - Assure unique random numbers after fork().
17309 - Make sure that concurrent threads access the global counter and
17310 md serializably so that we never lose entropy in them
17311 or use exactly the same state in multiple threads.
17312 Access to the large state is not always serializable because
17313 the additional locking could be a performance killer, and
17314 md should be large enough anyway.
17318 * New file `apps/app_rand.c` with commonly needed functionality
17319 for handling the random seed file.
17321 Use the random seed file in some applications that previously did not:
17323 dsaparam -genkey (which also ignored its '-rand' option),
17326 x509 (when signing).
17327 Except on systems with /dev/urandom, it is crucial to have a random
17328 seed file at least for key creation, DSA signing, and for DH exchanges;
17329 for RSA signatures we could do without one.
17331 gendh and gendsa (unlike genrsa) used to read only the first byte
17332 of each file listed in the '-rand' option. The function as previously
17333 found in genrsa is now in app_rand.c and is used by all programs
17334 that support '-rand'.
17338 * In RAND_write_file, use mode 0600 for creating files;
17339 don't just chmod when it may be too late.
17343 * Report an error from X509_STORE_load_locations
17344 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17348 * New function ASN1_mbstring_copy() this copies a string in either
17349 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17350 into an ASN1_STRING type. A mask of permissible types is passed
17351 and it chooses the "minimal" type to use or an error if not type
17356 * Add function equivalents to the various macros in asn1.h. The old
17357 macros are retained with an `M_` prefix. Code inside the library can
17358 use the `M_` macros. External code (including the openssl utility)
17359 should *NOT* in order to be "shared library friendly".
17363 * Add various functions that can check a certificate's extensions
17364 to see if it usable for various purposes such as SSL client,
17365 server or S/MIME and CAs of these types. This is currently
17366 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17367 verification. Also added a -purpose flag to x509 utility to
17368 print out all the purposes.
17372 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17377 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17378 for, obtain and decode and extension and obtain its critical flag.
17379 This allows all the necessary extension code to be handled in a
17380 single function call.
17384 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17385 platforms. See crypto/rc4/rc4_enc.c for further details.
17389 * New -noout option to asn1parse. This causes no output to be produced
17390 its main use is when combined with -strparse and -out to extract data
17391 from a file (which may not be in ASN.1 format).
17395 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17396 when producing the local key id.
17398 *Richard Levitte <levitte@stacken.kth.se>*
17400 * New option -dhparam in s_server. This allows a DH parameter file to be
17401 stated explicitly. If it is not stated then it tries the first server
17402 certificate file. The previous behaviour hard coded the filename
17407 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17408 a public key to be input or output. For example:
17409 openssl rsa -in key.pem -pubout -out pubkey.pem
17410 Also added necessary DSA public key functions to handle this.
17414 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17415 in the message. This was handled by allowing
17416 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17418 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17420 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17421 to the end of the strings whereas this didn't. This would cause problems
17422 if strings read with d2i_ASN1_bytes() were later modified.
17424 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17426 * Fix for base64 decode bug. When a base64 bio reads only one line of
17427 data and it contains EOF it will end up returning an error. This is
17428 caused by input 46 bytes long. The cause is due to the way base64
17429 BIOs find the start of base64 encoded data. They do this by trying a
17430 trial decode on each line until they find one that works. When they
17431 do a flag is set and it starts again knowing it can pass all the
17432 data directly through the decoder. Unfortunately it doesn't reset
17433 the context it uses. This means that if EOF is reached an attempt
17434 is made to pass two EOFs through the context and this causes the
17435 resulting error. This can also cause other problems as well. As is
17436 usual with these problems it takes *ages* to find and the fix is
17437 trivial: move one line.
17439 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17441 * Ugly workaround to get s_client and s_server working under Windows. The
17442 old code wouldn't work because it needed to select() on sockets and the
17443 tty (for keypresses and to see if data could be written). Win32 only
17444 supports select() on sockets so we select() with a 1s timeout on the
17445 sockets and then see if any characters are waiting to be read, if none
17446 are present then we retry, we also assume we can always write data to
17447 the tty. This isn't nice because the code then blocks until we've
17448 received a complete line of data and it is effectively polling the
17449 keyboard at 1s intervals: however it's quite a bit better than not
17450 working at all :-) A dedicated Windows application might handle this
17451 with an event loop for example.
17455 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17456 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17457 will be called when RSA_sign() and RSA_verify() are used. This is useful
17458 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17459 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17460 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17461 This necessitated the support of an extra signature type NID_md5_sha1
17462 for SSL signatures and modifications to the SSL library to use it instead
17463 of calling RSA_public_decrypt() and RSA_private_encrypt().
17467 * Add new -verify -CAfile and -CApath options to the crl program, these
17468 will lookup a CRL issuers certificate and verify the signature in a
17469 similar way to the verify program. Tidy up the crl program so it
17470 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17471 less strict. It will now permit CRL extensions even if it is not
17472 a V2 CRL: this will allow it to tolerate some broken CRLs.
17476 * Initialize all non-automatic variables each time one of the openssl
17477 sub-programs is started (this is necessary as they may be started
17478 multiple times from the "OpenSSL>" prompt).
17480 *Lennart Bang, Bodo Moeller*
17482 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17483 removing all other RSA functionality (this is what NO_RSA does). This
17484 is so (for example) those in the US can disable those operations covered
17485 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17490 * Non-copying interface to BIO pairs.
17491 (still largely untested)
17495 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17496 ASCII string. This was handled independently in various places before.
17500 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17501 UTF8 strings a character at a time.
17505 * Use client_version from client hello to select the protocol
17506 (s23_srvr.c) and for RSA client key exchange verification
17507 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17511 * Add various utility functions to handle SPKACs, these were previously
17512 handled by poking round in the structure internals. Added new function
17513 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17514 print, verify and generate SPKACs. Based on an original idea from
17515 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17519 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17523 * Allow the config file extension section to be overwritten on the
17524 command line. Based on an original idea from Massimiliano Pala
17525 <madwolf@comune.modena.it>. The new option is called -extensions
17526 and can be applied to ca, req and x509. Also -reqexts to override
17527 the request extensions in req and -crlexts to override the crl extensions
17532 * Add new feature to the SPKAC handling in ca. Now you can include
17533 the same field multiple times by preceding it by "XXXX." for example:
17536 this is the same syntax as used in the req config file.
17540 * Allow certificate extensions to be added to certificate requests. These
17541 are specified in a 'req_extensions' option of the req section of the
17542 config file. They can be printed out with the -text option to req but
17543 are otherwise ignored at present.
17547 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17548 data read consists of only the final block it would not decrypted because
17549 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17550 A misplaced 'break' also meant the decrypted final block might not be
17551 copied until the next read.
17555 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17556 a few extra parameters to the DH structure: these will be useful if
17557 for example we want the value of 'q' or implement X9.42 DH.
17561 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17562 provides hooks that allow the default DSA functions or functions on a
17563 "per key" basis to be replaced. This allows hardware acceleration and
17564 hardware key storage to be handled without major modification to the
17565 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17566 associated functions.
17570 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17571 as "read only": it can't be written to and the buffer it points to will
17572 not be freed. Reading from a read only BIO is much more efficient than
17573 a normal memory BIO. This was added because there are several times when
17574 an area of memory needs to be read from a BIO. The previous method was
17575 to create a memory BIO and write the data to it, this results in two
17576 copies of the data and an O(n^2) reading algorithm. There is a new
17577 function BIO_new_mem_buf() which creates a read only memory BIO from
17578 an area of memory. Also modified the PKCS#7 routines to use read only
17583 * Bugfix: ssl23_get_client_hello did not work properly when called in
17584 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17585 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17586 but a retry condition occurred while trying to read the rest.
17590 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17591 NID_pkcs7_encrypted by default: this was wrong since this should almost
17592 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17593 the encrypted data type: this is a more sensible place to put it and it
17594 allows the PKCS#12 code to be tidied up that duplicated this
17599 * Changed obj_dat.pl script so it takes its input and output files on
17600 the command line. This should avoid shell escape redirection problems
17605 * Initial support for certificate extension requests, these are included
17606 in things like Xenroll certificate requests. Included functions to allow
17607 extensions to be obtained and added.
17611 * -crlf option to s_client and s_server for sending newlines as
17612 CRLF (as required by many protocols).
17616 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17618 * Install libRSAglue.a when OpenSSL is built with RSAref.
17620 *Ralf S. Engelschall*
17622 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17624 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17626 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17631 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17632 DH parameters/keys (q is lost during that conversion, but the resulting
17633 DH parameters contain its length).
17635 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17636 much faster than DH_generate_parameters (which creates parameters
17637 where `p = 2*q + 1`), and also the smaller q makes DH computations
17638 much more efficient (160-bit exponentiation instead of 1024-bit
17639 exponentiation); so this provides a convenient way to support DHE
17640 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17641 utter importance to use
17642 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17644 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17645 when such DH parameters are used, because otherwise small subgroup
17646 attacks may become possible!
17650 * Avoid memory leak in i2d_DHparams.
17654 * Allow the -k option to be used more than once in the enc program:
17655 this allows the same encrypted message to be read by multiple recipients.
17659 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17660 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17661 it will always use the numerical form of the OID, even if it has a short
17666 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17667 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17668 otherwise bn_mod_exp was called. In the case of hardware keys for example
17669 no private key components need be present and it might store extra data
17670 in the RSA structure, which cannot be accessed from bn_mod_exp.
17671 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17672 private key operations.
17676 * Added support for SPARC Linux.
17680 * pem_password_cb function type incompatibly changed from
17681 typedef int pem_password_cb(char *buf, int size, int rwflag);
17683 ....(char *buf, int size, int rwflag, void *userdata);
17684 so that applications can pass data to their callbacks:
17685 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17686 additional void * argument, which is just handed through whenever
17687 the password callback is called.
17689 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17691 New function SSL_CTX_set_default_passwd_cb_userdata.
17693 Compatibility note: As many C implementations push function arguments
17694 onto the stack in reverse order, the new library version is likely to
17695 interoperate with programs that have been compiled with the old
17696 pem_password_cb definition (PEM_whatever takes some data that
17697 happens to be on the stack as its last argument, and the callback
17698 just ignores this garbage); but there is no guarantee whatsoever that
17701 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17702 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17703 problems not only on Windows, but also on some Unix platforms.
17704 To avoid problematic command lines, these definitions are now in an
17705 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17706 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17710 * MIPS III/IV assembler module is reimplemented.
17714 * More DES library cleanups: remove references to srand/rand and
17715 delete an unused file.
17719 * Add support for the free Netwide assembler (NASM) under Win32,
17720 since not many people have MASM (ml) and it can be hard to obtain.
17721 This is currently experimental but it seems to work OK and pass all
17722 the tests. Check out INSTALL.W32 for info.
17726 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17727 without temporary keys kept an extra copy of the server key,
17728 and connections with temporary keys did not free everything in case
17733 * New function RSA_check_key and new openssl rsa option -check
17734 for verifying the consistency of RSA keys.
17736 *Ulf Moeller, Bodo Moeller*
17738 * Various changes to make Win32 compile work:
17739 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17740 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17741 comparison" warnings.
17742 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17746 * Add a debugging option to PKCS#5 v2 key generation function: when
17747 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17748 derived keys are printed to stderr.
17752 * Copy the flags in ASN1_STRING_dup().
17754 *Roman E. Pavlov <pre@mo.msk.ru>*
17756 * The x509 application mishandled signing requests containing DSA
17757 keys when the signing key was also DSA and the parameters didn't match.
17759 It was supposed to omit the parameters when they matched the signing key:
17760 the verifying software was then supposed to automatically use the CA's
17761 parameters if they were absent from the end user certificate.
17763 Omitting parameters is no longer recommended. The test was also
17764 the wrong way round! This was probably due to unusual behaviour in
17765 EVP_cmp_parameters() which returns 1 if the parameters match.
17766 This meant that parameters were omitted when they *didn't* match and
17767 the certificate was useless. Certificates signed with 'ca' didn't have
17770 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17772 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17773 The interface is as follows:
17774 Applications can use
17775 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17776 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17777 "off" is now the default.
17778 The library internally uses
17779 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17780 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17781 to disable memory-checking temporarily.
17783 Some inconsistent states that previously were possible (and were
17784 even the default) are now avoided.
17786 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17787 with each memory chunk allocated; this is occasionally more helpful
17788 than just having a counter.
17790 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17792 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17797 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17798 which largely parallels "options", but is for changing API behaviour,
17799 whereas "options" are about protocol behaviour.
17800 Initial "mode" flags are:
17802 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17803 a single record has been written.
17804 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17805 retries use the same buffer location.
17806 (But all of the contents must be
17811 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17814 * Fix problems with no-hmac etc.
17816 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17818 * New functions RSA_get_default_method(), RSA_set_method() and
17819 RSA_get_method(). These allows replacement of RSA_METHODs without having
17820 to mess around with the internals of an RSA structure.
17824 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17825 Also really enable memory leak checks in openssl.c and in some
17828 *Chad C. Mulligan, Bodo Moeller*
17830 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17831 up the length of negative integers. This has now been simplified to just
17832 store the length when it is first determined and use it later, rather
17833 than trying to keep track of where data is copied and updating it to
17835 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17837 * Add a new function PKCS7_signatureVerify. This allows the verification
17838 of a PKCS#7 signature but with the signing certificate passed to the
17839 function itself. This contrasts with PKCS7_dataVerify which assumes the
17840 certificate is present in the PKCS#7 structure. This isn't always the
17841 case: certificates can be omitted from a PKCS#7 structure and be
17842 distributed by "out of band" means (such as a certificate database).
17846 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17847 function prototypes in pem.h, also change util/mkdef.pl to add the
17848 necessary function names.
17852 * mk1mf.pl (used by Windows builds) did not properly read the
17853 options set by Configure in the top level Makefile, and Configure
17854 was not even able to write more than one option correctly.
17855 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17859 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17860 file to be loaded from a BIO or FILE pointer. The BIO version will
17861 for example allow memory BIOs to contain config info.
17865 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17866 Whoever hopes to achieve shared-library compatibility across versions
17867 must use this, not the compile-time macro.
17868 (Exercise 0.9.4: Which is the minimum library version required by
17870 Note: All this applies only to multi-threaded programs, others don't
17875 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17876 through a BIO pair triggered the default case, i.e.
17877 SSLerr(...,SSL_R_UNKNOWN_STATE).
17881 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17882 can use the SSL library even if none of the specific BIOs is
17887 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17888 for the encoded length.
17890 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17892 * Add initial documentation of the X509V3 functions.
17896 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17897 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17898 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17899 secure PKCS#8 private key format with a high iteration count.
17903 * Fix determination of Perl interpreter: A perl or perl5
17904 *directory* in $PATH was also accepted as the interpreter.
17906 *Ralf S. Engelschall*
17908 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17909 wrong with it but it was very old and did things like calling
17910 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17911 unusual formatting.
17915 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17916 to use the new extension code.
17920 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17921 with macros. This should make it easier to change their form, add extra
17922 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17927 * Add to configuration table a new entry that can specify an alternative
17928 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17929 according to Mark Crispin <MRC@Panda.COM>.
17933 * DES CBC did not update the IV. Weird.
17937 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17938 Changing the behaviour of the former might break existing programs --
17939 where IV updating is needed, des_ncbc_encrypt can be used.
17942 * When bntest is run from "make test" it drives bc to check its
17943 calculations, as well as internally checking them. If an internal check
17944 fails, it needs to cause bc to give a non-zero result or make test carries
17945 on without noticing the failure. Fixed.
17949 * DES library cleanups.
17953 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17954 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17955 ciphers. NOTE: although the key derivation function has been verified
17956 against some published test vectors it has not been extensively tested
17957 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17962 * Instead of "mkdir -p", which is not fully portable, use new
17963 Perl script "util/mkdir-p.pl".
17967 * Rewrite the way password based encryption (PBE) is handled. It used to
17968 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17969 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17970 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17971 the 'parameter' field of the AlgorithmIdentifier is passed to the
17972 underlying key generation function so it must do its own ASN1 parsing.
17973 This has also changed the EVP_PBE_CipherInit() function which now has a
17974 'parameter' argument instead of literal salt and iteration count values
17975 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17979 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17980 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17981 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17982 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17983 value was just used as a "magic string" and not used directly its
17984 value doesn't matter.
17988 * Introduce some semblance of const correctness to BN. Shame C doesn't
17993 * "linux-sparc64" configuration (ultrapenguin).
17995 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17996 "linux-sparc" configuration.
17998 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18000 * config now generates no-xxx options for missing ciphers.
18004 * Support the EBCDIC character set (work in progress).
18005 File ebcdic.c not yet included because it has a different license.
18007 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18009 * Support BS2000/OSD-POSIX.
18011 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18013 * Make callbacks for key generation use `void *` instead of `char *`.
18017 * Make S/MIME samples compile (not yet tested).
18021 * Additional typesafe stacks.
18025 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18029 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18031 * New configuration variant "sco5-gcc".
18033 * Updated some demos.
18035 *Sean O Riordain, Wade Scholine*
18037 * Add missing BIO_free at exit of pkcs12 application.
18041 * Fix memory leak in conf.c.
18045 * Updates for Win32 to assembler version of MD5.
18049 * Set #! path to perl in `apps/der_chop` to where we found it
18050 instead of using a fixed path.
18054 * SHA library changes for irix64-mips4-cc.
18058 * Improvements for VMS support.
18062 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18064 * Bignum library bug fix. IRIX 6 passes "make test" now!
18065 This also avoids the problems with SC4.2 and unpatched SC5.
18067 *Andy Polyakov <appro@fy.chalmers.se>*
18069 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18070 These are required because of the typesafe stack would otherwise break
18071 existing code. If old code used a structure member which used to be STACK
18072 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18073 sk_num or sk_value it would produce an error because the num, data members
18074 are not present in STACK_OF. Now it just produces a warning. sk_set
18075 replaces the old method of assigning a value to sk_value
18076 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18077 that does this will no longer work (and should use sk_set instead) but
18078 this could be regarded as a "questionable" behaviour anyway.
18082 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18083 correctly handle encrypted S/MIME data.
18087 * Change type of various DES function arguments from des_cblock
18088 (which means, in function argument declarations, pointer to char)
18089 to des_cblock * (meaning pointer to array with 8 char elements),
18090 which allows the compiler to do more typechecking; it was like
18091 that back in SSLeay, but with lots of ugly casts.
18093 Introduce new type const_des_cblock.
18097 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18098 problems: find RecipientInfo structure that matches recipient certificate
18099 and initialise the ASN1 structures properly based on passed cipher.
18103 * Belatedly make the BN tests actually check the results.
18107 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18108 to and from BNs: it was completely broken. New compilation option
18109 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18110 key elements as negative integers.
18114 * Reorganize and speed up MD5.
18116 *Andy Polyakov <appro@fy.chalmers.se>*
18120 *Richard Levitte <richard@levitte.org>*
18122 * New option -out to asn1parse to allow the parsed structure to be
18123 output to a file. This is most useful when combined with the -strparse
18124 option to examine the output of things like OCTET STRINGS.
18128 * Make SSL library a little more fool-proof by not requiring any longer
18129 that `SSL_set_{accept,connect}_state` be called before
18130 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18131 in many applications because usually everything *appeared* to work as
18132 intended anyway -- now it really works as intended).
18136 * Move openssl.cnf out of lib/.
18140 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18141 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18142 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18144 *Ralf S. Engelschall*
18146 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18147 handle PKCS#7 enveloped data properly.
18149 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18151 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18152 copying pointers. The cert_st handling is changed by this in
18153 various ways (and thus what used to be known as ctx->default_cert
18154 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18155 any longer when s->cert does not give us what we need).
18156 ssl_cert_instantiate becomes obsolete by this change.
18157 As soon as we've got the new code right (possibly it already is?),
18158 we have solved a couple of bugs of the earlier code where s->cert
18159 was used as if it could not have been shared with other SSL structures.
18161 Note that using the SSL API in certain dirty ways now will result
18162 in different behaviour than observed with earlier library versions:
18163 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18164 does not influence s as it used to.
18166 In order to clean up things more thoroughly, inside SSL_SESSION
18167 we don't use CERT any longer, but a new structure SESS_CERT
18168 that holds per-session data (if available); currently, this is
18169 the peer's certificate chain and, for clients, the server's certificate
18170 and temporary key. CERT holds only those values that can have
18171 meaningful defaults in an SSL_CTX.
18175 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18176 from the internal representation. Various PKCS#7 fixes: remove some
18177 evil casts and set the enc_dig_alg field properly based on the signing
18182 * Allow PKCS#12 password to be set from the command line or the
18183 environment. Let 'ca' get its config file name from the environment
18184 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18189 * Allow certificate policies extension to use an IA5STRING for the
18190 organization field. This is contrary to the PKIX definition but
18191 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18196 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18197 without disallowing inline assembler and the like for non-pedantic builds.
18201 * Support Borland C++ builder.
18203 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18209 * SHA-1 cleanups and performance enhancements.
18211 *Andy Polyakov <appro@fy.chalmers.se>*
18213 * Sparc v8plus assembler for the bignum library.
18215 *Andy Polyakov <appro@fy.chalmers.se>*
18217 * Accept any -xxx and +xxx compiler options in Configure.
18221 * Update HPUX configuration.
18225 * Add missing `sk_<type>_unshift()` function to safestack.h
18227 *Ralf S. Engelschall*
18229 * New function SSL_CTX_use_certificate_chain_file that sets the
18230 "extra_cert"s in addition to the certificate. (This makes sense
18231 only for "PEM" format files, as chains as a whole are not
18236 * Support verify_depth from the SSL API.
18237 x509_vfy.c had what can be considered an off-by-one-error:
18238 Its depth (which was not part of the external interface)
18239 was actually counting the number of certificates in a chain;
18240 now it really counts the depth.
18244 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18245 instead of X509err, which often resulted in confusing error
18246 messages since the error codes are not globally unique
18247 (e.g. an alleged error in ssl3_accept when a certificate
18248 didn't match the private key).
18250 * New function SSL_CTX_set_session_id_context that allows to set a default
18251 value (so that you don't need SSL_set_session_id_context for each
18252 connection using the SSL_CTX).
18256 * OAEP decoding bug fix.
18260 * Support INSTALL_PREFIX for package builders, as proposed by
18265 * New Configure options "threads" and "no-threads". For systems
18266 where the proper compiler options are known (currently Solaris
18267 and Linux), "threads" is the default.
18271 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18275 * Install various scripts to $(OPENSSLDIR)/misc, not to
18276 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18277 such as /usr/local/bin.
18281 * "make linux-shared" to build shared libraries.
18283 *Niels Poppe <niels@netbox.org>*
18285 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18289 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18290 extension adding in x509 utility.
18294 * Remove NOPROTO sections and error code comments.
18298 * Partial rewrite of the DEF file generator to now parse the ANSI
18303 * New Configure options --prefix=DIR and --openssldir=DIR.
18307 * Complete rewrite of the error code script(s). It is all now handled
18308 by one script at the top level which handles error code gathering,
18309 header rewriting and C source file generation. It should be much better
18310 than the old method: it now uses a modified version of Ulf's parser to
18311 read the ANSI prototypes in all header files (thus the old K&R definitions
18312 aren't needed for error creation any more) and do a better job of
18313 translating function codes into names. The old 'ASN1 error code embedded
18314 in a comment' is no longer necessary and it doesn't use .err files which
18315 have now been deleted. Also the error code call doesn't have to appear all
18316 on one line (which resulted in some large lines...).
18320 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18324 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18325 0 (which usually indicates a closed connection), but continue reading.
18329 * Fix some race conditions.
18333 * Add support for CRL distribution points extension. Add Certificate
18334 Policies and CRL distribution points documentation.
18338 * Move the autogenerated header file parts to crypto/opensslconf.h.
18342 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18343 8 of keying material. Merlin has also confirmed interop with this fix
18344 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18346 *Merlin Hughes <merlin@baltimore.ie>*
18348 * Fix lots of warnings.
18350 *Richard Levitte <levitte@stacken.kth.se>*
18352 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18353 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18355 *Richard Levitte <levitte@stacken.kth.se>*
18357 * Fix problems with sizeof(long) == 8.
18359 *Andy Polyakov <appro@fy.chalmers.se>*
18361 * Change functions to ANSI C.
18365 * Fix typos in error codes.
18367 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18369 * Remove defunct assembler files from Configure.
18373 * SPARC v8 assembler BIGNUM implementation.
18375 *Andy Polyakov <appro@fy.chalmers.se>*
18377 * Support for Certificate Policies extension: both print and set.
18378 Various additions to support the r2i method this uses.
18382 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18383 return a const string when you are expecting an allocated buffer.
18387 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18388 types DirectoryString and DisplayText.
18392 * Add code to allow r2i extensions to access the configuration database,
18393 add an LHASH database driver and add several ctx helper functions.
18397 * Fix an evil bug in bn_expand2() which caused various BN functions to
18398 fail when they extended the size of a BIGNUM.
18402 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18403 support typesafe stack.
18407 * Fix typo in SSL_[gs]et_options().
18409 *Nils Frostberg <nils@medcom.se>*
18411 * Delete various functions and files that belonged to the (now obsolete)
18412 old X509V3 handling code.
18416 * New Configure option "rsaref".
18420 * Don't auto-generate pem.h.
18424 * Introduce type-safe ASN.1 SETs.
18428 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18430 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18432 * Introduce type-safe STACKs. This will almost certainly break lots of code
18433 that links with OpenSSL (well at least cause lots of warnings), but fear
18434 not: the conversion is trivial, and it eliminates loads of evil casts. A
18435 few STACKed things have been converted already. Feel free to convert more.
18436 In the fullness of time, I'll do away with the STACK type altogether.
18440 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18441 specified in `<certfile>` by updating the entry in the index.txt file.
18442 This way one no longer has to edit the index.txt file manually for
18443 revoking a certificate. The -revoke option does the gory details now.
18445 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18447 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18448 `-text` option at all and this way the `-noout -text` combination was
18449 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18451 *Ralf S. Engelschall*
18453 * Make sure a corresponding plain text error message exists for the
18454 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18455 verify callback function determined that a certificate was revoked.
18457 *Ralf S. Engelschall*
18459 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18460 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18461 all available ciphers including rc5, which was forgotten until now.
18462 In order to let the testing shell script know which algorithms
18463 are available, a new (up to now undocumented) command
18464 `openssl list-cipher-commands` is used.
18468 * Bugfix: s_client occasionally would sleep in select() when
18469 it should have checked SSL_pending() first.
18473 * New functions DSA_do_sign and DSA_do_verify to provide access to
18474 the raw DSA values prior to ASN.1 encoding.
18478 * Tweaks to Configure
18480 *Niels Poppe <niels@netbox.org>*
18482 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18487 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18491 * New config option to avoid instructions that are illegal on the 80386.
18492 The default code is faster, but requires at least a 486.
18496 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18497 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18498 same as SSL2_VERSION anyway.
18502 * New "-showcerts" option for s_client.
18506 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18507 application. Various cleanups and fixes.
18511 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18512 modify error routines to work internally. Add error codes and PBE init
18513 to library startup routines.
18517 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18518 packing functions to asn1 and evp. Changed function names and error
18519 codes along the way.
18523 * PKCS12 integration: and so it begins... First of several patches to
18524 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18525 objects to objects.h
18529 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18530 and display support for Thawte strong extranet extension.
18534 * Add LinuxPPC support.
18536 *Jeff Dubrule <igor@pobox.org>*
18538 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18539 bn_div_words in alpha.s.
18541 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18543 * Make sure the RSA OAEP test is skipped under -DRSAref because
18544 OAEP isn't supported when OpenSSL is built with RSAref.
18546 *Ulf Moeller <ulf@fitug.de>*
18548 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18549 so they no longer are missing under -DNOPROTO.
18551 *Soren S. Jorvang <soren@t.dk>*
18553 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18555 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18556 doesn't work when the session is reused. Coming soon!
18560 * Fix a security hole, that allows sessions to be reused in the wrong
18561 context thus bypassing client cert protection! All software that uses
18562 client certs and session caches in multiple contexts NEEDS PATCHING to
18563 allow session reuse! A fuller solution is in the works.
18565 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18567 * Some more source tree cleanups (removed obsolete files
18568 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18569 permission on "config" script to be executable) and a fix for the INSTALL
18572 *Ulf Moeller <ulf@fitug.de>*
18574 * Remove some legacy and erroneous uses of malloc, free instead of
18577 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18579 * Make rsa_oaep_test return non-zero on error.
18581 *Ulf Moeller <ulf@fitug.de>*
18583 * Add support for native Solaris shared libraries. Configure
18584 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18585 if someone would make that last step automatic.
18587 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18589 * ctx_size was not built with the right compiler during "make links". Fixed.
18593 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18594 except NULL ciphers". This means the default cipher list will no longer
18595 enable NULL ciphers. They need to be specifically enabled e.g. with
18596 the string "DEFAULT:eNULL".
18600 * Fix to RSA private encryption routines: if p < q then it would
18601 occasionally produce an invalid result. This will only happen with
18602 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18606 * Be less restrictive and allow also `perl util/perlpath.pl
18607 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18608 because this way one can also use an interpreter named `perl5` (which is
18609 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18610 installed as `perl`).
18612 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18614 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18616 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18618 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18619 advapi32.lib to Win32 build and change the pem test comparison
18620 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18621 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18622 and crypto/des/ede_cbcm_enc.c.
18626 * DES quad checksum was broken on big-endian architectures. Fixed.
18630 * Comment out two functions in bio.h that aren't implemented. Fix up the
18631 Win32 test batch file so it (might) work again. The Win32 test batch file
18632 is horrible: I feel ill....
18636 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18637 in e_os.h. Audit of header files to check ANSI and non ANSI
18638 sections: 10 functions were absent from non ANSI section and not exported
18639 from Windows DLLs. Fixed up libeay.num for new functions.
18643 * Make `openssl version` output lines consistent.
18645 *Ralf S. Engelschall*
18647 * Fix Win32 symbol export lists for BIO functions: Added
18648 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18649 to ms/libeay{16,32}.def.
18651 *Ralf S. Engelschall*
18653 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18654 fine under Unix and passes some trivial tests I've now added. But the
18655 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18656 added to make sure no one expects that this stuff really works in the
18657 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18658 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18661 *Ralf S. Engelschall*
18663 * Fix the generation of two part addresses in perl.
18665 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18667 * Add config entry for Linux on MIPS.
18669 *John Tobey <jtobey@channel1.com>*
18671 * Make links whenever Configure is run, unless we are on Windoze.
18675 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18676 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18681 * Add a useful kludge to allow package maintainers to specify compiler and
18682 other platforms details on the command line without having to patch the
18683 Configure script every time: One now can use
18684 `perl Configure <id>:<details>`,
18685 i.e. platform ids are allowed to have details appended
18686 to them (separated by colons). This is treated as there would be a static
18687 pre-configured entry in Configure's %table under key `<id>` with value
18688 `<details>` and `perl Configure <id>` is called. So, when you want to
18689 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18690 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18691 now, which overrides the FreeBSD-elf entry on-the-fly.
18693 *Ralf S. Engelschall*
18695 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18699 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18700 on the `perl Configure ...` command line. This way one can compile
18701 OpenSSL libraries with Position Independent Code (PIC) which is needed
18702 for linking it into DSOs.
18704 *Ralf S. Engelschall*
18706 * Remarkably, export ciphers were totally broken and no-one had noticed!
18711 * Cleaned up the LICENSE document: The official contact for any license
18712 questions now is the OpenSSL core team under openssl-core@openssl.org.
18713 And add a paragraph about the dual-license situation to make sure people
18714 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18715 to the OpenSSL toolkit.
18717 *Ralf S. Engelschall*
18719 * General source tree makefile cleanups: Made `making xxx in yyy...`
18720 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18721 Additionally cleaned up the `make links` target: Remove unnecessary
18722 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18723 to speed processing and no longer clutter the display with confusing
18724 stuff. Instead only the actually done links are displayed.
18726 *Ralf S. Engelschall*
18728 * Permit null encryption ciphersuites, used for authentication only. It used
18729 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18730 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18735 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18736 signed attributes when verifying signatures (this would break them),
18737 the detached data encoding was wrong and public keys obtained using
18738 X509_get_pubkey() weren't freed.
18742 * Add text documentation for the BUFFER functions. Also added a work around
18743 to a Win95 console bug. This was triggered by the password read stuff: the
18744 last character typed gets carried over to the next fread(). If you were
18745 generating a new cert request using 'req' for example then the last
18746 character of the passphrase would be CR which would then enter the first
18751 * Added the new 'Includes OpenSSL Cryptography Software' button as
18752 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18753 button and can be used by applications based on OpenSSL to show the
18754 relationship to the OpenSSL project.
18756 *Ralf S. Engelschall*
18758 * Remove confusing variables in function signatures in files
18759 ssl/ssl_lib.c and ssl/ssl.h.
18761 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18763 * Don't install bss_file.c under PREFIX/include/
18765 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18767 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18768 functions that return function pointers and has support for NT specific
18769 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18770 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18771 unsigned to signed types: this was killing the Win32 compile.
18775 * Add new certificate file to stack functions,
18776 SSL_add_dir_cert_subjects_to_stack() and
18777 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18778 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18779 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18780 This means that Apache-SSL and similar packages don't have to mess around
18781 to add as many CAs as they want to the preferred list.
18785 * Experiment with doxygen documentation. Currently only partially applied to
18787 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18788 openssl.doxy as the configuration file.
18792 * Get rid of remaining C++-style comments which strict C compilers hate.
18794 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18796 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18797 compiled in by default: it has problems with large keys.
18801 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18802 DH private keys and/or callback functions which directly correspond to
18803 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18804 is needed for applications which have to configure certificates on a
18805 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18807 For the RSA certificate situation is makes no difference, but
18808 for the DSA certificate situation this fixes the "no shared cipher"
18809 problem where the OpenSSL cipher selection procedure failed because the
18810 temporary keys were not overtaken from the context and the API provided
18811 no way to reconfigure them.
18812 The new functions now let applications reconfigure the stuff and they
18813 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18814 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18815 non-public-API function ssl_cert_instantiate() is used as a helper
18816 function and also to reduce code redundancy inside ssl_rsa.c.
18818 *Ralf S. Engelschall*
18820 * Move s_server -dcert and -dkey options out of the undocumented feature
18821 area because they are useful for the DSA situation and should be
18822 recognized by the users.
18824 *Ralf S. Engelschall*
18826 * Fix the cipher decision scheme for export ciphers: the export bits are
18827 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18828 SSL_EXP_MASK. So, the original variable has to be used instead of the
18829 already masked variable.
18831 *Richard Levitte <levitte@stacken.kth.se>*
18833 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18835 *Richard Levitte <levitte@stacken.kth.se>*
18837 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18838 from `int` to `unsigned int` because it is a length and initialized by
18839 EVP_DigestFinal() which expects an `unsigned int *`.
18841 *Richard Levitte <levitte@stacken.kth.se>*
18843 * Don't hard-code path to Perl interpreter on shebang line of Configure
18844 script. Instead use the usual Shell->Perl transition trick.
18846 *Ralf S. Engelschall*
18848 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18849 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18850 -noout -modulus` as it's already the case for `openssl rsa -noout
18851 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18852 currently the public key is printed (a decision which was already done by
18853 `openssl dsa -modulus` in the past) which serves a similar purpose.
18854 Additionally the NO_RSA no longer completely removes the whole -modulus
18855 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18858 *Ralf S. Engelschall*
18860 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18861 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18863 *Arne Ansper <arne@ats.cyber.ee>*
18865 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18866 to be added. Now both 'req' and 'ca' can use new objects defined in the
18871 * Add cool BIO that does syslog (or event log on NT).
18873 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18875 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18876 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18877 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18878 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18882 * Add preliminary config info for new extension code.
18886 * Make RSA_NO_PADDING really use no padding.
18888 *Ulf Moeller <ulf@fitug.de>*
18890 * Generate errors when private/public key check is done.
18894 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18895 for some CRL extensions and new objects added.
18899 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18900 key usage extension and fuller support for authority key id.
18904 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18905 padding method for RSA, which is recommended for new applications in PKCS
18906 #1 v2.0 (RFC 2437, October 1998).
18907 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18908 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18909 against Bleichbacher's attack on RSA.
18910 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18913 * Updates to the new SSL compression code
18915 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18917 * Fix so that the version number in the master secret, when passed
18918 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18919 (because the server will not accept higher), that the version number
18920 is 0x03,0x01, not 0x03,0x00
18922 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18924 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18925 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18926 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18930 * Support for RAW extensions where an arbitrary extension can be
18931 created by including its DER encoding. See `apps/openssl.cnf` for
18936 * Make sure latest Perl versions don't interpret some generated C array
18937 code as Perl array code in the crypto/err/err_genc.pl script.
18939 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18941 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18942 not many people have the assembler. Various Win32 compilation fixes and
18943 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18944 build instructions.
18948 * Modify configure script 'Configure' to automatically create crypto/date.h
18949 file under Win32 and also build pem.h from pem.org. New script
18950 util/mkfiles.pl to create the MINFO file on environments that can't do a
18951 'make files': perl util/mkfiles.pl >MINFO should work.
18955 * Major rework of DES function declarations, in the pursuit of correctness
18956 and purity. As a result, many evil casts evaporated, and some weirdness,
18957 too. You may find this causes warnings in your code. Zapping your evil
18958 casts will probably fix them. Mostly.
18962 * Fix for a typo in asn1.h. Bug fix to object creation script
18963 obj_dat.pl. It considered a zero in an object definition to mean
18964 "end of object": none of the objects in objects.h have any zeros
18965 so it wasn't spotted.
18967 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18969 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18970 Masking (CBCM). In the absence of test vectors, the best I have been able
18971 to do is check that the decrypt undoes the encrypt, so far. Send me test
18972 vectors if you have them.
18976 * Correct calculation of key length for export ciphers (too much space was
18977 allocated for null ciphers). This has not been tested!
18981 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18982 message is now correct (it understands "crypto" and "ssl" on its
18983 command line). There is also now an "update" option. This will update
18984 the util/ssleay.num and util/libeay.num files with any new functions.
18986 perl util/mkdef.pl crypto ssl update
18987 it will update them.
18991 * Overhauled the Perl interface:
18992 - ported BN stuff to OpenSSL's different BN library
18993 - made the perl/ source tree CVS-aware
18994 - renamed the package from SSLeay to OpenSSL (the files still contain
18995 their history because I've copied them in the repository)
18996 - removed obsolete files (the test scripts will be replaced
18997 by better Test::Harness variants in the future)
18999 *Ralf S. Engelschall*
19001 * First cut for a very conservative source tree cleanup:
19002 1. merge various obsolete readme texts into doc/ssleay.txt
19003 where we collect the old documents and readme texts.
19004 2. remove the first part of files where I'm already sure that we no
19005 longer need them because of three reasons: either they are just temporary
19006 files which were left by Eric or they are preserved original files where
19007 I've verified that the diff is also available in the CVS via "cvs diff
19008 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19009 the crypto/md/ stuff).
19011 *Ralf S. Engelschall*
19013 * More extension code. Incomplete support for subject and issuer alt
19014 name, issuer and authority key id. Change the i2v function parameters
19015 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19016 what that's for :-) Fix to ASN1 macro which messed up
19017 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19021 * Preliminary support for ENUMERATED type. This is largely copied from the
19026 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19028 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19030 * Make sure `make rehash` target really finds the `openssl` program.
19032 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19034 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19035 like to hear about it if this slows down other processors.
19039 * Add CygWin32 platform information to Configure script.
19041 *Alan Batie <batie@aahz.jf.intel.com>*
19043 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19045 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19047 * New program nseq to manipulate netscape certificate sequences
19051 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19056 * Fixes to BN code. Previously the default was to define BN_RECURSION
19057 but the BN code had some problems that would cause failures when
19058 doing certificate verification and some other functions.
19060 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19062 * Add ASN1 and PEM code to support netscape certificate sequences.
19066 * Add ASN1 and PEM code to support netscape certificate sequences.
19070 * Add several PKIX and private extended key usage OIDs.
19074 * Modify the 'ca' program to handle the new extension code. Modify
19075 openssl.cnf for new extension format, add comments.
19079 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19080 and add a sample to openssl.cnf so req -x509 now adds appropriate
19085 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19086 error code, add initial support to X509_print() and x509 application.
19090 * Takes a deep breath and start adding X509 V3 extension support code. Add
19091 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19092 stuff is currently isolated and isn't even compiled yet.
19096 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19097 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19098 Removed the versions check from X509 routines when loading extensions:
19099 this allows certain broken certificates that don't set the version
19100 properly to be processed.
19104 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19105 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19106 can still be regenerated with "make depend".
19110 * Spelling mistake in C version of CAST-128.
19112 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19114 * Changes to the error generation code. The perl script err-code.pl
19115 now reads in the old error codes and retains the old numbers, only
19116 adding new ones if necessary. It also only changes the .err files if new
19117 codes are added. The makefiles have been modified to only insert errors
19118 when needed (to avoid needlessly modifying header files). This is done
19119 by only inserting errors if the .err file is newer than the auto generated
19120 C file. To rebuild all the error codes from scratch (the old behaviour)
19121 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19122 or delete all the .err files.
19126 * CAST-128 was incorrectly implemented for short keys. The C version has
19127 been fixed, but is untested. The assembler versions are also fixed, but
19128 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19129 to regenerate it if needed.
19130 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19131 Hagino <itojun@kame.net>*
19133 * File was opened incorrectly in randfile.c.
19135 *Ulf Möller <ulf@fitug.de>*
19137 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19138 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19139 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19140 al: it's just almost always a UTCTime. Note this patch adds new error
19141 codes so do a "make errors" if there are problems.
19145 * Correct Linux 1 recognition in config.
19147 *Ulf Möller <ulf@fitug.de>*
19149 * Remove pointless MD5 hash when using DSA keys in ca.
19151 *Anonymous <nobody@replay.com>*
19153 * Generate an error if given an empty string as a cert directory. Also
19154 generate an error if handed NULL (previously returned 0 to indicate an
19155 error, but didn't set one).
19157 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19159 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19163 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19164 parameters. This was causing a warning which killed off the Win32 compile.
19168 * Remove C++ style comments from crypto/bn/bn_local.h.
19170 *Neil Costigan <neil.costigan@celocom.com>*
19172 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19173 based on a text string, looking up short and long names and finally
19174 "dot" format. The "dot" format stuff didn't work. Added new function
19175 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19176 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19177 OID is not part of the table.
19181 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19182 X509_LOOKUP_by_alias().
19186 * Sort openssl functions by name.
19190 * Get the `gendsa` command working and add it to the `list` command. Remove
19191 encryption from sample DSA keys (in case anyone is interested the password
19196 * Make *all* `*_free` functions accept a NULL pointer.
19198 *Frans Heymans <fheymans@isaserver.be>*
19200 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19203 *Anonymous <nobody@replay.com>*
19205 * s_server should send the CAfile as acceptable CAs, not its own cert.
19207 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19209 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19211 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19213 * Temp key "for export" tests were wrong in s3_srvr.c.
19215 *Anonymous <nobody@replay.com>*
19217 * Add prototype for temp key callback functions
19218 SSL_CTX_set_tmp_{rsa,dh}_callback().
19222 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19223 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19227 * X509_name_add_entry() freed the wrong thing after an error.
19229 *Arne Ansper <arne@ats.cyber.ee>*
19231 * rsa_eay.c would attempt to free a NULL context.
19233 *Arne Ansper <arne@ats.cyber.ee>*
19235 * BIO_s_socket() had a broken should_retry() on Windoze.
19237 *Arne Ansper <arne@ats.cyber.ee>*
19239 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19241 *Arne Ansper <arne@ats.cyber.ee>*
19243 * Make sure the already existing X509_STORE->depth variable is initialized
19244 in X509_STORE_new(), but document the fact that this variable is still
19245 unused in the certificate verification process.
19247 *Ralf S. Engelschall*
19249 * Fix the various library and `apps/` files to free up pkeys obtained from
19250 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19254 * Fix reference counting in X509_PUBKEY_get(). This makes
19255 demos/maurice/example2.c work, amongst others, probably.
19257 *Steve Henson and Ben Laurie*
19259 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19260 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19261 are no longer created. This way we have a single and consistent command
19262 line interface `openssl <command>`, similar to `cvs <command>`.
19264 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19266 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19267 BIT STRING wrapper always have zero unused bits.
19271 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19275 * Make the top-level INSTALL documentation easier to understand.
19279 * Makefiles updated to exit if an error occurs in a sub-directory
19280 make (including if user presses ^C) [Paul Sutton]
19282 * Make Montgomery context stuff explicit in RSA data structure.
19286 * Fix build order of pem and err to allow for generated pem.h.
19290 * Fix renumbering bug in X509_NAME_delete_entry().
19294 * Enhanced the err-ins.pl script so it makes the error library number
19295 global and can add a library name. This is needed for external ASN1 and
19296 other error libraries.
19300 * Fixed sk_insert which never worked properly.
19304 * Fix ASN1 macros so they can handle indefinite length constructed
19305 EXPLICIT tags. Some non standard certificates use these: they can now
19310 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19311 into a single doc/ssleay.txt bundle. This way the information is still
19312 preserved but no longer messes up this directory. Now it's new room for
19313 the new set of documentation files.
19315 *Ralf S. Engelschall*
19317 * SETs were incorrectly DER encoded. This was a major pain, because they
19318 shared code with SEQUENCEs, which aren't coded the same. This means that
19319 almost everything to do with SETs or SEQUENCEs has either changed name or
19320 number of arguments.
19322 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19324 * Fix test data to work with the above.
19328 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19329 was already fixed by Eric for 0.9.1 it seems.
19331 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19333 * Autodetect FreeBSD3.
19337 * Fix various bugs in Configure. This affects the following platforms:
19341 unixware-2.0-pentium
19346 * Eliminate generated files from CVS. Reorder tests to regenerate files
19347 before they are needed.
19351 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19355 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19357 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19358 changed SSLeay to OpenSSL in version strings.
19360 *Ralf S. Engelschall*
19362 * Some fixups to the top-level documents.
19366 * Fixed the nasty bug where rsaref.h was not found under compile-time
19367 because the symlink to include/ was missing.
19369 *Ralf S. Engelschall*
19371 * Incorporated the popular no-RSA/DSA-only patches
19372 which allow to compile a RSA-free SSLeay.
19374 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19376 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19377 when "ssleay" is still not found.
19379 *Ralf S. Engelschall*
19381 * Added more platforms to Configure: Cray T3E, HPUX 11,
19383 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19385 * Updated the README file.
19387 *Ralf S. Engelschall*
19389 * Added various .cvsignore files in the CVS repository subdirs
19390 to make a "cvs update" really silent.
19392 *Ralf S. Engelschall*
19394 * Recompiled the error-definition header files and added
19395 missing symbols to the Win32 linker tables.
19397 *Ralf S. Engelschall*
19399 * Cleaned up the top-level documents;
19400 o new files: CHANGES and LICENSE
19401 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19402 o merged COPYRIGHT into LICENSE
19403 o removed obsolete TODO file
19404 o renamed MICROSOFT to INSTALL.W32
19406 *Ralf S. Engelschall*
19408 * Removed dummy files from the 0.9.1b source tree:
19409 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19410 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19411 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19412 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19413 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19415 *Ralf S. Engelschall*
19417 * Added various platform portability fixes.
19421 * The Genesis of the OpenSSL rpject:
19422 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19423 Young and Tim J. Hudson created while they were working for C2Net until
19426 *The OpenSSL Project*
19428 ### Changes between 0.9.0b and 0.9.1b [not released]
19430 * Updated a few CA certificates under certs/
19434 * Changed some BIGNUM api stuff.
19438 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19439 DGUX x86, Linux Alpha, etc.
19443 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19444 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19449 * Add -strparse option to asn1pars program which parses nested
19452 *Dr Stephen Henson <shenson@bigfoot.com>*
19454 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19458 * DSA fix for "ca" program.
19462 * Added "-genkey" option to "dsaparam" program.
19466 * Added RIPE MD160 (rmd160) message digest.
19470 * Added -a (all) option to "ssleay version" command.
19474 * Added PLATFORM define which is the id given to Configure.
19478 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19482 * Extended the ASN.1 parser routines.
19486 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19490 * Added a BN_CTX to the BN library.
19494 * Fixed the weak key values in DES library
19498 * Changed API in EVP library for cipher aliases.
19502 * Added support for RC2/64bit cipher.
19506 * Converted the lhash library to the crypto/mem.c functions.
19510 * Added more recognized ASN.1 object ids.
19514 * Added more RSA padding checks for SSL/TLS.
19518 * Added BIO proxy/filter functionality.
19522 * Added extra_certs to SSL_CTX which can be used
19523 send extra CA certificates to the client in the CA cert chain sending
19524 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19528 * Now Fortezza is denied in the authentication phase because
19529 this is key exchange mechanism is not supported by SSLeay at all.
19533 * Additional PKCS1 checks.
19537 * Support the string "TLSv1" for all TLS v1 ciphers.
19541 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19542 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19546 * Fixed a few memory leaks.
19550 * Fixed various code and comment typos.
19554 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19555 bytes sent in the client random.
19557 *Edward Bishop <ebishop@spyglass.com>*
19561 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19562 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19563 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19564 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19565 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19566 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19567 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19568 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19569 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19570 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19571 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19572 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19573 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19574 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19575 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19576 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19577 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19578 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19579 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19580 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19581 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19582 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19583 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19584 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19585 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19586 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19587 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19588 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19589 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19590 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19591 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19592 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19593 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19594 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19595 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19596 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19597 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19598 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19599 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19600 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19601 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19602 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19603 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19604 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19605 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19606 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19607 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19608 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19609 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19610 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19611 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19612 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19613 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19614 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19615 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19616 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19617 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19618 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19619 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19620 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19621 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19622 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19623 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19624 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19625 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19626 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19627 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19628 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19629 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19630 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19631 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19632 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19633 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19634 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19635 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19636 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19637 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19638 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19639 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19640 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19641 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19642 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19643 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19644 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19645 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19646 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19647 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19648 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19649 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19650 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19651 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19652 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19653 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19654 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19655 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19656 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19657 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19658 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19659 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19660 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19661 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19662 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19663 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19664 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19665 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19666 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19667 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19668 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19669 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19670 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19671 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19672 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19673 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19674 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19675 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19676 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19677 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19678 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19679 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19680 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19681 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19682 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19683 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19684 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19685 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19686 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19687 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19688 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19689 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19690 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19691 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19692 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19693 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19694 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19695 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19696 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19697 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19698 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19699 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19700 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19701 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19702 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19703 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19704 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19705 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19706 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19707 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19708 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19709 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19710 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19711 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19712 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19713 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19714 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19715 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19716 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19717 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19718 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19719 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19720 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19721 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19722 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655