4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 alpha 1 [1 Dec 2022]
27 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
28 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
29 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
30 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
31 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
32 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
33 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
34 for legacy EC and SM2 keys is also changed similarly to honor the
35 equivalent conversion format flag as specified in the underlying
36 `EC_KEY` object being exported to a provider, when this function is
37 called through `EVP_PKEY_export()`.
41 * RNDR and RNDRRS support in provider functions to provide
42 random number generation for Arm CPUs (aarch64).
46 * s_client and s_server apps now explicitly say when the TLS version
47 does not include the renegotiation mechanism. This avoids confusion
48 between that scenario versus when the TLS version includes secure
49 renegotiation but the peer lacks support for it.
53 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
55 *Tomasz Kantecki, Andrey Matyukov*
57 * The various OBJ_* functions have been made thread safe.
61 * Parallel dual-prime 1536/2048-bit modular exponentiation for
62 AVX512_IFMA capable processors.
64 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
66 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
67 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
68 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
69 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
70 `OPENSSL_NO_DEPRECATED_3_1`.
72 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
73 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
74 definitions for these functions regardless of whether
75 `OPENSSL_NO_DEPRECATED_3_1` is defined.
77 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
78 functions regardless of whether they are using them. It is recommended that
79 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
83 * When generating safe-prime DH parameters set the recommended private key
84 length equivalent to minimum key lengths as in RFC 7919.
91 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
92 listed here are only a brief description.
93 The migration guide contains more detailed information related to new features,
94 breaking changes, and mappings for the large list of deprecated functions.
96 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
98 ### Changes between 3.0.6 and 3.0.7 [xx XXX xxxx]
100 * Fixed two buffer overflows in punycode decoding functions.
102 A buffer overrun can be triggered in X.509 certificate verification,
103 specifically in name constraint checking. Note that this occurs after
104 certificate chain signature verification and requires either a CA to
105 have signed the malicious certificate or for the application to continue
106 certificate verification despite failure to construct a path to a trusted
109 In a TLS client, this can be triggered by connecting to a malicious
110 server. In a TLS server, this can be triggered if the server requests
111 client authentication and a malicious client connects.
113 An attacker can craft a malicious email address to overflow
114 an arbitrary number of bytes containing the `.` character (decimal 46)
115 on the stack. This buffer overflow could result in a crash (causing a
119 An attacker can craft a malicious email address to overflow four
120 attacker-controlled bytes on the stack. This buffer overflow could
121 result in a crash (causing a denial of service) or potentially remote code
122 execution depending on stack layout for any given platform/compiler.
127 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
128 parameters in OpenSSL code.
129 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
130 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
131 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
132 Using these invalid names may cause algorithms to use slower methods
133 that ignore the CRT parameters.
137 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
142 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
143 data to be signed before signing the certificate.
147 * Added RIPEMD160 to the default provider.
151 * Ensured that the key share group sent or accepted for the key exchange
152 is allowed for the protocol version.
156 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
158 * OpenSSL supports creating a custom cipher via the legacy
159 EVP_CIPHER_meth_new() function and associated function calls. This function
160 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
161 to use the new provider mechanism in order to implement custom ciphers.
163 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
164 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
165 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
166 and decryption initialisation functions). Instead of using the custom cipher
167 directly it incorrectly tries to fetch an equivalent cipher from the
168 available providers. An equivalent cipher is found based on the NID passed to
169 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
170 given cipher. However it is possible for an application to incorrectly pass
171 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
172 is used in this way the OpenSSL encryption/decryption initialisation function
173 will match the NULL cipher as being equivalent and will fetch this from the
174 available providers. This will succeed if the default provider has been
175 loaded (or if a third party provider has been loaded that offers this
176 cipher). Using the NULL cipher means that the plaintext is emitted as the
179 Applications are only affected by this issue if they call
180 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
181 encryption/decryption initialisation function. Applications that only use
182 SSL/TLS are not impacted by this issue.
187 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
192 * Fixed the linux-mips64 Configure target which was missing the
193 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
198 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
203 * Correctly handle a retransmitted ClientHello in DTLS
207 * Fixed detection of ktls support in cross-compile environment on Linux
211 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
216 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
217 report correct results in some cases
221 * Fix UWP builds by defining VirtualLock
225 * For known safe primes use the minimum key length according to RFC 7919.
226 Longer private key sizes unnecessarily raise the cycles needed to compute the
227 shared secret without any increase of the real security. This fixes a
228 regression from 1.1.1 where these shorter keys were generated for the known
233 * Added the loongarch64 target
237 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
238 only passed to the FIPS provider and not to the default or legacy provider.
242 * Fixed reported performance degradation on aarch64. Restored the
243 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
244 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
245 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
246 The new algorithm is still used for 32 bit targets.
250 * Added a missing header for memcmp that caused compilation failure on some
255 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
257 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
258 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
259 This issue makes the RSA implementation with 2048 bit private keys
260 incorrect on such machines and memory corruption will happen during
261 the computation. As a consequence of the memory corruption an attacker
262 may be able to trigger a remote code execution on the machine performing
265 SSL/TLS servers or other servers using 2048 bit RSA private keys running
266 on machines supporting AVX512IFMA instructions of the X86_64 architecture
267 are affected by this issue.
272 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
273 implementation would not encrypt the entirety of the data under some
274 circumstances. This could reveal sixteen bytes of data that was
275 preexisting in the memory that wasn't written. In the special case of
276 "in place" encryption, sixteen bytes of the plaintext would be revealed.
278 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
279 they are both unaffected.
282 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
284 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
286 * In addition to the c_rehash shell command injection identified in
287 CVE-2022-1292, further bugs where the c_rehash script does not
288 properly sanitise shell metacharacters to prevent command injection have been
291 When the CVE-2022-1292 was fixed it was not discovered that there
292 are other places in the script where the file names of certificates
293 being hashed were possibly passed to a command executed through the shell.
295 This script is distributed by some operating systems in a manner where
296 it is automatically executed. On such operating systems, an attacker
297 could execute arbitrary commands with the privileges of the script.
299 Use of the c_rehash script is considered obsolete and should be replaced
300 by the OpenSSL rehash command line tool.
303 *Daniel Fiala, Tomáš Mráz*
305 * Case insensitive string comparison no longer uses locales. It has instead
306 been directly implemented.
310 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
312 * Case insensitive string comparison is reimplemented via new locale-agnostic
313 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
314 comparison. The previous implementation had problems when the Turkish locale
319 * Fixed a bug in the c_rehash script which was not properly sanitising shell
320 metacharacters to prevent command injection. This script is distributed by
321 some operating systems in a manner where it is automatically executed. On
322 such operating systems, an attacker could execute arbitrary commands with the
323 privileges of the script.
325 Use of the c_rehash script is considered obsolete and should be replaced
326 by the OpenSSL rehash command line tool.
331 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
332 certificate on an OCSP response. The bug caused the function in the case
333 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
334 response (meaning a successful verification) even in the case where the
335 response signing certificate fails to verify.
337 It is anticipated that most users of `OCSP_basic_verify` will not use the
338 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
339 a negative value (indicating a fatal error) in the case of a certificate
340 verification failure. The normal expected return value in this case would be
343 This issue also impacts the command line OpenSSL "ocsp" application. When
344 verifying an ocsp response with the "-no_cert_checks" option the command line
345 application will report that the verification is successful even though it
346 has in fact failed. In this case the incorrect successful response will also
347 be accompanied by error messages showing the failure and contradicting the
348 apparently successful result.
353 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
354 AAD data as the MAC key. This made the MAC key trivially predictable.
356 An attacker could exploit this issue by performing a man-in-the-middle attack
357 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
358 that the modified data would still pass the MAC integrity check.
360 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
361 endpoint will always be rejected by the recipient and the connection will
362 fail at that point. Many application protocols require data to be sent from
363 the client to the server first. Therefore, in such a case, only an OpenSSL
364 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
366 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
367 sent in both directions. In this case both clients and servers could be
368 affected, regardless of the application protocol.
370 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
371 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
372 the handshake when using this ciphersuite.
374 The confidentiality of data is not impacted by this issue, i.e. an attacker
375 cannot decrypt data that has been encrypted using this ciphersuite - they can
378 In order for this attack to work both endpoints must legitimately negotiate
379 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
380 OpenSSL 3.0, and is not available within the default provider or the default
381 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
382 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
383 following must have occurred:
385 1) OpenSSL must have been compiled with the (non-default) compile time option
386 enable-weak-ssl-ciphers
388 2) OpenSSL must have had the legacy provider explicitly loaded (either
389 through application code or via configuration)
391 3) The ciphersuite must have been explicitly added to the ciphersuite list
393 4) The libssl security level must have been set to 0 (default is 1)
395 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
397 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
398 others that both endpoints have in common
403 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
404 occupied by the removed hash table entries.
406 This function is used when decoding certificates or keys. If a long lived
407 process periodically decodes certificates or keys its memory usage will
408 expand without bounds and the process might be terminated by the operating
409 system causing a denial of service. Also traversing the empty hash table
410 entries will take increasingly more time.
412 Typically such long lived processes might be TLS clients or TLS servers
413 configured to accept client certificate authentication.
416 *Hugo Landau, Aliaksei Levin*
418 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
419 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
420 statistics are no longer supported. For compatibility, these statistics are
421 still listed in the output but are now always reported as zero.
425 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
427 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
428 for non-prime moduli.
430 Internally this function is used when parsing certificates that contain
431 elliptic curve public keys in compressed form or explicit elliptic curve
432 parameters with a base point encoded in compressed form.
434 It is possible to trigger the infinite loop by crafting a certificate that
435 has invalid explicit curve parameters.
437 Since certificate parsing happens prior to verification of the certificate
438 signature, any process that parses an externally supplied certificate may thus
439 be subject to a denial of service attack. The infinite loop can also be
440 reached when parsing crafted private keys as they can contain explicit
441 elliptic curve parameters.
443 Thus vulnerable situations include:
445 - TLS clients consuming server certificates
446 - TLS servers consuming client certificates
447 - Hosting providers taking certificates or private keys from customers
448 - Certificate authorities parsing certification requests from subscribers
449 - Anything else which parses ASN.1 elliptic curve parameters
451 Also any other applications that use the BN_mod_sqrt() where the attacker
452 can control the parameter values are vulnerable to this DoS issue.
457 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
458 to the list of ciphersuites providing Perfect Forward Secrecy as
459 required by SECLEVEL >= 3.
461 *Dmitry Belyavskiy, Nicola Tuveri*
463 * Made the AES constant time code for no-asm configurations
464 optional due to the resulting 95% performance degradation.
465 The AES constant time code can be enabled, for no assembly
466 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
470 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
475 * The negative return value handling of the certificate verification callback
476 was reverted. The replacement is to set the verification retry state with
477 the SSL_set_retry_verify() function.
481 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
483 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
484 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
485 verify a certificate supplied by a server. That function may return a
486 negative return value to indicate an internal error (for example out of
487 memory). Such a negative return value is mishandled by OpenSSL and will cause
488 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
489 success and a subsequent call to SSL_get_error() to return the value
490 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
491 returned by OpenSSL if the application has previously called
492 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
493 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
494 totally unexpected and applications may not behave correctly as a result. The
495 exact behaviour will depend on the application but it could result in
496 crashes, infinite loops or other similar incorrect responses.
498 This issue is made more serious in combination with a separate bug in OpenSSL
499 3.0 that will cause X509_verify_cert() to indicate an internal error when
500 processing a certificate chain. This will occur where a certificate does not
501 include the Subject Alternative Name extension but where a Certificate
502 Authority has enforced name constraints. This issue can occur even with valid
508 * Corrected a few file name and file reference bugs in the build,
509 installation and setup scripts, which lead to installation verification
510 failures. Slightly enhanced the installation verification script.
514 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
519 * Fixed PVK encoder to properly query for the passphrase.
523 * Multiple fixes in the OSSL_HTTP API functions.
527 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
528 OSSL_PARAM_INTEGER data type and return error on negative numbers
529 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
530 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
534 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
538 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
542 * Multiple threading fixes.
546 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
550 * Allow fetching an operation from the provider that owns an unexportable key
551 as a fallback if that is still allowed by the property query.
555 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
557 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
562 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
563 S390X capability vector to zero. This simplifies testing of different code
564 paths on S390X architecture.
568 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
569 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
570 SP 800-38D". The communication will fail at this point.
574 * The EC_GROUP_clear_free() function is deprecated as there is nothing
575 confidential in EC_GROUP data.
579 * The byte order mark (BOM) character is ignored if encountered at the
580 beginning of a PEM-formatted file.
584 * Added CMS support for the Russian GOST algorithms.
588 * Due to move of the implementation of cryptographic operations
589 to the providers, validation of various operation parameters can
590 be postponed until the actual operation is executed where previously
591 it happened immediately when an operation parameter was set.
593 For example when setting an unsupported curve with
594 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
595 fail but later keygen operations with the EVP_PKEY_CTX will fail.
597 *OpenSSL team members and many third party contributors*
599 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
600 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
601 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
602 instead to retrieve these algorithms from a provider.
606 * On build targets where the multilib postfix is set in the build
607 configuration the libdir directory was changing based on whether
608 the lib directory with the multilib postfix exists on the system
609 or not. This unpredictable behavior was removed and eventual
610 multilib postfix is now always added to the default libdir. Use
611 `--libdir=lib` to override the libdir if adding the postfix is
616 * The triple DES key wrap functionality now conforms to RFC 3217 but is
617 no longer interoperable with OpenSSL 1.1.1.
621 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
622 function codes, this function can only cause problems for calling
627 * Add a configurable flag to output date formats as ISO 8601. Does not
628 change the default date format.
632 * Version of MSVC earlier than 1300 could get link warnings, which could
633 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
634 Support for this flag has been removed.
638 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
639 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
640 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
641 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
642 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
646 * The signatures of the functions to get and set options on SSL and
647 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
648 Some source code changes may be required.
652 * The public definitions of conf_method_st and conf_st have been
653 deprecated. They will be made opaque in a future release.
655 *Rich Salz and Tomáš Mráz*
657 * Client-initiated renegotiation is disabled by default. To allow it, use
658 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
659 flag, or the "ClientRenegotiation" config parameter as appropriate.
663 * Add "abspath" and "includedir" pragma's to config files, to prevent,
664 or modify relative pathname inclusion.
668 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
669 validated. Please consult the README-FIPS and
670 README-PROVIDERS files, as well as the migration guide.
672 *OpenSSL team members and many third party contributors*
674 * For the key types DH and DHX the allowed settable parameters are now different.
678 * The openssl commands that read keys, certificates, and CRLs now
679 automatically detect the PEM or DER format of the input files.
681 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
683 * Added enhanced PKCS#12 APIs which accept a library context.
687 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
691 * Added support for Kernel TLS (KTLS).
693 *Boris Pismenny, John Baldwin and Andrew Gallatin*
695 * Support for RFC 5746 secure renegotiation is now required by default for
696 SSL or TLS connections to succeed.
700 * The signature of the `copy` functional parameter of the
701 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
702 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
703 the signature of the `pub_decode` functional parameter of the
704 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
705 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
709 * The error return values from some control calls (ctrl) have changed.
713 * A public key check is now performed during EVP_PKEY_derive_set_peer().
717 * Many functions in the EVP_ namespace that are getters of values from
718 implementations or contexts were renamed to include get or get0 in their
719 names. Old names are provided as macro aliases for compatibility and
724 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
725 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
726 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
731 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
734 * The output from the command line applications may have minor
739 * The output from numerous "printing" may have minor changes.
743 * Windows thread synchronization uses read/write primitives (SRWLock) when
744 supported by the OS, otherwise CriticalSection continues to be used.
748 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
749 work on read only BIO source/sinks that do not support these functions.
750 This allows piping or redirection of a file BIO using stdin to be buffered
751 into memory. This is used internally in OSSL_DECODER_from_bio().
755 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
756 this function would return one of the values OSSL_STORE_INFO_NAME,
757 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
758 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
759 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
760 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
761 using this function should be amended to handle the changed return value.
765 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
766 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
767 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
768 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
769 contains more than one certificate identifier: This means that all
770 certificates referenced there MUST be part of the validation chain.
774 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
775 RC5, DESX and DES have been moved to the legacy provider.
779 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
780 RIPEMD-160 have been moved to the legacy provider.
784 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
789 * The deprecated functions EVP_PKEY_get0_RSA(),
790 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
791 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
792 well as the similarly named "get1" functions behave differently in
797 * A number of functions handling low-level keys or engines were deprecated
798 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
799 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
800 EVP_PKEY_get0_siphash().
804 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
805 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
806 will need to load the legacy crypto provider. This includes these PBE
807 algorithms which use this KDF:
808 - NID_pbeWithMD2AndDES_CBC
809 - NID_pbeWithMD5AndDES_CBC
810 - NID_pbeWithSHA1AndRC2_CBC
811 - NID_pbeWithMD2AndRC2_CBC
812 - NID_pbeWithMD5AndRC2_CBC
813 - NID_pbeWithSHA1AndDES_CBC
817 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
818 BIO_debug_callback() functions.
822 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
823 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
827 * The RAND_METHOD APIs have been deprecated.
831 * The SRP APIs have been deprecated.
835 * Add a compile time option to prevent the caching of provider fetched
836 algorithms. This is enabled by including the no-cached-fetch option
837 at configuration time.
841 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
842 count of PKCS12_DEFAULT_ITER.
844 *Tomáš Mráz and Sahana Prasad*
846 * The openssl speed command does not use low-level API calls anymore.
850 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
853 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
855 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
859 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
860 providers may supply their own group implementations (using either the "key
861 exchange" or the "key encapsulation" methods) which will automatically be
862 detected and used by libssl.
864 *Matt Caswell, Nicola Tuveri*
866 * The undocumented function X509_certificate_type() has been deprecated;
870 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
874 * Removed RSA padding mode for SSLv23 (which was only used for
875 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
876 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
881 * Deprecated the obsolete X9.31 RSA key generation related functions.
883 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
884 is not allowed to return a value > 1, this is no more taken as failure.
886 *Viktor Dukhovni and David von Oheimb*
888 * Deprecated the obsolete X9.31 RSA key generation related functions
889 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
890 BN_X931_generate_prime_ex().
894 * The default key generation method for the regular 2-prime RSA keys was
895 changed to the FIPS 186-4 B.3.6 method.
899 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
903 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
907 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
908 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
910 *Rich Salz, Richard Levitte, and David von Oheimb*
912 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
916 * Deprecated `OCSP_parse_url()`.
920 * Validation of SM2 keys has been separated from the validation of regular EC
925 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
926 switches: a validation failure triggers an early exit, returning a failure
927 exit status to the parent process.
931 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
932 to ignore unknown ciphers.
936 * The `-cipher-commands` and `-digest-commands` options
937 of the command line utility `list` have been deprecated.
938 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
942 * Added convenience functions for generating asymmetric key pairs:
943 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
944 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
948 * All of the low-level EC_KEY functions have been deprecated.
950 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
952 * Deprecated all the libcrypto and libssl error string loading
957 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
958 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
963 * The `-crypt` option to the `passwd` command line tool has been removed.
967 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
972 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
976 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
977 EVP_PKEY_get1_tls_encodedpoint().
981 * The security callback, which can be customised by application code, supports
982 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
983 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
987 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
988 interface. Their functionality remains unchanged.
992 * Added new option for 'openssl list', '-providers', which will display the
993 list of loaded providers, their names, version and status. It optionally
994 displays their gettable parameters.
998 * Removed EVP_PKEY_set_alias_type().
1002 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1003 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1007 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1008 parameters are still checked for type safety at compile time via helper
1013 * Remove the RAND_DRBG API
1015 *Paul Dale and Matthias St. Pierre*
1017 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1018 as well as actual hostnames.
1022 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1023 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1024 conversely, silently ignore DTLS protocol version bounds when configuring
1025 TLS-based contexts. The commands can be repeated to set bounds of both
1026 types. The same applies with the corresponding "min_protocol" and
1027 "max_protocol" command-line switches, in case some application uses both TLS
1030 SSL_CTX instances that are created for a fixed protocol version (e.g.
1031 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1032 attempts to apply bounds to these protocol versions would result in an
1033 error. Now only the "version-flexible" SSL_CTX instances are subject to
1034 limits in configuration files in command-line options.
1038 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1043 * Reworked the recorded ERR codes to make better space for system errors.
1044 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1045 given code is a system error (true) or an OpenSSL error (false).
1049 * Reworked the test perl framework to better allow parallel testing.
1051 *Nicola Tuveri and David von Oheimb*
1053 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1054 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1058 * 'Configure' has been changed to figure out the configuration target if
1059 none is given on the command line. Consequently, the 'config' script is
1060 now only a mere wrapper. All documentation is changed to only mention
1063 *Rich Salz and Richard Levitte*
1065 * Added a library context `OSSL_LIB_CTX` that applications as well as
1066 other libraries can use to form a separate context within which
1067 libcrypto operations are performed.
1071 * Added various `_ex` functions to the OpenSSL API that support using
1072 a non-default `OSSL_LIB_CTX`.
1076 * Handshake now fails if Extended Master Secret extension is dropped
1081 * Dropped interactive mode from the `openssl` program.
1085 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1087 *David von Oheimb and Shane Lontis*
1089 * Deprecated `EC_METHOD_get_field_type()`.
1093 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1094 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1095 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1099 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1103 * Add CAdES-BES signature verification support, mostly derived
1104 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1106 *Filipe Raimundo da Silva*
1108 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1112 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1113 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1117 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1121 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1122 EC_KEY_precompute_mult().
1126 * Deprecated EC_POINTs_mul().
1130 * Removed FIPS_mode() and FIPS_mode_set().
1134 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1138 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1139 EC_POINT_get_Jprojective_coordinates_GFp().
1143 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1144 arrays to be more easily constructed via a series of utility functions.
1145 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1146 the various push functions and finally convert to a passable OSSL_PARAM
1147 array using OSSL_PARAM_BLD_to_param().
1151 * The security strength of SHA1 and MD5 based signatures in TLS has been
1156 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1157 contain a provider side internal key.
1161 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1165 * Project text documents not yet having a proper file name extension
1166 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1167 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1171 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1172 have been converted to Markdown with the goal to produce documents
1173 which not only look pretty when viewed online in the browser, but
1174 remain well readable inside a plain text editor.
1176 To achieve this goal, a 'minimalistic' Markdown style has been applied
1177 which avoids formatting elements that interfere too much with the
1178 reading flow in the text file. For example, it
1180 * avoids [ATX headings][] and uses [setext headings][] instead
1181 (which works for `<h1>` and `<h2>` headings only).
1182 * avoids [inline links][] and uses [reference links][] instead.
1183 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1185 [ATX headings]: https://github.github.com/gfm/#atx-headings
1186 [setext headings]: https://github.github.com/gfm/#setext-headings
1187 [inline links]: https://github.github.com/gfm/#inline-link
1188 [reference links]: https://github.github.com/gfm/#reference-link
1189 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1190 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1192 *Matthias St. Pierre*
1194 * The test suite is changed to preserve results of each test recipe.
1195 A new directory test-runs/ with subdirectories named like the
1196 test recipes are created in the build tree for this purpose.
1200 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1201 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1202 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1204 *David von Oheimb, Martin Peylo*
1206 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1207 It supports arbitrary request and response content types, GET redirection,
1208 TLS, connections via HTTP(S) proxies, connections and exchange via
1209 user-defined BIOs (allowing implicit connections), persistent connections,
1210 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1211 The legacy OCSP-focused (and only partly documented) API
1212 is retained for backward compatibility, while most of it is deprecated.
1216 * Added `util/check-format.pl`, a tool for checking adherence to the
1217 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1218 The checks performed are incomplete and yield some false positives.
1219 Still the tool should be useful for detecting most typical glitches.
1223 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1224 If domain name resolution yields multiple IP addresses all of them are tried
1225 after `connect()` failures.
1229 * All of the low-level RSA functions have been deprecated.
1233 * X509 certificates signed using SHA1 are no longer allowed at security
1238 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1239 modified to use PKEY APIs. These commands are now in maintenance mode
1240 and no new features will be added to them.
1244 * The command line utility rsautl has been deprecated.
1248 * The command line utilities genrsa and rsa have been modified to use PKEY
1249 APIs. They now write PKCS#8 keys by default. These commands are now in
1250 maintenance mode and no new features will be added to them.
1254 * All of the low-level DH functions have been deprecated.
1256 *Paul Dale and Matt Caswell*
1258 * All of the low-level DSA functions have been deprecated.
1262 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1263 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1267 * Deprecated low-level ECDH and ECDSA functions.
1271 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1275 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1276 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1277 a new formulation to include all the things it can be used for,
1278 as well as words of caution.
1282 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1286 * All of the low-level HMAC functions have been deprecated.
1288 *Paul Dale and David von Oheimb*
1290 * Over two thousand fixes were made to the documentation, including:
1291 - Common options (such as -rand/-writerand, TLS version control, etc)
1292 were refactored and point to newly-enhanced descriptions in openssl.pod.
1293 - Added style conformance for all options (with help from Richard Levitte),
1294 documented all reported missing options, added a CI build to check
1295 that all options are documented and that no unimplemented options
1297 - Documented some internals, such as all use of environment variables.
1298 - Addressed all internal broken L<> references.
1302 * All of the low-level CMAC functions have been deprecated.
1306 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1307 functions have been deprecated.
1309 *Paul Dale and David von Oheimb*
1311 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1312 set of functions. The documentation mentioned negative values for some
1313 errors, but this was never the case, so the mention of negative values
1316 Code that followed the documentation and thereby check with something
1317 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1321 * All of the low-level cipher functions have been deprecated.
1323 *Matt Caswell and Paul Dale*
1325 * Removed include/openssl/opensslconf.h.in and replaced it with
1326 include/openssl/configuration.h.in, which differs in not including
1327 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1328 was added to include both.
1330 This allows internal hacks where one might need to modify the set
1331 of configured macros, for example this if deprecated symbols are
1332 still supposed to be available internally:
1334 #include <openssl/configuration.h>
1336 #undef OPENSSL_NO_DEPRECATED
1337 #define OPENSSL_SUPPRESS_DEPRECATED
1339 #include <openssl/macros.h>
1341 This should not be used by applications that use the exported
1342 symbols, as that will lead to linking errors.
1346 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1347 used in exponentiation with 512-bit moduli. No EC algorithms are
1348 affected. Analysis suggests that attacks against 2-prime RSA1024,
1349 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1350 difficult to perform and are not believed likely. Attacks against DH512
1351 are considered just feasible. However, for an attack the target would
1352 have to re-use the DH512 private key, which is not recommended anyway.
1353 Also applications directly using the low-level API BN_mod_exp may be
1354 affected if they use BN_FLG_CONSTTIME.
1359 * Most memory-debug features have been deprecated, and the functionality
1360 replaced with no-ops.
1364 * Added documentation for the STACK API.
1368 * Introduced a new method type and API, OSSL_ENCODER, to represent
1369 generic encoders. These do the same sort of job that PEM writers
1370 and d2i functions do, but with support for methods supplied by
1371 providers, and the possibility for providers to support other
1376 * Introduced a new method type and API, OSSL_DECODER, to represent
1377 generic decoders. These do the same sort of job that PEM readers
1378 and i2d functions do, but with support for methods supplied by
1379 providers, and the possibility for providers to support other
1384 * Added a .pragma directive to the syntax of configuration files, to
1385 allow varying behavior in a supported and predictable manner.
1386 Currently added pragma:
1390 This allows dollar signs to be a keyword character unless it's
1391 followed by a opening brace or parenthesis. This is useful for
1392 platforms where dollar signs are commonly used in names, such as
1393 volume names and system directory names on VMS.
1397 * Added functionality to create an EVP_PKEY from user data.
1401 * Change the interpretation of the '--api' configuration option to
1402 mean that this is a desired API compatibility level with no
1403 further meaning. The previous interpretation, that this would
1404 also mean to remove all deprecated symbols up to and including
1405 the given version, no requires that 'no-deprecated' is also used
1406 in the configuration.
1408 When building applications, the desired API compatibility level
1409 can be set with the OPENSSL_API_COMPAT macro like before. For
1410 API compatibility version below 3.0, the old style numerical
1411 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1412 For version 3.0 and on, the value is expected to be the decimal
1413 value calculated from the major and minor version like this:
1415 MAJOR * 10000 + MINOR * 100
1419 -DOPENSSL_API_COMPAT=30000 For 3.0
1420 -DOPENSSL_API_COMPAT=30200 For 3.2
1422 To hide declarations that are deprecated up to and including the
1423 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1424 given when building the application as well.
1428 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1429 access to certificate and CRL stores via URIs and OSSL_STORE
1432 This adds the following functions:
1434 - X509_LOOKUP_store()
1435 - X509_STORE_load_file()
1436 - X509_STORE_load_path()
1437 - X509_STORE_load_store()
1438 - SSL_add_store_cert_subjects_to_stack()
1439 - SSL_CTX_set_default_verify_store()
1440 - SSL_CTX_load_verify_file()
1441 - SSL_CTX_load_verify_dir()
1442 - SSL_CTX_load_verify_store()
1446 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1447 The presence of this system service is determined at run-time.
1451 * Added functionality to create an EVP_PKEY context based on data
1452 for methods from providers. This takes an algorithm name and a
1453 property query string and simply stores them, with the intent
1454 that any operation that uses this context will use those strings
1455 to fetch the needed methods implicitly, thereby making the port
1456 of application written for pre-3.0 OpenSSL easier.
1460 * The undocumented function NCONF_WIN32() has been deprecated; for
1461 conversion details see the HISTORY section of doc/man5/config.pod
1465 * Introduced the new functions EVP_DigestSignInit_ex() and
1466 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1467 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1468 pages for further details.
1472 * Over two thousand fixes were made to the documentation, including:
1473 adding missing command flags, better style conformance, documentation
1476 *Rich Salz, Richard Levitte*
1478 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1479 X25519, X448, Ed25519 and Ed448.
1483 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1488 * Deprecated the public definition of `ERR_STATE` as well as the function
1489 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1494 * Added ERR functionality to give callers access to the stored function
1495 names that have replaced the older function code based functions.
1497 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1498 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1499 ERR_peek_error_all() and ERR_peek_last_error_all().
1501 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1502 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1503 ERR_func_error_string().
1507 * Extended testing to be verbose for failing tests only. The make variables
1508 VERBOSE_FAILURE or VF can be used to enable this:
1510 $ make VF=1 test # Unix
1511 $ mms /macro=(VF=1) test ! OpenVMS
1512 $ nmake VF=1 test # Windows
1516 * Added the `-copy_extensions` option to the `x509` command for use with
1517 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1518 all extensions in the request are copied to the certificate or vice versa.
1520 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1522 * Added the `-copy_extensions` option to the `req` command for use with
1523 `-x509`. When given with the `copy` or `copyall` argument,
1524 all extensions in the certification request are copied to the certificate.
1528 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1529 they generate are by default RFC 5280 compliant in the following sense:
1530 There is a subjectKeyIdentifier extension with a hash value of the public key
1531 and for not self-signed certs there is an authorityKeyIdentifier extension
1532 with a keyIdentifier field or issuer information identifying the signing key.
1533 This is done unless some configuration overrides the new default behavior,
1534 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1538 * Added several checks to `X509_verify_cert()` according to requirements in
1539 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1540 (which may be done by using the CLI option `-x509_strict`):
1541 * The basicConstraints of CA certificates must be marked critical.
1542 * CA certificates must explicitly include the keyUsage extension.
1543 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1544 * The issuer name of any certificate must not be empty.
1545 * The subject name of CA certs, certs with keyUsage crlSign,
1546 and certs without subjectAlternativeName must not be empty.
1547 * If a subjectAlternativeName extension is given it must not be empty.
1548 * The signatureAlgorithm field and the cert signature must be consistent.
1549 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1550 must not be marked critical.
1551 * The authorityKeyIdentifier must be given for X.509v3 certs
1552 unless they are self-signed.
1553 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1557 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1558 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1562 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1563 used even when parsing explicit parameters, when loading a encoded key
1564 or calling `EC_GROUP_new_from_ecpkparameters()`/
1565 `EC_GROUP_new_from_ecparameters()`.
1566 This prevents bypass of security hardening and performance gains,
1567 especially for curves with specialized EC_METHODs.
1568 By default, if a key encoded with explicit parameters is loaded and later
1569 encoded, the output is still encoded with explicit parameters, even if
1570 internally a "named" EC_GROUP is used for computation.
1574 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1575 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1576 NULL. After this change, only the cofactor parameter can be NULL. It also
1577 does some minimal sanity checks on the passed order.
1582 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1583 An attack is simple, if the first CMS_recipientInfo is valid but the
1584 second CMS_recipientInfo is chosen ciphertext. If the second
1585 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1586 encryption key will be replaced by garbage, and the message cannot be
1587 decoded, but if the RSA decryption fails, the correct encryption key is
1588 used and the recipient will not notice the attack.
1589 As a work around for this potential attack the length of the decrypted
1590 key must be equal to the cipher default key length, in case the
1591 certificate is not given and all recipientInfo are tried out.
1592 The old behaviour can be re-enabled in the CMS code by setting the
1593 CMS_DEBUG_DECRYPT flag.
1597 * Early start up entropy quality from the DEVRANDOM seed source has been
1598 improved for older Linux systems. The RAND subsystem will wait for
1599 /dev/random to be producing output before seeding from /dev/urandom.
1600 The seeded state is stored for future library initialisations using
1601 a system global shared memory segment. The shared memory identifier
1602 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1603 the desired value. The default identifier is 114.
1607 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1608 when primes for RSA keys are computed.
1609 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1610 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1611 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1612 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1613 This avoids possible fingerprinting of newly generated RSA modules.
1617 * Correct the extended master secret constant on EBCDIC systems. Without this
1618 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1619 negotiate EMS will fail. Unfortunately this also means that TLS connections
1620 between EBCDIC systems with this fix, and EBCDIC systems without this
1621 fix will fail if they negotiate EMS.
1625 * Changed the library initialisation so that the config file is now loaded
1626 by default. This was already the case for libssl. It now occurs for both
1627 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1628 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1632 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1633 where the former acts as a replacement for `ERR_put_error()`, and the
1634 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1635 `ERR_raise_data()` adds more flexibility by taking a format string and
1636 an arbitrary number of arguments following it, to be processed with
1641 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1642 to check if a named provider is loaded and available. When called, it
1643 will also activate all fallback providers if such are still present.
1647 * Enforce a minimum DH modulus size of 512 bits.
1651 * Changed DH parameters to generate the order q subgroup instead of 2q.
1652 Previously generated DH parameters are still accepted by DH_check
1653 but DH_generate_key works around that by clearing bit 0 of the
1654 private key for those. This avoids leaking bit 0 of the private key.
1658 * Significantly reduce secure memory usage by the randomness pools.
1662 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1667 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1668 algorithms. An implementation of a key exchange algorithm can be obtained
1669 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1670 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1671 the older EVP_PKEY_derive_init() function. See the man pages for the new
1672 functions for further details.
1676 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1680 * Removed the function names from error messages and deprecated the
1685 * Removed NextStep support and the macro OPENSSL_UNISTD
1689 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1690 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1691 Also removed "export var as function" capability; we do not export
1692 variables, only functions.
1696 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1697 an error and 1 indicating success. In previous versions of OpenSSL this
1698 was a void type. If a key was set longer than the maximum possible this
1703 * Support SM2 signing and verification schemes with X509 certificate.
1707 * Use SHA256 as the default digest for TS query in the `ts` app.
1711 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1715 * Default cipher lists/suites are now available via a function, the
1716 #defines are deprecated.
1720 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1721 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1722 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1726 * Join the directories crypto/x509 and crypto/x509v3
1730 * Added command 'openssl kdf' that uses the EVP_KDF API.
1734 * Added command 'openssl mac' that uses the EVP_MAC API.
1738 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1739 as default directories. Also added the command 'openssl info'
1740 for scripting purposes.
1744 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1749 * Add prediction resistance to the DRBG reseeding process.
1753 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1754 mandated by IEEE Std 1619-2018.
1758 * Added newline escaping functionality to a filename when using openssl dgst.
1759 This output format is to replicate the output format found in the `*sum`
1760 checksum programs. This aims to preserve backward compatibility.
1762 *Matt Eaton, Richard Levitte, and Paul Dale*
1764 * Removed the heartbeat message in DTLS feature, as it has very
1765 little usage and doesn't seem to fulfill a valuable purpose.
1766 The configuration option is now deprecated.
1770 * Changed the output of 'openssl {digestname} < file' to display the
1771 digest name in its output.
1775 * Added a new generic trace API which provides support for enabling
1776 instrumentation through trace output.
1778 *Richard Levitte & Matthias St. Pierre*
1780 * Added build tests for C++. These are generated files that only do one
1781 thing, to include one public OpenSSL head file each. This tests that
1782 the public header files can be usefully included in a C++ application.
1784 This test isn't enabled by default. It can be enabled with the option
1785 'enable-buildtest-c++'.
1789 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1793 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1797 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1801 * Added KMAC to EVP_MAC.
1805 * Added property based algorithm implementation selection framework to
1810 * Added SCA hardening for modular field inversion in EC_GROUP through
1811 a new dedicated field_inv() pointer in EC_METHOD.
1812 This also addresses a leakage affecting conversions from projective
1813 to affine coordinates.
1815 *Billy Bob Brumley, Nicola Tuveri*
1817 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1818 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1819 those algorithms that were already supported through the EVP_PKEY API
1820 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1821 and scrypt are now wrappers that call EVP_KDF.
1825 * Build devcrypto engine as a dynamic engine.
1827 *Eneas U de Queiroz*
1829 * Add keyed BLAKE2 to EVP_MAC.
1833 * Fix a bug in the computation of the endpoint-pair shared secret used
1834 by DTLS over SCTP. This breaks interoperability with older versions
1835 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1836 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1837 interoperability with such broken implementations. However, enabling
1838 this switch breaks interoperability with correct implementations.
1840 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1841 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1845 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1849 * Changed the license to the Apache License v2.0.
1853 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1855 - Major releases (indicated by incrementing the MAJOR release number)
1856 may introduce incompatible API/ABI changes.
1857 - Minor releases (indicated by incrementing the MINOR release number)
1858 may introduce new features but retain API/ABI compatibility.
1859 - Patch releases (indicated by incrementing the PATCH number)
1860 are intended for bug fixes and other improvements of existing
1861 features only (like improving performance or adding documentation)
1862 and retain API/ABI compatibility.
1866 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1870 * Remove the 'dist' target and add a tarball building script. The
1871 'dist' target has fallen out of use, and it shouldn't be
1872 necessary to configure just to create a source distribution.
1876 * Recreate the OS390-Unix config target. It no longer relies on a
1877 special script like it did for OpenSSL pre-1.1.0.
1881 * Instead of having the source directories listed in Configure, add
1882 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1887 * Add GMAC to EVP_MAC.
1891 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1895 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1896 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1897 to facilitate the continued use of MACs through raw private keys in
1898 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1902 * Deprecate ECDH_KDF_X9_62().
1906 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1907 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1908 are retained for backwards compatibility.
1912 * AES-XTS mode now enforces that its two keys are different to mitigate
1913 the attacked described in "Efficient Instantiations of Tweakable
1914 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1915 Details of this attack can be obtained from:
1916 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1920 * Rename the object files, i.e. give them other names than in previous
1921 versions. Their names now include the name of the final product, as
1922 well as its type mnemonic (bin, lib, shlib).
1926 * Added new option for 'openssl list', '-objects', which will display the
1927 list of built in objects, i.e. OIDs with names.
1931 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1932 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1937 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1938 improves application performance by removing data copies and providing
1939 applications with zero-copy system calls such as sendfile and splice.
1943 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1947 * `PKCS12_parse` now maintains the order of the parsed certificates
1948 when outputting them via `*ca` (rather than reversing it).
1952 * Deprecated pthread fork support methods.
1956 * Added support for FFDHE key exchange in TLS 1.3.
1960 * Added a new concept for OpenSSL plugability: providers. This
1961 functionality is designed to replace the ENGINE API and ENGINE
1962 implementations, and to be much more dynamic, allowing provider
1963 authors to introduce new algorithms among other things, as long as
1964 there's an API that supports the algorithm type.
1966 With this concept comes a new core API for interaction between
1967 libcrypto and provider implementations. Public libcrypto functions
1968 that want to use providers do so through this core API.
1970 The main documentation for this core API is found in
1971 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1972 refer to other manuals describing the API specific for supported
1973 algorithm types (also called operations).
1980 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1982 * Avoid loading of a dynamic engine twice.
1986 * Prioritise DANE TLSA issuer certs over peer certs
1990 * Fixed random API for MacOS prior to 10.12
1992 These MacOS versions don't support the CommonCrypto APIs
1996 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1998 * Fixed an SM2 Decryption Buffer Overflow.
2000 In order to decrypt SM2 encrypted data an application is expected to
2001 call the API function EVP_PKEY_decrypt(). Typically an application will
2002 call this function twice. The first time, on entry, the "out" parameter
2003 can be NULL and, on exit, the "outlen" parameter is populated with the
2004 buffer size required to hold the decrypted plaintext. The application
2005 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2006 again, but this time passing a non-NULL value for the "out" parameter.
2008 A bug in the implementation of the SM2 decryption code means that the
2009 calculation of the buffer size required to hold the plaintext returned
2010 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2011 size required by the second call. This can lead to a buffer overflow
2012 when EVP_PKEY_decrypt() is called by the application a second time with
2013 a buffer that is too small.
2015 A malicious attacker who is able present SM2 content for decryption to
2016 an application could cause attacker chosen data to overflow the buffer
2017 by up to a maximum of 62 bytes altering the contents of other data held
2018 after the buffer, possibly changing application behaviour or causing
2019 the application to crash. The location of the buffer is application
2020 dependent but is typically heap allocated.
2025 * Fixed various read buffer overruns processing ASN.1 strings
2027 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2028 structure which contains a buffer holding the string data and a field
2029 holding the buffer length. This contrasts with normal C strings which
2030 are represented as a buffer for the string data which is terminated
2031 with a NUL (0) byte.
2033 Although not a strict requirement, ASN.1 strings that are parsed using
2034 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2035 well as any string whose value has been set with the ASN1_STRING_set()
2036 function will additionally NUL terminate the byte array in the
2037 ASN1_STRING structure.
2039 However, it is possible for applications to directly construct valid
2040 ASN1_STRING structures which do not NUL terminate the byte array by
2041 directly setting the "data" and "length" fields in the ASN1_STRING
2042 array. This can also happen by using the ASN1_STRING_set0() function.
2044 Numerous OpenSSL functions that print ASN.1 data have been found to
2045 assume that the ASN1_STRING byte array will be NUL terminated, even
2046 though this is not guaranteed for strings that have been directly
2047 constructed. Where an application requests an ASN.1 structure to be
2048 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2049 been directly constructed by the application without NUL terminating
2050 the "data" field, then a read buffer overrun can occur.
2052 The same thing can also occur during name constraints processing
2053 of certificates (for example if a certificate has been directly
2054 constructed by the application instead of loading it via the OpenSSL
2055 parsing functions, and the certificate contains non NUL terminated
2056 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2057 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2059 If a malicious actor can cause an application to directly construct an
2060 ASN1_STRING and then process it through one of the affected OpenSSL
2061 functions then this issue could be hit. This might result in a crash
2062 (causing a Denial of Service attack). It could also result in the
2063 disclosure of private memory contents (such as private keys, or
2064 sensitive plaintext).
2069 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2071 * Fixed a problem with verifying a certificate chain when using the
2072 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2073 the certificates present in a certificate chain. It is not set by default.
2075 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2076 the chain that have explicitly encoded elliptic curve parameters was added
2077 as an additional strict check.
2079 An error in the implementation of this check meant that the result of a
2080 previous check to confirm that certificates in the chain are valid CA
2081 certificates was overwritten. This effectively bypasses the check
2082 that non-CA certificates must not be able to issue other certificates.
2084 If a "purpose" has been configured then there is a subsequent opportunity
2085 for checks that the certificate is a valid CA. All of the named "purpose"
2086 values implemented in libcrypto perform this check. Therefore, where
2087 a purpose is set the certificate chain will still be rejected even when the
2088 strict flag has been used. A purpose is set by default in libssl client and
2089 server certificate verification routines, but it can be overridden or
2090 removed by an application.
2092 In order to be affected, an application must explicitly set the
2093 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2094 for the certificate verification or, in the case of TLS client or server
2095 applications, override the default purpose.
2100 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2101 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2102 renegotiation ClientHello omits the signature_algorithms extension (where it
2103 was present in the initial ClientHello), but includes a
2104 signature_algorithms_cert extension then a NULL pointer dereference will
2105 result, leading to a crash and a denial of service attack.
2107 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2108 (which is the default configuration). OpenSSL TLS clients are not impacted by
2112 *Peter Kästle and Samuel Sapalski*
2114 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2116 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2117 create a unique hash value based on the issuer and serial number data
2118 contained within an X509 certificate. However, it was failing to correctly
2119 handle any errors that may occur while parsing the issuer field (which might
2120 occur if the issuer field is maliciously constructed). This may subsequently
2121 result in a NULL pointer deref and a crash leading to a potential denial of
2127 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2128 padding mode to correctly check for rollback attacks. This is considered a
2129 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2134 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2135 functions. Previously they could overflow the output length argument in some
2136 cases where the input length is close to the maximum permissible length for
2137 an integer on the platform. In such cases the return value from the function
2138 call would be 1 (indicating success), but the output length value would be
2139 negative. This could cause applications to behave incorrectly or crash.
2144 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2145 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2146 could be exploited in a side channel attack to recover the password. Since
2147 the attack is local host only this is outside of the current OpenSSL
2148 threat model and therefore no CVE is assigned.
2150 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2155 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2157 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2158 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2159 If an attacker can control both items being compared then this could lead
2160 to a possible denial of service attack. OpenSSL itself uses the
2161 GENERAL_NAME_cmp function for two purposes:
2162 1) Comparing CRL distribution point names between an available CRL and a
2163 CRL distribution point embedded in an X509 certificate
2164 2) When verifying that a timestamp response token signer matches the
2165 timestamp authority name (exposed via the API functions
2166 TS_RESP_verify_response and TS_RESP_verify_token)
2171 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2173 * Certificates with explicit curve parameters are now disallowed in
2174 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2178 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2179 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2180 conversely, silently ignore DTLS protocol version bounds when configuring
2181 TLS-based contexts. The commands can be repeated to set bounds of both
2182 types. The same applies with the corresponding "min_protocol" and
2183 "max_protocol" command-line switches, in case some application uses both TLS
2186 SSL_CTX instances that are created for a fixed protocol version (e.g.
2187 TLSv1_server_method()) also silently ignore version bounds. Previously
2188 attempts to apply bounds to these protocol versions would result in an
2189 error. Now only the "version-flexible" SSL_CTX instances are subject to
2190 limits in configuration files in command-line options.
2194 * Handshake now fails if Extended Master Secret extension is dropped
2199 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2201 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2203 * Fixed segmentation fault in SSL_check_chain()
2204 Server or client applications that call the SSL_check_chain() function
2205 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2206 dereference as a result of incorrect handling of the
2207 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2208 or unrecognised signature algorithm is received from the peer. This could
2209 be exploited by a malicious peer in a Denial of Service attack.
2214 * Added AES consttime code for no-asm configurations
2215 an optional constant time support for AES was added
2216 when building openssl for no-asm.
2217 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2218 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2219 At this time this feature is by default disabled.
2220 It will be enabled by default in 3.0.
2224 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2226 * Revert the change of EOF detection while reading in libssl to avoid
2227 regressions in applications depending on the current way of reporting
2228 the EOF. As the existing method is not fully accurate the change to
2229 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2230 branch and will be present in the 3.0 release.
2234 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2235 when primes for RSA keys are computed.
2236 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2237 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2238 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2239 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2240 This avoids possible fingerprinting of newly generated RSA modules.
2244 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2246 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2247 while reading in libssl then we would report an error back to the
2248 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2249 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2250 therefore give a hint as to what went wrong.
2254 * Check that ed25519 and ed448 are allowed by the security level. Previously
2255 signature algorithms not using an MD were not being checked that they were
2256 allowed by the security level.
2260 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2261 was not quite right. The behaviour was not consistent between resumption
2262 and normal handshakes, and also not quite consistent with historical
2263 behaviour. The behaviour in various scenarios has been clarified and
2264 it has been updated to make it match historical behaviour as closely as
2269 * *[VMS only]* The header files that the VMS compilers include automatically,
2270 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2271 that the C++ compiler doesn't understand. This is a shortcoming in the
2272 compiler, but can be worked around with `__cplusplus` guards.
2274 C++ applications that use OpenSSL libraries must be compiled using the
2275 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2276 functions. Otherwise, only functions with symbols of less than 31
2277 characters can be used, as the linker will not be able to successfully
2278 resolve symbols with longer names.
2282 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2283 The presence of this system service is determined at run-time.
2287 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2292 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2294 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2295 number generator (RNG). This was intended to include protection in the
2296 event of a fork() system call in order to ensure that the parent and child
2297 processes did not share the same RNG state. However, this protection was not
2298 being used in the default case.
2300 A partial mitigation for this issue is that the output from a high
2301 precision timer is mixed into the RNG state so the likelihood of a parent
2302 and child process sharing state is significantly reduced.
2304 If an application already calls OPENSSL_init_crypto() explicitly using
2305 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2308 *Matthias St. Pierre*
2310 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2311 used even when parsing explicit parameters, when loading a encoded key
2312 or calling `EC_GROUP_new_from_ecpkparameters()`/
2313 `EC_GROUP_new_from_ecparameters()`.
2314 This prevents bypass of security hardening and performance gains,
2315 especially for curves with specialized EC_METHODs.
2316 By default, if a key encoded with explicit parameters is loaded and later
2317 encoded, the output is still encoded with explicit parameters, even if
2318 internally a "named" EC_GROUP is used for computation.
2322 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2323 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2324 NULL. After this change, only the cofactor parameter can be NULL. It also
2325 does some minimal sanity checks on the passed order.
2330 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2331 An attack is simple, if the first CMS_recipientInfo is valid but the
2332 second CMS_recipientInfo is chosen ciphertext. If the second
2333 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2334 encryption key will be replaced by garbage, and the message cannot be
2335 decoded, but if the RSA decryption fails, the correct encryption key is
2336 used and the recipient will not notice the attack.
2337 As a work around for this potential attack the length of the decrypted
2338 key must be equal to the cipher default key length, in case the
2339 certificate is not given and all recipientInfo are tried out.
2340 The old behaviour can be re-enabled in the CMS code by setting the
2341 CMS_DEBUG_DECRYPT flag.
2346 * Early start up entropy quality from the DEVRANDOM seed source has been
2347 improved for older Linux systems. The RAND subsystem will wait for
2348 /dev/random to be producing output before seeding from /dev/urandom.
2349 The seeded state is stored for future library initialisations using
2350 a system global shared memory segment. The shared memory identifier
2351 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2352 the desired value. The default identifier is 114.
2356 * Correct the extended master secret constant on EBCDIC systems. Without this
2357 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2358 negotiate EMS will fail. Unfortunately this also means that TLS connections
2359 between EBCDIC systems with this fix, and EBCDIC systems without this
2360 fix will fail if they negotiate EMS.
2364 * Use Windows installation paths in the mingw builds
2366 Mingw isn't a POSIX environment per se, which means that Windows
2367 paths should be used for installation.
2372 * Changed DH_check to accept parameters with order q and 2q subgroups.
2373 With order 2q subgroups the bit 0 of the private key is not secret
2374 but DH_generate_key works around that by clearing bit 0 of the
2375 private key for those. This avoids leaking bit 0 of the private key.
2379 * Significantly reduce secure memory usage by the randomness pools.
2383 * Revert the DEVRANDOM_WAIT feature for Linux systems
2385 The DEVRANDOM_WAIT feature added a select() call to wait for the
2386 /dev/random device to become readable before reading from the
2387 /dev/urandom device.
2389 It turned out that this change had negative side effects on
2390 performance which were not acceptable. After some discussion it
2391 was decided to revert this feature and leave it up to the OS
2392 resp. the platform maintainer to ensure a proper initialization
2393 during early boot time.
2395 *Matthias St. Pierre*
2397 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2399 * Add build tests for C++. These are generated files that only do one
2400 thing, to include one public OpenSSL head file each. This tests that
2401 the public header files can be usefully included in a C++ application.
2403 This test isn't enabled by default. It can be enabled with the option
2404 'enable-buildtest-c++'.
2408 * Enable SHA3 pre-hashing for ECDSA and DSA.
2412 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2413 This changes the size when using the `genpkey` command when no size is given.
2414 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2415 generation commands to use 2048 bits by default.
2419 * Reorganize the manual pages to consistently have RETURN VALUES,
2420 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2421 util/fix-doc-nits accordingly.
2423 *Paul Yang, Joshua Lock*
2425 * Add the missing accessor EVP_PKEY_get0_engine()
2429 * Have commands like `s_client` and `s_server` output the signature scheme
2430 along with other cipher suite parameters when debugging.
2434 * Make OPENSSL_config() error agnostic again.
2438 * Do the error handling in RSA decryption constant time.
2442 * Prevent over long nonces in ChaCha20-Poly1305.
2444 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2445 for every encryption operation. RFC 7539 specifies that the nonce value
2446 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2447 and front pads the nonce with 0 bytes if it is less than 12
2448 bytes. However it also incorrectly allows a nonce to be set of up to 16
2449 bytes. In this case only the last 12 bytes are significant and any
2450 additional leading bytes are ignored.
2452 It is a requirement of using this cipher that nonce values are
2453 unique. Messages encrypted using a reused nonce value are susceptible to
2454 serious confidentiality and integrity attacks. If an application changes
2455 the default nonce length to be longer than 12 bytes and then makes a
2456 change to the leading bytes of the nonce expecting the new value to be a
2457 new unique nonce then such an application could inadvertently encrypt
2458 messages with a reused nonce.
2460 Additionally the ignored bytes in a long nonce are not covered by the
2461 integrity guarantee of this cipher. Any application that relies on the
2462 integrity of these ignored leading bytes of a long nonce may be further
2463 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2464 is safe because no such use sets such a long nonce value. However user
2465 applications that use this cipher directly and set a non-default nonce
2466 length to be longer than 12 bytes may be vulnerable.
2468 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2474 * Add DEVRANDOM_WAIT feature for Linux systems
2476 On older Linux systems where the getrandom() system call is not available,
2477 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2478 Contrary to getrandom(), the /dev/urandom device will not block during
2479 early boot when the kernel CSPRNG has not been seeded yet.
2481 To mitigate this known weakness, use select() to wait for /dev/random to
2482 become readable before reading from /dev/urandom.
2484 * Ensure that SM2 only uses SM3 as digest algorithm
2488 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2490 * Change the info callback signals for the start and end of a post-handshake
2491 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2492 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2493 confused by this and assume that a TLSv1.2 renegotiation has started. This
2494 can break KeyUpdate handling. Instead we no longer signal the start and end
2495 of a post handshake message exchange (although the messages themselves are
2496 still signalled). This could break some applications that were expecting
2497 the old signals. However without this KeyUpdate is not usable for many
2502 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2504 * Timing vulnerability in DSA signature generation
2506 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2507 timing side channel attack. An attacker could use variations in the signing
2508 algorithm to recover the private key.
2510 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2515 * Timing vulnerability in ECDSA signature generation
2517 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2518 timing side channel attack. An attacker could use variations in the signing
2519 algorithm to recover the private key.
2521 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2526 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2527 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2528 of two gigabytes and the error handling improved.
2530 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2531 categorized as a normal bug, not a security issue, because the DRBG reseeds
2532 automatically and is fully functional even without additional randomness
2533 provided by the application.
2535 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2537 * Add a new ClientHello callback. Provides a callback interface that gives
2538 the application the ability to adjust the nascent SSL object at the
2539 earliest stage of ClientHello processing, immediately after extensions have
2540 been collected but before they have been processed. In particular, this
2541 callback can adjust the supported TLS versions in response to the contents
2546 * Add SM2 base algorithm support.
2550 * s390x assembly pack: add (improved) hardware-support for the following
2551 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2552 aes-cfb/cfb8, aes-ecb.
2556 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2557 parameter is no longer accepted, as it leads to a corrupt table. NULL
2558 pem_str is reserved for alias entries only.
2562 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2563 step for prime curves. The new implementation is based on formulae from
2564 differential addition-and-doubling in homogeneous projective coordinates
2565 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2566 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2567 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2568 to work in projective coordinates.
2570 *Billy Bob Brumley, Nicola Tuveri*
2572 * Change generating and checking of primes so that the error rate of not
2573 being prime depends on the intended use based on the size of the input.
2574 For larger primes this will result in more rounds of Miller-Rabin.
2575 The maximal error rate for primes with more than 1080 bits is lowered
2578 *Kurt Roeckx, Annie Yousar*
2580 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2584 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2585 moving between systems, and to avoid confusion when a Windows build is
2586 done with mingw vs with MSVC. For POSIX installs, there's still a
2587 symlink or copy named 'tsget' to avoid that confusion as well.
2591 * Revert blinding in ECDSA sign and instead make problematic addition
2592 length-invariant. Switch even to fixed-length Montgomery multiplication.
2596 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2597 step for binary curves. The new implementation is based on formulae from
2598 differential addition-and-doubling in mixed Lopez-Dahab projective
2599 coordinates, modified to independently blind the operands.
2601 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2603 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2604 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2605 EC_METHODs to implement their own specialized "ladder step", to take
2606 advantage of more favorable coordinate systems or more efficient
2607 differential addition-and-doubling algorithms.
2609 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2611 * Modified the random device based seed sources to keep the relevant
2612 file descriptors open rather than reopening them on each access.
2613 This allows such sources to operate in a chroot() jail without
2614 the associated device nodes being available. This behaviour can be
2615 controlled using RAND_keep_random_devices_open().
2619 * Numerous side-channel attack mitigations have been applied. This may have
2620 performance impacts for some algorithms for the benefit of improved
2621 security. Specific changes are noted in this change log by their respective
2626 * AIX shared library support overhaul. Switch to AIX "natural" way of
2627 handling shared libraries, which means collecting shared objects of
2628 different versions and bitnesses in one common archive. This allows to
2629 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2630 doesn't affect the way 3rd party applications are linked, only how
2631 multi-version installation is managed.
2635 * Make ec_group_do_inverse_ord() more robust and available to other
2636 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2637 mitigations are applied to the fallback BN_mod_inverse().
2638 When using this function rather than BN_mod_inverse() directly, new
2639 EC cryptosystem implementations are then safer-by-default.
2643 * Add coordinate blinding for EC_POINT and implement projective
2644 coordinate blinding for generic prime curves as a countermeasure to
2645 chosen point SCA attacks.
2647 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2649 * Add blinding to ECDSA and DSA signatures to protect against side channel
2650 attacks discovered by Keegan Ryan (NCC Group).
2654 * Enforce checking in the `pkeyutl` command to ensure that the input
2655 length does not exceed the maximum supported digest length when performing
2656 a sign, verify or verifyrecover operation.
2660 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2661 I/O in combination with something like select() or poll() will hang. This
2662 can be turned off again using SSL_CTX_clear_mode().
2663 Many applications do not properly handle non-application data records, and
2664 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2665 around the problems in those applications, but can also break some.
2666 It's recommended to read the manpages about SSL_read(), SSL_write(),
2667 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2668 SSL_CTX_set_read_ahead() again.
2672 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2673 now allow empty (zero character) pass phrases.
2677 * Apply blinding to binary field modular inversion and remove patent
2678 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2682 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2683 binary and prime elliptic curves.
2687 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2688 constant time fixed point multiplication.
2692 * Revise elliptic curve scalar multiplication with timing attack
2693 defenses: ec_wNAF_mul redirects to a constant time implementation
2694 when computing fixed point and variable point multiplication (which
2695 in OpenSSL are mostly used with secret scalars in keygen, sign,
2696 ECDH derive operations).
2697 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2700 * Updated CONTRIBUTING
2704 * Updated DRBG / RAND to request nonce and additional low entropy
2705 randomness from the system.
2707 *Matthias St. Pierre*
2709 * Updated 'openssl rehash' to use OpenSSL consistent default.
2713 * Moved the load of the ssl_conf module to libcrypto, which helps
2714 loading engines that libssl uses before libssl is initialised.
2718 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2722 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2724 *Ingo Schwarze, Rich Salz*
2726 * Added output of accepting IP address and port for 'openssl s_server'
2730 * Added a new API for TLSv1.3 ciphersuites:
2731 SSL_CTX_set_ciphersuites()
2732 SSL_set_ciphersuites()
2736 * Memory allocation failures consistently add an error to the error
2741 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2742 in libcrypto when run as setuid/setgid.
2746 * Load any config file by default when libssl is used.
2750 * Added new public header file <openssl/rand_drbg.h> and documentation
2751 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2753 *Matthias St. Pierre*
2755 * QNX support removed (cannot find contributors to get their approval
2756 for the license change).
2760 * TLSv1.3 replay protection for early data has been implemented. See the
2761 SSL_read_early_data() man page for further details.
2765 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2766 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2767 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2768 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2769 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2770 configuration has been separated out. See the ciphers man page or the
2771 SSL_CTX_set_ciphersuites() man page for more information.
2775 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2776 in responder mode now supports the new "-multi" option, which
2777 spawns the specified number of child processes to handle OCSP
2778 requests. The "-timeout" option now also limits the OCSP
2779 responder's patience to wait to receive the full client request
2780 on a newly accepted connection. Child processes are respawned
2781 as needed, and the CA index file is automatically reloaded
2782 when changed. This makes it possible to run the "ocsp" responder
2783 as a long-running service, making the OpenSSL CA somewhat more
2784 feature-complete. In this mode, most diagnostic messages logged
2785 after entering the event loop are logged via syslog(3) rather than
2790 * Added support for X448 and Ed448. Heavily based on original work by
2795 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2796 objects loaded. This adds the functions OSSL_STORE_expect() and
2797 OSSL_STORE_find() as well as needed tools to construct searches and
2798 get the search data out of them.
2802 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2803 version of OpenSSL should review their configuration settings to ensure
2804 that they are still appropriate for TLSv1.3. For further information see:
2805 <https://wiki.openssl.org/index.php/TLS1.3>
2809 * Grand redesign of the OpenSSL random generator
2811 The default RAND method now utilizes an AES-CTR DRBG according to
2812 NIST standard SP 800-90Ar1. The new random generator is essentially
2813 a port of the default random generator from the OpenSSL FIPS 2.0
2814 object module. It is a hybrid deterministic random bit generator
2815 using an AES-CTR bit stream and which seeds and reseeds itself
2816 automatically using trusted system entropy sources.
2818 Some of its new features are:
2819 - Support for multiple DRBG instances with seed chaining.
2820 - The default RAND method makes use of a DRBG.
2821 - There is a public and private DRBG instance.
2822 - The DRBG instances are fork-safe.
2823 - Keep all global DRBG instances on the secure heap if it is enabled.
2824 - The public and private DRBG instance are per thread for lock free
2827 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2829 * Changed Configure so it only says what it does and doesn't dump
2830 so much data. Instead, ./configdata.pm should be used as a script
2831 to display all sorts of configuration data.
2835 * Added processing of "make variables" to Configure.
2839 * Added SHA512/224 and SHA512/256 algorithm support.
2843 * The last traces of Netware support, first removed in 1.1.0, have
2848 * Get rid of Makefile.shared, and in the process, make the processing
2849 of certain files (rc.obj, or the .def/.map/.opt files produced from
2850 the ordinal files) more visible and hopefully easier to trace and
2851 debug (or make silent).
2855 * Make it possible to have environment variable assignments as
2856 arguments to config / Configure.
2860 * Add multi-prime RSA (RFC 8017) support.
2864 * Add SM3 implemented according to GB/T 32905-2016
2865 *Jack Lloyd <jack.lloyd@ribose.com>,*
2866 *Ronald Tse <ronald.tse@ribose.com>,*
2867 *Erick Borsboom <erick.borsboom@ribose.com>*
2869 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2870 as documented in RFC6066.
2871 Based on a patch from Tomasz Moń
2873 *Filipe Raimundo da Silva*
2875 * Add SM4 implemented according to GB/T 32907-2016.
2876 *Jack Lloyd <jack.lloyd@ribose.com>,*
2877 *Ronald Tse <ronald.tse@ribose.com>,*
2878 *Erick Borsboom <erick.borsboom@ribose.com>*
2880 * Reimplement -newreq-nodes and ERR_error_string_n; the
2881 original author does not agree with the license change.
2885 * Add ARIA AEAD TLS support.
2889 * Some macro definitions to support VS6 have been removed. Visual
2890 Studio 6 has not worked since 1.1.0
2894 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2895 without clearing the errors.
2899 * Add "atfork" functions. If building on a system that without
2900 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2901 requirements. The RAND facility now uses/requires this.
2909 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2910 not possible to disable entirely. However, it's still possible to
2911 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2914 To disable, configure with 'no-ui-console'. 'no-ui' is still
2915 possible to use as an alias. Check at compile time with the
2916 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2917 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2921 * Add a STORE module, which implements a uniform and URI based reader of
2922 stores that can contain keys, certificates, CRLs and numerous other
2923 objects. The main API is loosely based on a few stdio functions,
2924 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2925 OSSL_STORE_error and OSSL_STORE_close.
2926 The implementation uses backends called "loaders" to implement arbitrary
2927 URI schemes. There is one built in "loader" for the 'file' scheme.
2931 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2932 then adjusted to work on FreeBSD 8.4 as well.
2933 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2934 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2938 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2939 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2940 error code calls like this:
2942 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2944 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2945 that can be encoded in C. For the foreseeable future, this will only
2948 *Richard Levitte and Tim Hudson*
2950 * Removed BSD cryptodev engine.
2954 * Add a build target 'build_all_generated', to build all generated files
2955 and only that. This can be used to prepare everything that requires
2956 things like perl for a system that lacks perl and then move everything
2957 to that system and do the rest of the build there.
2961 * In the UI interface, make it possible to duplicate the user data. This
2962 can be used by engines that need to retain the data for a longer time
2963 than just the call where this user data is passed.
2967 * Ignore the '-named_curve auto' value for compatibility of applications
2970 *Tomáš Mráz <tmraz@fedoraproject.org>*
2972 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2973 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2974 alerts across multiple records (some of which could be empty). In practice
2975 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2976 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2977 support this at all. Supporting it adds significant complexity to the
2978 record layer, and its removal is unlikely to cause interoperability
2983 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2984 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2985 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2990 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2991 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2993 *Richard Levitte, Andy Polyakov*
2995 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3000 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3001 platform rather than 'mingw'.
3005 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3006 success if they are asked to add an object which already exists
3007 in the store. This change cascades to other functions which load
3008 certificates and CRLs.
3012 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3013 facilitate stack unwinding even from assembly subroutines.
3017 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3018 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3022 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3023 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3024 which is the minimum version we support.
3028 * Certificate time validation (X509_cmp_time) enforces stricter
3029 compliance with RFC 5280. Fractional seconds and timezone offsets
3030 are no longer allowed.
3034 * Add support for ARIA
3038 * s_client will now send the Server Name Indication (SNI) extension by
3039 default unless the new "-noservername" option is used. The server name is
3040 based on the host provided to the "-connect" option unless overridden by
3041 using "-servername".
3045 * Add support for SipHash
3049 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3050 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3051 prevent issues where no progress is being made and the peer continually
3052 sends unrecognised record types, using up resources processing them.
3056 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3057 using the algorithm defined in
3058 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3062 * Heartbeat support has been removed; the ABI is changed for now.
3064 *Richard Levitte, Rich Salz*
3066 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3070 * The RSA "null" method, which was partially supported to avoid patent
3071 issues, has been replaced to always returns NULL.
3078 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3080 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3081 used even when parsing explicit parameters, when loading a encoded key
3082 or calling `EC_GROUP_new_from_ecpkparameters()`/
3083 `EC_GROUP_new_from_ecparameters()`.
3084 This prevents bypass of security hardening and performance gains,
3085 especially for curves with specialized EC_METHODs.
3086 By default, if a key encoded with explicit parameters is loaded and later
3087 encoded, the output is still encoded with explicit parameters, even if
3088 internally a "named" EC_GROUP is used for computation.
3092 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3093 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3094 NULL. After this change, only the cofactor parameter can be NULL. It also
3095 does some minimal sanity checks on the passed order.
3100 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3101 An attack is simple, if the first CMS_recipientInfo is valid but the
3102 second CMS_recipientInfo is chosen ciphertext. If the second
3103 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3104 encryption key will be replaced by garbage, and the message cannot be
3105 decoded, but if the RSA decryption fails, the correct encryption key is
3106 used and the recipient will not notice the attack.
3107 As a work around for this potential attack the length of the decrypted
3108 key must be equal to the cipher default key length, in case the
3109 certificate is not given and all recipientInfo are tried out.
3110 The old behaviour can be re-enabled in the CMS code by setting the
3111 CMS_DEBUG_DECRYPT flag.
3116 * Use Windows installation paths in the mingw builds
3118 Mingw isn't a POSIX environment per se, which means that Windows
3119 paths should be used for installation.
3124 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3126 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3127 This changes the size when using the `genpkey` command when no size is given.
3128 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3129 generation commands to use 2048 bits by default.
3133 * Prevent over long nonces in ChaCha20-Poly1305.
3135 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3136 for every encryption operation. RFC 7539 specifies that the nonce value
3137 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3138 and front pads the nonce with 0 bytes if it is less than 12
3139 bytes. However it also incorrectly allows a nonce to be set of up to 16
3140 bytes. In this case only the last 12 bytes are significant and any
3141 additional leading bytes are ignored.
3143 It is a requirement of using this cipher that nonce values are
3144 unique. Messages encrypted using a reused nonce value are susceptible to
3145 serious confidentiality and integrity attacks. If an application changes
3146 the default nonce length to be longer than 12 bytes and then makes a
3147 change to the leading bytes of the nonce expecting the new value to be a
3148 new unique nonce then such an application could inadvertently encrypt
3149 messages with a reused nonce.
3151 Additionally the ignored bytes in a long nonce are not covered by the
3152 integrity guarantee of this cipher. Any application that relies on the
3153 integrity of these ignored leading bytes of a long nonce may be further
3154 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3155 is safe because no such use sets such a long nonce value. However user
3156 applications that use this cipher directly and set a non-default nonce
3157 length to be longer than 12 bytes may be vulnerable.
3159 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3165 * Added SCA hardening for modular field inversion in EC_GROUP through
3166 a new dedicated field_inv() pointer in EC_METHOD.
3167 This also addresses a leakage affecting conversions from projective
3168 to affine coordinates.
3170 *Billy Bob Brumley, Nicola Tuveri*
3172 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3173 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3177 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3181 * Remove the 'dist' target and add a tarball building script. The
3182 'dist' target has fallen out of use, and it shouldn't be
3183 necessary to configure just to create a source distribution.
3187 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3189 * Timing vulnerability in DSA signature generation
3191 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3192 timing side channel attack. An attacker could use variations in the signing
3193 algorithm to recover the private key.
3195 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3200 * Timing vulnerability in ECDSA signature generation
3202 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3203 timing side channel attack. An attacker could use variations in the signing
3204 algorithm to recover the private key.
3206 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3211 * Add coordinate blinding for EC_POINT and implement projective
3212 coordinate blinding for generic prime curves as a countermeasure to
3213 chosen point SCA attacks.
3215 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3217 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3219 * Client DoS due to large DH parameter
3221 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3222 malicious server can send a very large prime value to the client. This will
3223 cause the client to spend an unreasonably long period of time generating a
3224 key for this prime resulting in a hang until the client has finished. This
3225 could be exploited in a Denial Of Service attack.
3227 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3232 * Cache timing vulnerability in RSA Key Generation
3234 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3235 a cache timing side channel attack. An attacker with sufficient access to
3236 mount cache timing attacks during the RSA key generation process could
3237 recover the private key.
3239 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3240 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3245 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3246 parameter is no longer accepted, as it leads to a corrupt table. NULL
3247 pem_str is reserved for alias entries only.
3251 * Revert blinding in ECDSA sign and instead make problematic addition
3252 length-invariant. Switch even to fixed-length Montgomery multiplication.
3256 * Change generating and checking of primes so that the error rate of not
3257 being prime depends on the intended use based on the size of the input.
3258 For larger primes this will result in more rounds of Miller-Rabin.
3259 The maximal error rate for primes with more than 1080 bits is lowered
3262 *Kurt Roeckx, Annie Yousar*
3264 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3268 * Add blinding to ECDSA and DSA signatures to protect against side channel
3269 attacks discovered by Keegan Ryan (NCC Group).
3273 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3274 now allow empty (zero character) pass phrases.
3278 * Certificate time validation (X509_cmp_time) enforces stricter
3279 compliance with RFC 5280. Fractional seconds and timezone offsets
3280 are no longer allowed.
3284 * Fixed a text canonicalisation bug in CMS
3286 Where a CMS detached signature is used with text content the text goes
3287 through a canonicalisation process first prior to signing or verifying a
3288 signature. This process strips trailing space at the end of lines, converts
3289 line terminators to CRLF and removes additional trailing line terminators
3290 at the end of a file. A bug in the canonicalisation process meant that
3291 some characters, such as form-feed, were incorrectly treated as whitespace
3292 and removed. This is contrary to the specification (RFC5485). This fix
3293 could mean that detached text data signed with an earlier version of
3294 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3295 signed with a fixed OpenSSL may fail to verify with an earlier version of
3296 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3297 and use the "-binary" flag (for the "cms" command line application) or set
3298 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3302 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3304 * Constructed ASN.1 types with a recursive definition could exceed the stack
3306 Constructed ASN.1 types with a recursive definition (such as can be found
3307 in PKCS7) could eventually exceed the stack given malicious input with
3308 excessive recursion. This could result in a Denial Of Service attack. There
3309 are no such structures used within SSL/TLS that come from untrusted sources
3310 so this is considered safe.
3312 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3318 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3320 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3321 effectively reduced to only comparing the least significant bit of each
3322 byte. This allows an attacker to forge messages that would be considered as
3323 authenticated in an amount of tries lower than that guaranteed by the
3324 security claims of the scheme. The module can only be compiled by the
3325 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3327 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3333 * Add a build target 'build_all_generated', to build all generated files
3334 and only that. This can be used to prepare everything that requires
3335 things like perl for a system that lacks perl and then move everything
3336 to that system and do the rest of the build there.
3340 * Backport SSL_OP_NO_RENGOTIATION
3342 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3343 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3344 changes this is no longer possible in 1.1.0. Therefore, the new
3345 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3346 1.1.0 to provide equivalent functionality.
3348 Note that if an application built against 1.1.0h headers (or above) is run
3349 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3350 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3354 * Removed the OS390-Unix config target. It relied on a script that doesn't
3359 * rsaz_1024_mul_avx2 overflow bug on x86_64
3361 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3362 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3363 Analysis suggests that attacks against RSA and DSA as a result of this
3364 defect would be very difficult to perform and are not believed likely.
3365 Attacks against DH1024 are considered just feasible, because most of the
3366 work necessary to deduce information about a private key may be performed
3367 offline. The amount of resources required for such an attack would be
3368 significant. However, for an attack on TLS to be meaningful, the server
3369 would have to share the DH1024 private key among multiple clients, which is
3370 no longer an option since CVE-2016-0701.
3372 This only affects processors that support the AVX2 but not ADX extensions
3373 like Intel Haswell (4th generation).
3375 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3376 was originally found via the OSS-Fuzz project.
3381 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3383 * bn_sqrx8x_internal carry bug on x86_64
3385 There is a carry propagating bug in the x86_64 Montgomery squaring
3386 procedure. No EC algorithms are affected. Analysis suggests that attacks
3387 against RSA and DSA as a result of this defect would be very difficult to
3388 perform and are not believed likely. Attacks against DH are considered just
3389 feasible (although very difficult) because most of the work necessary to
3390 deduce information about a private key may be performed offline. The amount
3391 of resources required for such an attack would be very significant and
3392 likely only accessible to a limited number of attackers. An attacker would
3393 additionally need online access to an unpatched system using the target
3394 private key in a scenario with persistent DH parameters and a private
3395 key that is shared between multiple clients.
3397 This only affects processors that support the BMI1, BMI2 and ADX extensions
3398 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3400 This issue was reported to OpenSSL by the OSS-Fuzz project.
3405 * Malformed X.509 IPAddressFamily could cause OOB read
3407 If an X.509 certificate has a malformed IPAddressFamily extension,
3408 OpenSSL could do a one-byte buffer overread. The most likely result
3409 would be an erroneous display of the certificate in text format.
3411 This issue was reported to OpenSSL by the OSS-Fuzz project.
3416 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3418 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3419 platform rather than 'mingw'.
3423 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3424 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3425 which is the minimum version we support.
3429 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3431 * Encrypt-Then-Mac renegotiation crash
3433 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3434 negotiated where it was not in the original handshake (or vice-versa) then
3435 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3436 and servers are affected.
3438 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3443 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3445 * Truncated packet could crash via OOB read
3447 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3448 cipher is being used, then a truncated packet can cause that host to
3449 perform an out-of-bounds read, usually resulting in a crash.
3451 This issue was reported to OpenSSL by Robert Święcki of Google.
3456 * Bad (EC)DHE parameters cause a client crash
3458 If a malicious server supplies bad parameters for a DHE or ECDHE key
3459 exchange then this can result in the client attempting to dereference a
3460 NULL pointer leading to a client crash. This could be exploited in a Denial
3463 This issue was reported to OpenSSL by Guido Vranken.
3468 * BN_mod_exp may produce incorrect results on x86_64
3470 There is a carry propagating bug in the x86_64 Montgomery squaring
3471 procedure. No EC algorithms are affected. Analysis suggests that attacks
3472 against RSA and DSA as a result of this defect would be very difficult to
3473 perform and are not believed likely. Attacks against DH are considered just
3474 feasible (although very difficult) because most of the work necessary to
3475 deduce information about a private key may be performed offline. The amount
3476 of resources required for such an attack would be very significant and
3477 likely only accessible to a limited number of attackers. An attacker would
3478 additionally need online access to an unpatched system using the target
3479 private key in a scenario with persistent DH parameters and a private
3480 key that is shared between multiple clients. For example this can occur by
3481 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3482 similar to CVE-2015-3193 but must be treated as a separate problem.
3484 This issue was reported to OpenSSL by the OSS-Fuzz project.
3489 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3491 * ChaCha20/Poly1305 heap-buffer-overflow
3493 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3494 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3495 crash. This issue is not considered to be exploitable beyond a DoS.
3497 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3502 * CMS Null dereference
3504 Applications parsing invalid CMS structures can crash with a NULL pointer
3505 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3506 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3507 structure callback if an attempt is made to free certain invalid encodings.
3508 Only CHOICE structures using a callback which do not handle NULL value are
3511 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3516 * Montgomery multiplication may produce incorrect results
3518 There is a carry propagating bug in the Broadwell-specific Montgomery
3519 multiplication procedure that handles input lengths divisible by, but
3520 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3521 and DH private keys are impossible. This is because the subroutine in
3522 question is not used in operations with the private key itself and an input
3523 of the attacker's direct choice. Otherwise the bug can manifest itself as
3524 transient authentication and key negotiation failures or reproducible
3525 erroneous outcome of public-key operations with specially crafted input.
3526 Among EC algorithms only Brainpool P-512 curves are affected and one
3527 presumably can attack ECDH key negotiation. Impact was not analyzed in
3528 detail, because pre-requisites for attack are considered unlikely. Namely
3529 multiple clients have to choose the curve in question and the server has to
3530 share the private key among them, neither of which is default behaviour.
3531 Even then only clients that chose the curve will be affected.
3533 This issue was publicly reported as transient failures and was not
3534 initially recognized as a security issue. Thanks to Richard Morgan for
3535 providing reproducible case.
3540 * Removed automatic addition of RPATH in shared libraries and executables,
3541 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3545 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3547 * Fix Use After Free for large message sizes
3549 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3550 message larger than approx 16k is received then the underlying buffer to
3551 store the incoming message is reallocated and moved. Unfortunately a
3552 dangling pointer to the old location is left which results in an attempt to
3553 write to the previously freed location. This is likely to result in a
3554 crash, however it could potentially lead to execution of arbitrary code.
3556 This issue only affects OpenSSL 1.1.0a.
3558 This issue was reported to OpenSSL by Robert Święcki.
3563 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3565 * OCSP Status Request extension unbounded memory growth
3567 A malicious client can send an excessively large OCSP Status Request
3568 extension. If that client continually requests renegotiation, sending a
3569 large OCSP Status Request extension each time, then there will be unbounded
3570 memory growth on the server. This will eventually lead to a Denial Of
3571 Service attack through memory exhaustion. Servers with a default
3572 configuration are vulnerable even if they do not support OCSP. Builds using
3573 the "no-ocsp" build time option are not affected.
3575 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3580 * SSL_peek() hang on empty record
3582 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3583 sends an empty record. This could be exploited by a malicious peer in a
3584 Denial Of Service attack.
3586 This issue was reported to OpenSSL by Alex Gaynor.
3591 * Excessive allocation of memory in tls_get_message_header() and
3592 dtls1_preprocess_fragment()
3594 A (D)TLS message includes 3 bytes for its length in the header for the
3595 message. This would allow for messages up to 16Mb in length. Messages of
3596 this length are excessive and OpenSSL includes a check to ensure that a
3597 peer is sending reasonably sized messages in order to avoid too much memory
3598 being consumed to service a connection. A flaw in the logic of version
3599 1.1.0 means that memory for the message is allocated too early, prior to
3600 the excessive message length check. Due to way memory is allocated in
3601 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3602 to service a connection. This could lead to a Denial of Service through
3603 memory exhaustion. However, the excessive message length check still takes
3604 place, and this would cause the connection to immediately fail. Assuming
3605 that the application calls SSL_free() on the failed connection in a timely
3606 manner then the 21Mb of allocated memory will then be immediately freed
3607 again. Therefore, the excessive memory allocation will be transitory in
3608 nature. This then means that there is only a security impact if:
3610 1) The application does not call SSL_free() in a timely manner in the event
3611 that the connection fails
3613 2) The application is working in a constrained environment where there is
3614 very little free memory
3616 3) The attacker initiates multiple connection attempts such that there are
3617 multiple connections in a state where memory has been allocated for the
3618 connection; SSL_free() has not yet been called; and there is insufficient
3619 memory to service the multiple requests.
3621 Except in the instance of (1) above any Denial Of Service is likely to be
3622 transitory because as soon as the connection fails the memory is
3623 subsequently freed again in the SSL_free() call. However there is an
3624 increased risk during this period of application crashes due to the lack of
3625 memory - which would then mean a more serious Denial of Service.
3627 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3628 (CVE-2016-6307 and CVE-2016-6308)
3632 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3633 had to be removed. Primary reason is that vendor assembler can't
3634 assemble our modules with -KPIC flag. As result it, assembly
3635 support, was not even available as option. But its lack means
3636 lack of side-channel resistant code, which is incompatible with
3637 security by todays standards. Fortunately gcc is readily available
3638 prepackaged option, which we firmly point at...
3642 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3644 * Windows command-line tool supports UTF-8 opt-in option for arguments
3645 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3646 (to any value) allows Windows user to access PKCS#12 file generated
3647 with Windows CryptoAPI and protected with non-ASCII password, as well
3648 as files generated under UTF-8 locale on Linux also protected with
3653 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3654 have been disabled by default and removed from DEFAULT, just like RC4.
3655 See the RC4 item below to re-enable both.
3659 * The method for finding the storage location for the Windows RAND seed file
3660 has changed. First we check %RANDFILE%. If that is not set then we check
3661 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3662 all else fails we fall back to C:\.
3666 * The EVP_EncryptUpdate() function has had its return type changed from void
3667 to int. A return of 0 indicates and error while a return of 1 indicates
3672 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3673 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3674 off the constant time implementation for RSA, DSA and DH have been made
3675 no-ops and deprecated.
3679 * Windows RAND implementation was simplified to only get entropy by
3680 calling CryptGenRandom(). Various other RAND-related tickets
3683 *Joseph Wylie Yandle, Rich Salz*
3685 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3686 and `OPENSSL_LH_`, respectively. The old names are available
3687 with API compatibility. They new names are now completely documented.
3691 * Unify TYPE_up_ref(obj) methods signature.
3692 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3693 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3694 int (instead of void) like all others TYPE_up_ref() methods.
3695 So now these methods also check the return value of CRYPTO_atomic_add(),
3696 and the validity of object reference counter.
3698 *fdasilvayy@gmail.com*
3700 * With Windows Visual Studio builds, the .pdb files are installed
3701 alongside the installed libraries and executables. For a static
3702 library installation, ossl_static.pdb is the associate compiler
3703 generated .pdb file to be used when linking programs.
3707 * Remove openssl.spec. Packaging files belong with the packagers.
3711 * Automatic Darwin/OSX configuration has had a refresh, it will now
3712 recognise x86_64 architectures automatically. You can still decide
3713 to build for a different bitness with the environment variable
3714 KERNEL_BITS (can be 32 or 64), for example:
3716 KERNEL_BITS=32 ./config
3720 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3721 256 bit AES and HMAC with SHA256.
3725 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3729 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3733 * To enable users to have their own config files and build file templates,
3734 Configure looks in the directory indicated by the environment variable
3735 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3736 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3737 name and is used as is.
3741 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3742 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3743 X509_CERT_FILE_CTX was removed.
3747 * "shared" builds are now the default. To create only static libraries use
3748 the "no-shared" Configure option.
3752 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3753 All of these option have not worked for some while and are fundamental
3758 * Make various cleanup routines no-ops and mark them as deprecated. Most
3759 global cleanup functions are no longer required because they are handled
3760 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3761 Explicitly de-initing can cause problems (e.g. where a library that uses
3762 OpenSSL de-inits, but an application is still using it). The affected
3763 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3764 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3765 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3766 COMP_zlib_cleanup().
3770 * --strict-warnings no longer enables runtime debugging options
3771 such as REF_DEBUG. Instead, debug options are automatically
3772 enabled with '--debug' builds.
3774 *Andy Polyakov, Emilia Käsper*
3776 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3777 have been moved out of the public header files. New functions for managing
3778 these have been added.
3782 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3783 objects have been moved out of the public header files. New
3784 functions for managing these have been added.
3788 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3789 have been moved out of the public header files. New functions for managing
3790 these have been added.
3794 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3795 moved out of the public header files. New functions for managing these
3800 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3804 * Removed the mk1mf build scripts.
3808 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3809 it is always safe to #include a header now.
3813 * Removed the aged BC-32 config and all its supporting scripts
3817 * Removed support for Ultrix, Netware, and OS/2.
3821 * Add support for HKDF.
3823 *Alessandro Ghedini*
3825 * Add support for blake2b and blake2s
3829 * Added support for "pipelining". Ciphers that have the
3830 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3831 encryptions/decryptions simultaneously. There are currently no built-in
3832 ciphers with this property but the expectation is that engines will be able
3833 to offer it to significantly improve throughput. Support has been extended
3834 into libssl so that multiple records for a single connection can be
3835 processed in one go (for >=TLS 1.1).
3839 * Added the AFALG engine. This is an async capable engine which is able to
3840 offload work to the Linux kernel. In this initial version it only supports
3841 AES128-CBC. The kernel must be version 4.1.0 or greater.
3845 * OpenSSL now uses a new threading API. It is no longer necessary to
3846 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3847 are two supported threading models: pthreads and windows threads. It is
3848 also possible to configure OpenSSL at compile time for "no-threads". The
3849 old threading API should no longer be used. The functions have been
3850 replaced with "no-op" compatibility macros.
3852 *Alessandro Ghedini, Matt Caswell*
3854 * Modify behavior of ALPN to invoke callback after SNI/servername
3855 callback, such that updates to the SSL_CTX affect ALPN.
3859 * Add SSL_CIPHER queries for authentication and key-exchange.
3863 * Changes to the DEFAULT cipherlist:
3864 - Prefer (EC)DHE handshakes over plain RSA.
3865 - Prefer AEAD ciphers over legacy ciphers.
3866 - Prefer ECDSA over RSA when both certificates are available.
3867 - Prefer TLSv1.2 ciphers/PRF.
3868 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3873 * Change the ECC default curve list to be this, in order: x25519,
3874 secp256r1, secp521r1, secp384r1.
3878 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3879 disabled by default. They can be re-enabled using the
3880 enable-weak-ssl-ciphers option to Configure.
3884 * If the server has ALPN configured, but supports no protocols that the
3885 client advertises, send a fatal "no_application_protocol" alert.
3886 This behaviour is SHALL in RFC 7301, though it isn't universally
3887 implemented by other servers.
3891 * Add X25519 support.
3892 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3893 for public and private key encoding using the format documented in
3894 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3895 key generation and key derivation.
3897 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3902 * Deprecate SRP_VBASE_get_by_user.
3903 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3904 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3905 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3906 seed, even if the seed is configured.
3908 Users should use SRP_VBASE_get1_by_user instead. Note that in
3909 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3910 also that even though configuring the SRP seed attempts to hide
3911 invalid usernames by continuing the handshake with fake
3912 credentials, this behaviour is not constant time and no strong
3913 guarantees are made that the handshake is indistinguishable from
3914 that of a valid user.
3918 * Configuration change; it's now possible to build dynamic engines
3919 without having to build shared libraries and vice versa. This
3920 only applies to the engines in `engines/`, those in `crypto/engine/`
3921 will always be built into libcrypto (i.e. "static").
3923 Building dynamic engines is enabled by default; to disable, use
3924 the configuration option "disable-dynamic-engine".
3926 The only requirements for building dynamic engines are the
3927 presence of the DSO module and building with position independent
3928 code, so they will also automatically be disabled if configuring
3929 with "disable-dso" or "disable-pic".
3931 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3932 are also taken away from openssl/opensslconf.h, as they are
3937 * Configuration change; if there is a known flag to compile
3938 position independent code, it will always be applied on the
3939 libcrypto and libssl object files, and never on the application
3940 object files. This means other libraries that use routines from
3941 libcrypto / libssl can be made into shared libraries regardless
3942 of how OpenSSL was configured.
3944 If this isn't desirable, the configuration options "disable-pic"
3945 or "no-pic" can be used to disable the use of PIC. This will
3946 also disable building shared libraries and dynamic engines.
3950 * Removed JPAKE code. It was experimental and has no wide use.
3954 * The INSTALL_PREFIX Makefile variable has been renamed to
3955 DESTDIR. That makes for less confusion on what this variable
3956 is for. Also, the configuration option --install_prefix is
3961 * Heartbeat for TLS has been removed and is disabled by default
3962 for DTLS; configure with enable-heartbeats. Code that uses the
3963 old #define's might need to be updated.
3965 *Emilia Käsper, Rich Salz*
3967 * Rename REF_CHECK to REF_DEBUG.
3971 * New "unified" build system
3973 The "unified" build system is aimed to be a common system for all
3974 platforms we support. With it comes new support for VMS.
3976 This system builds supports building in a different directory tree
3977 than the source tree. It produces one Makefile (for unix family
3978 or lookalikes), or one descrip.mms (for VMS).
3980 The source of information to make the Makefile / descrip.mms is
3981 small files called 'build.info', holding the necessary
3982 information for each directory with source to compile, and a
3983 template in Configurations, like unix-Makefile.tmpl or
3986 With this change, the library names were also renamed on Windows
3987 and on VMS. They now have names that are closer to the standard
3988 on Unix, and include the major version number, and in certain
3989 cases, the architecture they are built for. See "Notes on shared
3990 libraries" in INSTALL.
3992 We rely heavily on the perl module Text::Template.
3996 * Added support for auto-initialisation and de-initialisation of the library.
3997 OpenSSL no longer requires explicit init or deinit routines to be called,
3998 except in certain circumstances. See the OPENSSL_init_crypto() and
3999 OPENSSL_init_ssl() man pages for further information.
4003 * The arguments to the DTLSv1_listen function have changed. Specifically the
4004 "peer" argument is now expected to be a BIO_ADDR object.
4006 * Rewrite of BIO networking library. The BIO library lacked consistent
4007 support of IPv6, and adding it required some more extensive
4008 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4009 which hold all types of addresses and chains of address information.
4010 It also introduces a new API, with functions like BIO_socket,
4011 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4012 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4013 have been adapted accordingly.
4017 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4022 * CRIME protection: disable compression by default, even if OpenSSL is
4023 compiled with zlib enabled. Applications can still enable compression
4024 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4025 using the SSL_CONF library to configure compression.
4029 * The signature of the session callback configured with
4030 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4031 was explicitly marked as `const unsigned char*` instead of
4036 * Always DPURIFY. Remove the use of uninitialized memory in the
4037 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4041 * Removed many obsolete configuration items, including
4042 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4043 MD2_CHAR, MD2_INT, MD2_LONG
4045 IDEA_SHORT, IDEA_LONG
4046 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4048 *Rich Salz, with advice from Andy Polyakov*
4050 * Many BN internals have been moved to an internal header file.
4052 *Rich Salz with help from Andy Polyakov*
4054 * Configuration and writing out the results from it has changed.
4055 Files such as Makefile include/openssl/opensslconf.h and are now
4056 produced through general templates, such as Makefile.in and
4057 crypto/opensslconf.h.in and some help from the perl module
4060 Also, the center of configuration information is no longer
4061 Makefile. Instead, Configure produces a perl module in
4062 configdata.pm which holds most of the config data (in the hash
4063 table %config), the target data that comes from the target
4064 configuration in one of the `Configurations/*.conf` files (in
4069 * To clarify their intended purposes, the Configure options
4070 --prefix and --openssldir change their semantics, and become more
4071 straightforward and less interdependent.
4073 --prefix shall be used exclusively to give the location INSTALLTOP
4074 where programs, scripts, libraries, include files and manuals are
4075 going to be installed. The default is now /usr/local.
4077 --openssldir shall be used exclusively to give the default
4078 location OPENSSLDIR where certificates, private keys, CRLs are
4079 managed. This is also where the default openssl.cnf gets
4081 If the directory given with this option is a relative path, the
4082 values of both the --prefix value and the --openssldir value will
4083 be combined to become OPENSSLDIR.
4084 The default for --openssldir is INSTALLTOP/ssl.
4086 Anyone who uses --openssldir to specify where OpenSSL is to be
4087 installed MUST change to use --prefix instead.
4091 * The GOST engine was out of date and therefore it has been removed. An up
4092 to date GOST engine is now being maintained in an external repository.
4093 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4094 support for GOST ciphersuites (these are only activated if a GOST engine
4099 * EGD is no longer supported by default; use enable-egd when
4102 *Ben Kaduk and Rich Salz*
4104 * The distribution now has Makefile.in files, which are used to
4105 create Makefile's when Configure is run. *Configure must be run
4106 before trying to build now.*
4110 * The return value for SSL_CIPHER_description() for error conditions
4115 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4117 Obtaining and performing DNSSEC validation of TLSA records is
4118 the application's responsibility. The application provides
4119 the TLSA records of its choice to OpenSSL, and these are then
4120 used to authenticate the peer.
4122 The TLSA records need not even come from DNS. They can, for
4123 example, be used to implement local end-entity certificate or
4124 trust-anchor "pinning", where the "pin" data takes the form
4125 of TLSA records, which can augment or replace verification
4126 based on the usual WebPKI public certification authorities.
4130 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4131 continues to support deprecated interfaces in default builds.
4132 However, applications are strongly advised to compile their
4133 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4134 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4135 or the 1.1.0 releases.
4137 In environments in which all applications have been ported to
4138 not use any deprecated interfaces OpenSSL's Configure script
4139 should be used with the --api=1.1.0 option to entirely remove
4140 support for the deprecated features from the library and
4141 unconditionally disable them in the installed headers.
4142 Essentially the same effect can be achieved with the "no-deprecated"
4143 argument to Configure, except that this will always restrict
4144 the build to just the latest API, rather than a fixed API
4147 As applications are ported to future revisions of the API,
4148 they should update their compile-time OPENSSL_API_COMPAT define
4149 accordingly, but in most cases should be able to continue to
4150 compile with later releases.
4152 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4153 0x10000000L and 0x00908000L, respectively. However those
4154 versions did not support the OPENSSL_API_COMPAT feature, and
4155 so applications are not typically tested for explicit support
4156 of just the undeprecated features of either release.
4160 * Add support for setting the minimum and maximum supported protocol.
4161 It can bet set via the SSL_set_min_proto_version() and
4162 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4163 MaxProtocol. It's recommended to use the new APIs to disable
4164 protocols instead of disabling individual protocols using
4165 SSL_set_options() or SSL_CONF's Protocol. This change also
4166 removes support for disabling TLS 1.2 in the OpenSSL TLS
4167 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4171 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4175 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4176 and integrates ECDSA and ECDH functionality into EC. Implementations can
4177 now redirect key generation and no longer need to convert to or from
4180 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4181 include the ec.h header file instead.
4185 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4186 ciphers who are no longer supported and drops support the ephemeral RSA key
4187 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4191 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4192 opaque. For HMAC_CTX, the following constructors and destructors
4195 HMAC_CTX *HMAC_CTX_new(void);
4196 void HMAC_CTX_free(HMAC_CTX *ctx);
4198 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4199 destroy such methods has been added. See EVP_MD_meth_new(3) and
4200 EVP_CIPHER_meth_new(3) for documentation.
4203 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4204 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4205 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4206 an already created structure.
4207 2) For consistency with the majority of our object creators and
4208 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4209 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4210 for deprecated builds.
4214 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4215 cryptographic operations to be performed asynchronously as long as an
4216 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4217 further details. Libssl has also had this capability integrated with the
4218 introduction of the new mode SSL_MODE_ASYNC and associated error
4219 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4220 pages. This work was developed in partnership with Intel Corp.
4224 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4225 always enabled now. If you want to disable the support you should
4226 exclude it using the list of supported ciphers. This also means that the
4227 "-no_ecdhe" option has been removed from s_server.
4231 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4232 SSL_{CTX_}set1_curves() which can set a list.
4236 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4237 curve you want to support using SSL_{CTX_}set1_curves().
4241 * State machine rewrite. The state machine code has been significantly
4242 refactored in order to remove much duplication of code and solve issues
4243 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4244 further details). This change does have some associated API changes.
4245 Notably the SSL_state() function has been removed and replaced by
4246 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4247 SSL_set_state() has been removed altogether. The previous handshake states
4248 defined in ssl.h and ssl3.h have also been removed.
4252 * All instances of the string "ssleay" in the public API were replaced
4253 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4254 Some error codes related to internal RSA_eay API's were renamed.
4258 * The demo files in crypto/threads were moved to demo/threads.
4262 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4265 *Matt Caswell, Rich Salz*
4267 * New ASN.1 embed macro.
4269 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4270 structure is not allocated: it is part of the parent. That is instead of
4278 This reduces memory fragmentation and make it impossible to accidentally
4279 set a mandatory field to NULL.
4281 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4282 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4283 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4288 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4292 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4293 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4294 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4295 DES and RC4 ciphersuites.
4299 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4300 This changes the decoding behaviour for some invalid messages,
4301 though the change is mostly in the more lenient direction, and
4302 legacy behaviour is preserved as much as possible.
4306 * Fix no-stdio build.
4307 *David Woodhouse <David.Woodhouse@intel.com> and also*
4308 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4310 * New testing framework
4311 The testing framework has been largely rewritten and is now using
4312 perl and the perl modules Test::Harness and an extended variant of
4313 Test::More called OpenSSL::Test to do its work. All test scripts in
4314 test/ have been rewritten into test recipes, and all direct calls to
4315 executables in test/Makefile have become individual recipes using the
4316 simplified testing OpenSSL::Test::Simple.
4318 For documentation on our testing modules, do:
4320 perldoc test/testlib/OpenSSL/Test/Simple.pm
4321 perldoc test/testlib/OpenSSL/Test.pm
4325 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4326 are used; the latter aborts on memory leaks (usually checked on exit).
4327 Some undocumented "set malloc, etc., hooks" functions were removed
4328 and others were changed. All are now documented.
4332 * In DSA_generate_parameters_ex, if the provided seed is too short,
4335 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4337 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4338 from RFC4279, RFC4785, RFC5487, RFC5489.
4340 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4341 original RSA_PSK patch.
4345 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4346 era flag was never set throughout the codebase (only read). Also removed
4347 SSL3_FLAGS_POP_BUFFER which was only used if
4348 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4352 * Changed the default name options in the "ca", "crl", "req" and "x509"
4353 to be "oneline" instead of "compat".
4357 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4358 not aware of clients that still exhibit this bug, and the workaround
4359 hasn't been working properly for a while.
4363 * The return type of BIO_number_read() and BIO_number_written() as well as
4364 the corresponding num_read and num_write members in the BIO structure has
4365 changed from unsigned long to uint64_t. On platforms where an unsigned
4366 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4371 * Given the pervasive nature of TLS extensions it is inadvisable to run
4372 OpenSSL without support for them. It also means that maintaining
4373 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4374 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4378 * Removed support for the two export grade static DH ciphersuites
4379 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4380 were newly added (along with a number of other static DH ciphersuites) to
4381 1.0.2. However the two export ones have *never* worked since they were
4382 introduced. It seems strange in any case to be adding new export
4383 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4387 * Version negotiation has been rewritten. In particular SSLv23_method(),
4388 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4389 and turned into macros which simply call the new preferred function names
4390 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4391 should use the new names instead. Also as part of this change the ssl23.h
4392 header file has been removed.
4396 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4397 code and the associated standard is no longer considered fit-for-purpose.
4401 * RT2547 was closed. When generating a private key, try to make the
4402 output file readable only by the owner. This behavior change might
4403 be noticeable when interacting with other software.
4405 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4410 * Added HTTP GET support to the ocsp command.
4414 * Changed default digest for the dgst and enc commands from MD5 to
4419 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4423 * Added support for TLS extended master secret from
4424 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4425 initial patch which was a great help during development.
4429 * All libssl internal structures have been removed from the public header
4430 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4431 now redundant). Users should not attempt to access internal structures
4432 directly. Instead they should use the provided API functions.
4436 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4437 Access to deprecated functions can be re-enabled by running config with
4438 "enable-deprecated". In addition applications wishing to use deprecated
4439 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4440 will, by default, disable some transitive includes that previously existed
4441 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4445 * Added support for OCB mode. OpenSSL has been granted a patent license
4446 compatible with the OpenSSL license for use of OCB. Details are available
4447 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4448 for OCB can be removed by calling config with no-ocb.
4452 * SSLv2 support has been removed. It still supports receiving an SSLv2
4453 compatible client hello.
4457 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4458 done while fixing the error code for the key-too-small case.
4460 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4462 * CA.sh has been removed; use CA.pl instead.
4466 * Removed old DES API.
4470 * Remove various unsupported platforms:
4476 Sinix/ReliantUNIX RM400
4481 16-bit platforms such as WIN16
4485 * Clean up OPENSSL_NO_xxx #define's
4486 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4487 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4488 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4489 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4490 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4491 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4492 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4493 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4494 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4495 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4499 * Cleaned up dead code
4500 Remove all but one '#ifdef undef' which is to be looked at.
4504 * Clean up calling of xxx_free routines.
4505 Just like free(), fix most of the xxx_free routines to accept
4506 NULL. Remove the non-null checks from callers. Save much code.
4510 * Add secure heap for storage of private keys (when possible).
4511 Add BIO_s_secmem(), CBIGNUM, etc.
4512 Contributed by Akamai Technologies under our Corporate CLA.
4516 * Experimental support for a new, fast, unbiased prime candidate generator,
4517 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4519 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4521 * New output format NSS in the sess_id command line tool. This allows
4522 exporting the session id and the master key in NSS keylog format.
4524 *Martin Kaiser <martin@kaiser.cx>*
4526 * Harmonize version and its documentation. -f flag is used to display
4529 *mancha <mancha1@zoho.com>*
4531 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4532 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4534 *mancha <mancha1@zoho.com>*
4536 * Fix some double frees. These are not thought to be exploitable.
4538 *mancha <mancha1@zoho.com>*
4540 * A missing bounds check in the handling of the TLS heartbeat extension
4541 can be used to reveal up to 64k of memory to a connected client or
4544 Thanks for Neel Mehta of Google Security for discovering this bug and to
4545 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4546 preparing the fix ([CVE-2014-0160])
4548 *Adam Langley, Bodo Moeller*
4550 * Fix for the attack described in the paper "Recovering OpenSSL
4551 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4552 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4553 <http://eprint.iacr.org/2014/140>
4555 Thanks to Yuval Yarom and Naomi Benger for discovering this
4556 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4558 *Yuval Yarom and Naomi Benger*
4560 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4561 this fixes a limitation in previous versions of OpenSSL.
4565 * Experimental encrypt-then-mac support.
4567 Experimental support for encrypt then mac from
4568 draft-gutmann-tls-encrypt-then-mac-02.txt
4570 To enable it set the appropriate extension number (0x42 for the test
4571 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4573 For non-compliant peers (i.e. just about everything) this should have no
4576 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4580 * Add EVP support for key wrapping algorithms, to avoid problems with
4581 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4582 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4583 algorithms and include tests cases.
4587 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4592 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4593 MGF1 digest and OAEP label.
4597 * Make openssl verify return errors.
4599 *Chris Palmer <palmer@google.com> and Ben Laurie*
4601 * New function ASN1_TIME_diff to calculate the difference between two
4602 ASN1_TIME structures or one structure and the current time.
4606 * Update fips_test_suite to support multiple command line options. New
4607 test to induce all self test errors in sequence and check expected
4612 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4613 sign or verify all in one operation.
4617 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4618 test programs and fips_test_suite. Includes functionality to parse
4619 the minimal script output of fipsalgest.pl directly.
4623 * Add authorisation parameter to FIPS_module_mode_set().
4627 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4631 * Use separate DRBG fields for internal and external flags. New function
4632 FIPS_drbg_health_check() to perform on demand health checking. Add
4633 generation tests to fips_test_suite with reduced health check interval to
4634 demonstrate periodic health checking. Add "nodh" option to
4635 fips_test_suite to skip very slow DH test.
4639 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4644 * More extensive health check for DRBG checking many more failure modes.
4645 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4646 combination: call this in fips_test_suite.
4650 * Add support for canonical generation of DSA parameter 'g'. See
4653 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4654 POST to handle HMAC cases.
4658 * Add functions FIPS_module_version() and FIPS_module_version_text()
4659 to return numerical and string versions of the FIPS module number.
4663 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4664 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4665 outside the validated module in the FIPS capable OpenSSL.
4669 * Minor change to DRBG entropy callback semantics. In some cases
4670 there is no multiple of the block length between min_len and
4671 max_len. Allow the callback to return more than max_len bytes
4672 of entropy but discard any extra: it is the callback's responsibility
4673 to ensure that the extra data discarded does not impact the
4674 requested amount of entropy.
4678 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4679 information in FIPS186-3, SP800-57 and SP800-131A.
4683 * CCM support via EVP. Interface is very similar to GCM case except we
4684 must supply all data in one chunk (i.e. no update, final) and the
4685 message length must be supplied if AAD is used. Add algorithm test
4690 * Initial version of POST overhaul. Add POST callback to allow the status
4691 of POST to be monitored and/or failures induced. Modify fips_test_suite
4692 to use callback. Always run all selftests even if one fails.
4696 * XTS support including algorithm test driver in the fips_gcmtest program.
4697 Note: this does increase the maximum key length from 32 to 64 bytes but
4698 there should be no binary compatibility issues as existing applications
4699 will never use XTS mode.
4703 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4704 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4705 performs algorithm blocking for unapproved PRNG types. Also do not
4706 set PRNG type in FIPS_mode_set(): leave this to the application.
4707 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4708 the standard OpenSSL PRNG: set additional data to a date time vector.
4712 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4713 This shouldn't present any incompatibility problems because applications
4714 shouldn't be using these directly and any that are will need to rethink
4715 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4719 * Extensive self tests and health checking required by SP800-90 DRBG.
4720 Remove strength parameter from FIPS_drbg_instantiate and always
4721 instantiate at maximum supported strength.
4725 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4729 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4733 * New function DH_compute_key_padded() to compute a DH key and pad with
4734 leading zeroes if needed: this complies with SP800-56A et al.
4738 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4739 anything, incomplete, subject to change and largely untested at present.
4743 * Modify fipscanisteronly build option to only build the necessary object
4744 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4748 * Add experimental option FIPSSYMS to give all symbols in
4749 fipscanister.o and FIPS or fips prefix. This will avoid
4750 conflicts with future versions of OpenSSL. Add perl script
4751 util/fipsas.pl to preprocess assembly language source files
4752 and rename any affected symbols.
4756 * Add selftest checks and algorithm block of non-fips algorithms in
4757 FIPS mode. Remove DES2 from selftests.
4761 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4762 return internal method without any ENGINE dependencies. Add new
4763 tiny fips sign and verify functions.
4767 * New build option no-ec2m to disable characteristic 2 code.
4771 * New build option "fipscanisteronly". This only builds fipscanister.o
4772 and (currently) associated fips utilities. Uses the file Makefile.fips
4773 instead of Makefile.org as the prototype.
4777 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4778 Update fips_gcmtest to use IV generator.
4782 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4783 setting output buffer to NULL. The `*Final` function must be
4784 called although it will not retrieve any additional data. The tag
4785 can be set or retrieved with a ctrl. The IV length is by default 12
4786 bytes (96 bits) but can be set to an alternative value. If the IV
4787 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4792 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4793 underlying do_cipher function handles all cipher semantics itself
4794 including padding and finalisation. This is useful if (for example)
4795 an ENGINE cipher handles block padding itself. The behaviour of
4796 do_cipher is subtly changed if this flag is set: the return value
4797 is the number of characters written to the output buffer (zero is
4798 no longer an error code) or a negative error code. Also if the
4799 input buffer is NULL and length 0 finalisation should be performed.
4803 * If a candidate issuer certificate is already part of the constructed
4804 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4808 * Improve forward-security support: add functions
4810 void SSL_CTX_set_not_resumable_session_callback(
4811 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4812 void SSL_set_not_resumable_session_callback(
4813 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4815 for use by SSL/TLS servers; the callback function will be called whenever a
4816 new session is created, and gets to decide whether the session may be
4817 cached to make it resumable (return 0) or not (return 1). (As by the
4818 SSL/TLS protocol specifications, the session_id sent by the server will be
4819 empty to indicate that the session is not resumable; also, the server will
4820 not generate RFC 4507 (RFC 5077) session tickets.)
4822 A simple reasonable callback implementation is to return is_forward_secure.
4823 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4824 by the SSL/TLS server library, indicating whether it can provide forward
4827 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4829 * New -verify_name option in command line utilities to set verification
4834 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4835 Add CMAC pkey methods.
4839 * Experimental renegotiation in s_server -www mode. If the client
4840 browses /reneg connection is renegotiated. If /renegcert it is
4841 renegotiated requesting a certificate.
4845 * Add an "external" session cache for debugging purposes to s_server. This
4846 should help trace issues which normally are only apparent in deployed
4847 multi-process servers.
4851 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4852 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4853 BIO_set_cipher() and some obscure PEM functions were changed so they
4854 can now return an error. The RAND changes required a change to the
4855 RAND_METHOD structure.
4859 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4860 a gcc attribute to warn if the result of a function is ignored. This
4861 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4862 whose return value is often ignored.
4866 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4867 These allow SCTs (signed certificate timestamps) to be requested and
4868 validated when establishing a connection.
4870 *Rob Percival <robpercival@google.com>*
4875 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4877 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4878 used even when parsing explicit parameters, when loading a encoded key
4879 or calling `EC_GROUP_new_from_ecpkparameters()`/
4880 `EC_GROUP_new_from_ecparameters()`.
4881 This prevents bypass of security hardening and performance gains,
4882 especially for curves with specialized EC_METHODs.
4883 By default, if a key encoded with explicit parameters is loaded and later
4884 encoded, the output is still encoded with explicit parameters, even if
4885 internally a "named" EC_GROUP is used for computation.
4889 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4890 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4891 NULL. After this change, only the cofactor parameter can be NULL. It also
4892 does some minimal sanity checks on the passed order.
4897 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4898 An attack is simple, if the first CMS_recipientInfo is valid but the
4899 second CMS_recipientInfo is chosen ciphertext. If the second
4900 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4901 encryption key will be replaced by garbage, and the message cannot be
4902 decoded, but if the RSA decryption fails, the correct encryption key is
4903 used and the recipient will not notice the attack.
4904 As a work around for this potential attack the length of the decrypted
4905 key must be equal to the cipher default key length, in case the
4906 certificate is not given and all recipientInfo are tried out.
4907 The old behaviour can be re-enabled in the CMS code by setting the
4908 CMS_DEBUG_DECRYPT flag.
4913 * Document issue with installation paths in diverse Windows builds
4915 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4916 binaries and run-time config file.
4921 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4923 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4924 This changes the size when using the `genpkey` command when no size is given.
4925 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4926 generation commands to use 2048 bits by default.
4930 * Add FIPS support for Android Arm 64-bit
4932 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4933 Module in Version 2.0.10. For some reason, the corresponding target
4934 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4935 built with FIPS support on Android Arm 64-bit. This omission has been
4938 *Matthias St. Pierre*
4940 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4942 * 0-byte record padding oracle
4944 If an application encounters a fatal protocol error and then calls
4945 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4946 then OpenSSL can respond differently to the calling application if a 0 byte
4947 record is received with invalid padding compared to if a 0 byte record is
4948 received with an invalid MAC. If the application then behaves differently
4949 based on that in a way that is detectable to the remote peer, then this
4950 amounts to a padding oracle that could be used to decrypt data.
4952 In order for this to be exploitable "non-stitched" ciphersuites must be in
4953 use. Stitched ciphersuites are optimised implementations of certain
4954 commonly used ciphersuites. Also the application must call SSL_shutdown()
4955 twice even if a protocol error has occurred (applications should not do
4956 this but some do anyway).
4958 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4959 Aviram, with additional investigation by Steven Collison and Andrew
4960 Hourselt. It was reported to OpenSSL on 10th December 2018.
4965 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4969 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4971 * Microarchitecture timing vulnerability in ECC scalar multiplication
4973 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4974 shown to be vulnerable to a microarchitecture timing side channel attack.
4975 An attacker with sufficient access to mount local timing attacks during
4976 ECDSA signature generation could recover the private key.
4978 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4979 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4985 * Timing vulnerability in DSA signature generation
4987 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4988 timing side channel attack. An attacker could use variations in the signing
4989 algorithm to recover the private key.
4991 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4996 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4997 Module, accidentally introduced while backporting security fixes from the
4998 development branch and hindering the use of ECC in FIPS mode.
5002 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5004 * Client DoS due to large DH parameter
5006 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5007 malicious server can send a very large prime value to the client. This will
5008 cause the client to spend an unreasonably long period of time generating a
5009 key for this prime resulting in a hang until the client has finished. This
5010 could be exploited in a Denial Of Service attack.
5012 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5017 * Cache timing vulnerability in RSA Key Generation
5019 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5020 a cache timing side channel attack. An attacker with sufficient access to
5021 mount cache timing attacks during the RSA key generation process could
5022 recover the private key.
5024 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5025 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5030 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5031 parameter is no longer accepted, as it leads to a corrupt table. NULL
5032 pem_str is reserved for alias entries only.
5036 * Revert blinding in ECDSA sign and instead make problematic addition
5037 length-invariant. Switch even to fixed-length Montgomery multiplication.
5041 * Change generating and checking of primes so that the error rate of not
5042 being prime depends on the intended use based on the size of the input.
5043 For larger primes this will result in more rounds of Miller-Rabin.
5044 The maximal error rate for primes with more than 1080 bits is lowered
5047 *Kurt Roeckx, Annie Yousar*
5049 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5053 * Add blinding to ECDSA and DSA signatures to protect against side channel
5054 attacks discovered by Keegan Ryan (NCC Group).
5058 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5059 now allow empty (zero character) pass phrases.
5063 * Certificate time validation (X509_cmp_time) enforces stricter
5064 compliance with RFC 5280. Fractional seconds and timezone offsets
5065 are no longer allowed.
5069 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5071 * Constructed ASN.1 types with a recursive definition could exceed the stack
5073 Constructed ASN.1 types with a recursive definition (such as can be found
5074 in PKCS7) could eventually exceed the stack given malicious input with
5075 excessive recursion. This could result in a Denial Of Service attack. There
5076 are no such structures used within SSL/TLS that come from untrusted sources
5077 so this is considered safe.
5079 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5085 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5087 * Read/write after SSL object in error state
5089 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5090 mechanism. The intent was that if a fatal error occurred during a handshake
5091 then OpenSSL would move into the error state and would immediately fail if
5092 you attempted to continue the handshake. This works as designed for the
5093 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5094 SSL_connect()), however due to a bug it does not work correctly if
5095 SSL_read() or SSL_write() is called directly. In that scenario, if the
5096 handshake fails then a fatal error will be returned in the initial function
5097 call. If SSL_read()/SSL_write() is subsequently called by the application
5098 for the same SSL object then it will succeed and the data is passed without
5099 being decrypted/encrypted directly from the SSL/TLS record layer.
5101 In order to exploit this issue an application bug would have to be present
5102 that resulted in a call to SSL_read()/SSL_write() being issued after having
5103 already received a fatal error.
5105 This issue was reported to OpenSSL by David Benjamin (Google).
5110 * rsaz_1024_mul_avx2 overflow bug on x86_64
5112 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5113 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5114 Analysis suggests that attacks against RSA and DSA as a result of this
5115 defect would be very difficult to perform and are not believed likely.
5116 Attacks against DH1024 are considered just feasible, because most of the
5117 work necessary to deduce information about a private key may be performed
5118 offline. The amount of resources required for such an attack would be
5119 significant. However, for an attack on TLS to be meaningful, the server
5120 would have to share the DH1024 private key among multiple clients, which is
5121 no longer an option since CVE-2016-0701.
5123 This only affects processors that support the AVX2 but not ADX extensions
5124 like Intel Haswell (4th generation).
5126 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5127 was originally found via the OSS-Fuzz project.
5132 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5134 * bn_sqrx8x_internal carry bug on x86_64
5136 There is a carry propagating bug in the x86_64 Montgomery squaring
5137 procedure. No EC algorithms are affected. Analysis suggests that attacks
5138 against RSA and DSA as a result of this defect would be very difficult to
5139 perform and are not believed likely. Attacks against DH are considered just
5140 feasible (although very difficult) because most of the work necessary to
5141 deduce information about a private key may be performed offline. The amount
5142 of resources required for such an attack would be very significant and
5143 likely only accessible to a limited number of attackers. An attacker would
5144 additionally need online access to an unpatched system using the target
5145 private key in a scenario with persistent DH parameters and a private
5146 key that is shared between multiple clients.
5148 This only affects processors that support the BMI1, BMI2 and ADX extensions
5149 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5151 This issue was reported to OpenSSL by the OSS-Fuzz project.
5156 * Malformed X.509 IPAddressFamily could cause OOB read
5158 If an X.509 certificate has a malformed IPAddressFamily extension,
5159 OpenSSL could do a one-byte buffer overread. The most likely result
5160 would be an erroneous display of the certificate in text format.
5162 This issue was reported to OpenSSL by the OSS-Fuzz project.
5166 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5168 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5169 platform rather than 'mingw'.
5173 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5175 * Truncated packet could crash via OOB read
5177 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5178 cipher is being used, then a truncated packet can cause that host to
5179 perform an out-of-bounds read, usually resulting in a crash.
5181 This issue was reported to OpenSSL by Robert Święcki of Google.
5186 * BN_mod_exp may produce incorrect results on x86_64
5188 There is a carry propagating bug in the x86_64 Montgomery squaring
5189 procedure. No EC algorithms are affected. Analysis suggests that attacks
5190 against RSA and DSA as a result of this defect would be very difficult to
5191 perform and are not believed likely. Attacks against DH are considered just
5192 feasible (although very difficult) because most of the work necessary to
5193 deduce information about a private key may be performed offline. The amount
5194 of resources required for such an attack would be very significant and
5195 likely only accessible to a limited number of attackers. An attacker would
5196 additionally need online access to an unpatched system using the target
5197 private key in a scenario with persistent DH parameters and a private
5198 key that is shared between multiple clients. For example this can occur by
5199 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5200 similar to CVE-2015-3193 but must be treated as a separate problem.
5202 This issue was reported to OpenSSL by the OSS-Fuzz project.
5207 * Montgomery multiplication may produce incorrect results
5209 There is a carry propagating bug in the Broadwell-specific Montgomery
5210 multiplication procedure that handles input lengths divisible by, but
5211 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5212 and DH private keys are impossible. This is because the subroutine in
5213 question is not used in operations with the private key itself and an input
5214 of the attacker's direct choice. Otherwise the bug can manifest itself as
5215 transient authentication and key negotiation failures or reproducible
5216 erroneous outcome of public-key operations with specially crafted input.
5217 Among EC algorithms only Brainpool P-512 curves are affected and one
5218 presumably can attack ECDH key negotiation. Impact was not analyzed in
5219 detail, because pre-requisites for attack are considered unlikely. Namely
5220 multiple clients have to choose the curve in question and the server has to
5221 share the private key among them, neither of which is default behaviour.
5222 Even then only clients that chose the curve will be affected.
5224 This issue was publicly reported as transient failures and was not
5225 initially recognized as a security issue. Thanks to Richard Morgan for
5226 providing reproducible case.
5231 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5232 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5233 prevent issues where no progress is being made and the peer continually
5234 sends unrecognised record types, using up resources processing them.
5238 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5240 * Missing CRL sanity check
5242 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5243 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5244 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5246 This issue only affects the OpenSSL 1.0.2i
5251 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5253 * OCSP Status Request extension unbounded memory growth
5255 A malicious client can send an excessively large OCSP Status Request
5256 extension. If that client continually requests renegotiation, sending a
5257 large OCSP Status Request extension each time, then there will be unbounded
5258 memory growth on the server. This will eventually lead to a Denial Of
5259 Service attack through memory exhaustion. Servers with a default
5260 configuration are vulnerable even if they do not support OCSP. Builds using
5261 the "no-ocsp" build time option are not affected.
5263 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5268 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5271 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5277 * OOB write in MDC2_Update()
5279 An overflow can occur in MDC2_Update() either if called directly or
5280 through the EVP_DigestUpdate() function using MDC2. If an attacker
5281 is able to supply very large amounts of input data after a previous
5282 call to EVP_EncryptUpdate() with a partial block then a length check
5283 can overflow resulting in a heap corruption.
5285 The amount of data needed is comparable to SIZE_MAX which is impractical
5288 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5293 * Malformed SHA512 ticket DoS
5295 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5296 DoS attack where a malformed ticket will result in an OOB read which will
5299 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5300 a custom server callback and ticket lookup mechanism.
5302 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5307 * OOB write in BN_bn2dec()
5309 The function BN_bn2dec() does not check the return value of BN_div_word().
5310 This can cause an OOB write if an application uses this function with an
5311 overly large BIGNUM. This could be a problem if an overly large certificate
5312 or CRL is printed out from an untrusted source. TLS is not affected because
5313 record limits will reject an oversized certificate before it is parsed.
5315 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5320 * OOB read in TS_OBJ_print_bio()
5322 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5323 the total length the OID text representation would use and not the amount
5324 of data written. This will result in OOB reads when large OIDs are
5327 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5332 * Pointer arithmetic undefined behaviour
5334 Avoid some undefined pointer arithmetic
5336 A common idiom in the codebase is to check limits in the following manner:
5339 Where "p" points to some malloc'd data of SIZE bytes and
5342 "len" here could be from some externally supplied data (e.g. from a TLS
5345 The rules of C pointer arithmetic are such that "p + len" is only well
5346 defined where len <= SIZE. Therefore the above idiom is actually
5347 undefined behaviour.
5349 For example this could cause problems if some malloc implementation
5350 provides an address for "p" such that "p + len" actually overflows for
5351 values of len that are too big and therefore p + len < limit.
5353 This issue was reported to OpenSSL by Guido Vranken
5358 * Constant time flag not preserved in DSA signing
5360 Operations in the DSA signing algorithm should run in constant time in
5361 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5362 implementation means that a non-constant time codepath is followed for
5363 certain operations. This has been demonstrated through a cache-timing
5364 attack to be sufficient for an attacker to recover the private DSA key.
5366 This issue was reported by César Pereida (Aalto University), Billy Brumley
5367 (Tampere University of Technology), and Yuval Yarom (The University of
5368 Adelaide and NICTA).
5373 * DTLS buffered message DoS
5375 In a DTLS connection where handshake messages are delivered out-of-order
5376 those messages that OpenSSL is not yet ready to process will be buffered
5377 for later use. Under certain circumstances, a flaw in the logic means that
5378 those messages do not get removed from the buffer even though the handshake
5379 has been completed. An attacker could force up to approx. 15 messages to
5380 remain in the buffer when they are no longer required. These messages will
5381 be cleared when the DTLS connection is closed. The default maximum size for
5382 a message is 100k. Therefore, the attacker could force an additional 1500k
5383 to be consumed per connection. By opening many simultaneous connections an
5384 attacker could cause a DoS attack through memory exhaustion.
5386 This issue was reported to OpenSSL by Quan Luo.
5391 * DTLS replay protection DoS
5393 A flaw in the DTLS replay attack protection mechanism means that records
5394 that arrive for future epochs update the replay protection "window" before
5395 the MAC for the record has been validated. This could be exploited by an
5396 attacker by sending a record for the next epoch (which does not have to
5397 decrypt or have a valid MAC), with a very large sequence number. This means
5398 that all subsequent legitimate packets are dropped causing a denial of
5399 service for a specific DTLS connection.
5401 This issue was reported to OpenSSL by the OCAP audit team.
5406 * Certificate message OOB reads
5408 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5409 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5410 theoretical DoS risk but this has not been observed in practice on common
5413 The messages affected are client certificate, client certificate request
5414 and server certificate. As a result the attack can only be performed
5415 against a client or a server which enables client authentication.
5417 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5422 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5424 * Prevent padding oracle in AES-NI CBC MAC check
5426 A MITM attacker can use a padding oracle attack to decrypt traffic
5427 when the connection uses an AES CBC cipher and the server support
5430 This issue was introduced as part of the fix for Lucky 13 padding
5431 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5432 constant time by making sure that always the same bytes are read and
5433 compared against either the MAC or padding bytes. But it no longer
5434 checked that there was enough data to have both the MAC and padding
5437 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5441 * Fix EVP_EncodeUpdate overflow
5443 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5444 Base64 encoding of binary data. If an attacker is able to supply very large
5445 amounts of input data then a length check can overflow resulting in a heap
5448 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5449 the `PEM_write_bio*` family of functions. These are mainly used within the
5450 OpenSSL command line applications, so any application which processes data
5451 from an untrusted source and outputs it as a PEM file should be considered
5452 vulnerable to this issue. User applications that call these APIs directly
5453 with large amounts of untrusted data may also be vulnerable.
5455 This issue was reported by Guido Vranken.
5460 * Fix EVP_EncryptUpdate overflow
5462 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5463 is able to supply very large amounts of input data after a previous call to
5464 EVP_EncryptUpdate() with a partial block then a length check can overflow
5465 resulting in a heap corruption. Following an analysis of all OpenSSL
5466 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5467 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5468 the first called function after an EVP_EncryptInit(), and therefore that
5469 specific call must be safe. The second form is where the length passed to
5470 EVP_EncryptUpdate() can be seen from the code to be some small value and
5471 therefore there is no possibility of an overflow. Since all instances are
5472 one of these two forms, it is believed that there can be no overflows in
5473 internal code due to this problem. It should be noted that
5474 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5475 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5476 of these calls have also been analysed too and it is believed there are no
5477 instances in internal usage where an overflow could occur.
5479 This issue was reported by Guido Vranken.
5484 * Prevent ASN.1 BIO excessive memory allocation
5486 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5487 a short invalid encoding can cause allocation of large amounts of memory
5488 potentially consuming excessive resources or exhausting memory.
5490 Any application parsing untrusted data through d2i BIO functions is
5491 affected. The memory based functions such as d2i_X509() are *not* affected.
5492 Since the memory based functions are used by the TLS library, TLS
5493 applications are not affected.
5495 This issue was reported by Brian Carpenter.
5502 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5503 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5504 in arbitrary stack data being returned in the buffer.
5506 This issue was reported by Guido Vranken.
5511 * Modify behavior of ALPN to invoke callback after SNI/servername
5512 callback, such that updates to the SSL_CTX affect ALPN.
5516 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5521 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5522 methods are enabled and ssl2 is disabled the methods return NULL.
5526 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5528 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5529 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5530 provide any "EXPORT" or "LOW" strength ciphers.
5534 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5535 is by default disabled at build-time. Builds that are not configured with
5536 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5537 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5538 will need to explicitly call either of:
5540 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5542 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5544 as appropriate. Even if either of those is used, or the application
5545 explicitly uses the version-specific SSLv2_method() or its client and
5546 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5547 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5548 ciphers, and SSLv2 56-bit DES are no longer available.
5553 * Fix a double-free in DSA code
5555 A double free bug was discovered when OpenSSL parses malformed DSA private
5556 keys and could lead to a DoS attack or memory corruption for applications
5557 that receive DSA private keys from untrusted sources. This scenario is
5560 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5566 * Disable SRP fake user seed to address a server memory leak.
5568 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5570 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5571 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5572 was changed to ignore the "fake user" SRP seed, even if the seed
5575 Users should use SRP_VBASE_get1_by_user instead. Note that in
5576 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5577 also that even though configuring the SRP seed attempts to hide
5578 invalid usernames by continuing the handshake with fake
5579 credentials, this behaviour is not constant time and no strong
5580 guarantees are made that the handshake is indistinguishable from
5581 that of a valid user.
5586 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5588 In the BN_hex2bn function the number of hex digits is calculated using an
5589 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5590 large values of `i` this can result in `bn_expand` not allocating any
5591 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5592 field as NULL leading to a subsequent NULL ptr deref. For very large values
5593 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5594 In this case memory is allocated to the internal BIGNUM data field, but it
5595 is insufficiently sized leading to heap corruption. A similar issue exists
5596 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5597 is ever called by user applications with very large untrusted hex/dec data.
5598 This is anticipated to be a rare occurrence.
5600 All OpenSSL internal usage of these functions use data that is not expected
5601 to be untrusted, e.g. config file data or application command line
5602 arguments. If user developed applications generate config file data based
5603 on untrusted data then it is possible that this could also lead to security
5604 consequences. This is also anticipated to be rare.
5606 This issue was reported to OpenSSL by Guido Vranken.
5611 * Fix memory issues in `BIO_*printf` functions
5613 The internal `fmtstr` function used in processing a "%s" format string in
5614 the `BIO_*printf` functions could overflow while calculating the length of a
5615 string and cause an OOB read when printing very long strings.
5617 Additionally the internal `doapr_outch` function can attempt to write to an
5618 OOB memory location (at an offset from the NULL pointer) in the event of a
5619 memory allocation failure. In 1.0.2 and below this could be caused where
5620 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5621 could be in processing a very long "%s" format string. Memory leaks can
5624 The first issue may mask the second issue dependent on compiler behaviour.
5625 These problems could enable attacks where large amounts of untrusted data
5626 is passed to the `BIO_*printf` functions. If applications use these functions
5627 in this way then they could be vulnerable. OpenSSL itself uses these
5628 functions when printing out human-readable dumps of ASN.1 data. Therefore
5629 applications that print this data could be vulnerable if the data is from
5630 untrusted sources. OpenSSL command line applications could also be
5631 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5632 as command line arguments.
5634 Libssl is not considered directly vulnerable. Additionally certificates etc
5635 received via remote connections via libssl are also unlikely to be able to
5636 trigger these issues because of message size limits enforced within libssl.
5638 This issue was reported to OpenSSL Guido Vranken.
5643 * Side channel attack on modular exponentiation
5645 A side-channel attack was found which makes use of cache-bank conflicts on
5646 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5647 of RSA keys. The ability to exploit this issue is limited as it relies on
5648 an attacker who has control of code in a thread running on the same
5649 hyper-threaded core as the victim thread which is performing decryptions.
5651 This issue was reported to OpenSSL by Yuval Yarom, The University of
5652 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5653 Nadia Heninger, University of Pennsylvania with more information at
5654 <http://cachebleed.info>.
5659 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5660 if no keysize is specified with default_bits. This fixes an
5661 omission in an earlier change that changed all RSA/DSA key generation
5662 commands to use 2048 bits by default.
5666 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5668 * DH small subgroups
5670 Historically OpenSSL only ever generated DH parameters based on "safe"
5671 primes. More recently (in version 1.0.2) support was provided for
5672 generating X9.42 style parameter files such as those required for RFC 5114
5673 support. The primes used in such files may not be "safe". Where an
5674 application is using DH configured with parameters based on primes that are
5675 not "safe" then an attacker could use this fact to find a peer's private
5676 DH exponent. This attack requires that the attacker complete multiple
5677 handshakes in which the peer uses the same private DH exponent. For example
5678 this could be used to discover a TLS server's private DH exponent if it's
5679 reusing the private DH exponent or it's using a static DH ciphersuite.
5681 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5682 TLS. It is not on by default. If the option is not set then the server
5683 reuses the same private DH exponent for the life of the server process and
5684 would be vulnerable to this attack. It is believed that many popular
5685 applications do set this option and would therefore not be at risk.
5687 The fix for this issue adds an additional check where a "q" parameter is
5688 available (as is the case in X9.42 based parameters). This detects the
5689 only known attack, and is the only possible defense for static DH
5690 ciphersuites. This could have some performance impact.
5692 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5693 default and cannot be disabled. This could have some performance impact.
5695 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5700 * SSLv2 doesn't block disabled ciphers
5702 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5703 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5704 been disabled, provided that the SSLv2 protocol was not also disabled via
5707 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5708 and Sebastian Schinzel.
5713 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5715 * BN_mod_exp may produce incorrect results on x86_64
5717 There is a carry propagating bug in the x86_64 Montgomery squaring
5718 procedure. No EC algorithms are affected. Analysis suggests that attacks
5719 against RSA and DSA as a result of this defect would be very difficult to
5720 perform and are not believed likely. Attacks against DH are considered just
5721 feasible (although very difficult) because most of the work necessary to
5722 deduce information about a private key may be performed offline. The amount
5723 of resources required for such an attack would be very significant and
5724 likely only accessible to a limited number of attackers. An attacker would
5725 additionally need online access to an unpatched system using the target
5726 private key in a scenario with persistent DH parameters and a private
5727 key that is shared between multiple clients. For example this can occur by
5728 default in OpenSSL DHE based SSL/TLS ciphersuites.
5730 This issue was reported to OpenSSL by Hanno Böck.
5735 * Certificate verify crash with missing PSS parameter
5737 The signature verification routines will crash with a NULL pointer
5738 dereference if presented with an ASN.1 signature using the RSA PSS
5739 algorithm and absent mask generation function parameter. Since these
5740 routines are used to verify certificate signature algorithms this can be
5741 used to crash any certificate verification operation and exploited in a
5742 DoS attack. Any application which performs certificate verification is
5743 vulnerable including OpenSSL clients and servers which enable client
5746 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5751 * X509_ATTRIBUTE memory leak
5753 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5754 memory. This structure is used by the PKCS#7 and CMS routines so any
5755 application which reads PKCS#7 or CMS data from untrusted sources is
5756 affected. SSL/TLS is not affected.
5758 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5764 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5765 This changes the decoding behaviour for some invalid messages,
5766 though the change is mostly in the more lenient direction, and
5767 legacy behaviour is preserved as much as possible.
5771 * In DSA_generate_parameters_ex, if the provided seed is too short,
5774 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5776 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5778 * Alternate chains certificate forgery
5780 During certificate verification, OpenSSL will attempt to find an
5781 alternative certificate chain if the first attempt to build such a chain
5782 fails. An error in the implementation of this logic can mean that an
5783 attacker could cause certain checks on untrusted certificates to be
5784 bypassed, such as the CA flag, enabling them to use a valid leaf
5785 certificate to act as a CA and "issue" an invalid certificate.
5787 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5792 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5794 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5795 incompatibility in the handling of HMAC. The previous ABI has now been
5800 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5802 * Malformed ECParameters causes infinite loop
5804 When processing an ECParameters structure OpenSSL enters an infinite loop
5805 if the curve specified is over a specially malformed binary polynomial
5808 This can be used to perform denial of service against any
5809 system which processes public keys, certificate requests or
5810 certificates. This includes TLS clients and TLS servers with
5811 client authentication enabled.
5813 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5818 * Exploitable out-of-bounds read in X509_cmp_time
5820 X509_cmp_time does not properly check the length of the ASN1_TIME
5821 string and can read a few bytes out of bounds. In addition,
5822 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5825 An attacker can use this to craft malformed certificates and CRLs of
5826 various sizes and potentially cause a segmentation fault, resulting in
5827 a DoS on applications that verify certificates or CRLs. TLS clients
5828 that verify CRLs are affected. TLS clients and servers with client
5829 authentication enabled may be affected if they use custom verification
5832 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5833 independently by Hanno Böck.
5838 * PKCS7 crash with missing EnvelopedContent
5840 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5841 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5842 with missing content and trigger a NULL pointer dereference on parsing.
5844 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5845 structures from untrusted sources are affected. OpenSSL clients and
5846 servers are not affected.
5848 This issue was reported to OpenSSL by Michal Zalewski (Google).
5853 * CMS verify infinite loop with unknown hash function
5855 When verifying a signedData message the CMS code can enter an infinite loop
5856 if presented with an unknown hash function OID. This can be used to perform
5857 denial of service against any system which verifies signedData messages using
5859 This issue was reported to OpenSSL by Johannes Bauer.
5864 * Race condition handling NewSessionTicket
5866 If a NewSessionTicket is received by a multi-threaded client when attempting to
5867 reuse a previous ticket then a race condition can occur potentially leading to
5868 a double free of the ticket data.
5873 * Only support 256-bit or stronger elliptic curves with the
5874 'ecdh_auto' setting (server) or by default (client). Of supported
5875 curves, prefer P-256 (both).
5879 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5881 * ClientHello sigalgs DoS fix
5883 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5884 invalid signature algorithms extension a NULL pointer dereference will
5885 occur. This can be exploited in a DoS attack against the server.
5887 This issue was was reported to OpenSSL by David Ramos of Stanford
5891 *Stephen Henson and Matt Caswell*
5893 * Multiblock corrupted pointer fix
5895 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5896 feature only applies on 64 bit x86 architecture platforms that support AES
5897 NI instructions. A defect in the implementation of "multiblock" can cause
5898 OpenSSL's internal write buffer to become incorrectly set to NULL when
5899 using non-blocking IO. Typically, when the user application is using a
5900 socket BIO for writing, this will only result in a failed connection.
5901 However if some other BIO is used then it is likely that a segmentation
5902 fault will be triggered, thus enabling a potential DoS attack.
5904 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5909 * Segmentation fault in DTLSv1_listen fix
5911 The DTLSv1_listen function is intended to be stateless and processes the
5912 initial ClientHello from many peers. It is common for user code to loop
5913 over the call to DTLSv1_listen until a valid ClientHello is received with
5914 an associated cookie. A defect in the implementation of DTLSv1_listen means
5915 that state is preserved in the SSL object from one invocation to the next
5916 that can lead to a segmentation fault. Errors processing the initial
5917 ClientHello can trigger this scenario. An example of such an error could be
5918 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5921 This issue was reported to OpenSSL by Per Allansson.
5926 * Segmentation fault in ASN1_TYPE_cmp fix
5928 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5929 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5930 certificate signature algorithm consistency this can be used to crash any
5931 certificate verification operation and exploited in a DoS attack. Any
5932 application which performs certificate verification is vulnerable including
5933 OpenSSL clients and servers which enable client authentication.
5938 * Segmentation fault for invalid PSS parameters fix
5940 The signature verification routines will crash with a NULL pointer
5941 dereference if presented with an ASN.1 signature using the RSA PSS
5942 algorithm and invalid parameters. Since these routines are used to verify
5943 certificate signature algorithms this can be used to crash any
5944 certificate verification operation and exploited in a DoS attack. Any
5945 application which performs certificate verification is vulnerable including
5946 OpenSSL clients and servers which enable client authentication.
5948 This issue was was reported to OpenSSL by Brian Carpenter.
5953 * ASN.1 structure reuse memory corruption fix
5955 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5956 memory corruption via an invalid write. Such reuse is and has been
5957 strongly discouraged and is believed to be rare.
5959 Applications that parse structures containing CHOICE or ANY DEFINED BY
5960 components may be affected. Certificate parsing (d2i_X509 and related
5961 functions) are however not affected. OpenSSL clients and servers are
5967 * PKCS7 NULL pointer dereferences fix
5969 The PKCS#7 parsing code does not handle missing outer ContentInfo
5970 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5971 missing content and trigger a NULL pointer dereference on parsing.
5973 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5974 otherwise parse PKCS#7 structures from untrusted sources are
5975 affected. OpenSSL clients and servers are not affected.
5977 This issue was reported to OpenSSL by Michal Zalewski (Google).
5982 * DoS via reachable assert in SSLv2 servers fix
5984 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5985 servers that both support SSLv2 and enable export cipher suites by sending
5986 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5988 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5989 (OpenSSL development team).
5994 * Empty CKE with client auth and DHE fix
5996 If client auth is used then a server can seg fault in the event of a DHE
5997 ciphersuite being selected and a zero length ClientKeyExchange message
5998 being sent by the client. This could be exploited in a DoS attack.
6003 * Handshake with unseeded PRNG fix
6005 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6006 with an unseeded PRNG. The conditions are:
6007 - The client is on a platform where the PRNG has not been seeded
6008 automatically, and the user has not seeded manually
6009 - A protocol specific client method version has been used (i.e. not
6010 SSL_client_methodv23)
6011 - A ciphersuite is used that does not require additional random data from
6012 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6014 If the handshake succeeds then the client random that has been used will
6015 have been generated from a PRNG with insufficient entropy and therefore the
6016 output may be predictable.
6018 For example using the following command with an unseeded openssl will
6019 succeed on an unpatched platform:
6021 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6026 * Use After Free following d2i_ECPrivatekey error fix
6028 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6029 could cause a use after free condition. This, in turn, could cause a double
6030 free in several private key parsing functions (such as d2i_PrivateKey
6031 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6032 for applications that receive EC private keys from untrusted
6033 sources. This scenario is considered rare.
6035 This issue was discovered by the BoringSSL project and fixed in their
6041 * X509_to_X509_REQ NULL pointer deref fix
6043 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6044 the certificate key is invalid. This function is rarely used in practice.
6046 This issue was discovered by Brian Carpenter.
6051 * Removed the export ciphers from the DEFAULT ciphers
6055 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6057 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6058 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6059 So far those who have to target multiple platforms would compromise
6060 and argue that binary targeting say ARMv5 would still execute on
6061 ARMv8. "Universal" build resolves this compromise by providing
6062 near-optimal performance even on newer platforms.
6066 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6067 (other platforms pending).
6069 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6071 * Add support for the SignedCertificateTimestampList certificate and
6072 OCSP response extensions from RFC6962.
6076 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6077 for corner cases. (Certain input points at infinity could lead to
6078 bogus results, with non-infinity inputs mapped to infinity too.)
6082 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6083 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6084 common cases are optimized and there still is room for further
6085 improvements. Vector Permutation AES for Altivec is also added.
6089 * Add support for little-endian ppc64 Linux target.
6091 *Marcelo Cerri (IBM)*
6093 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6094 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6095 are optimized and there still is room for further improvements.
6096 Both 32- and 64-bit modes are supported.
6098 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6100 * Improved ARMv7 NEON support.
6104 * Support for SPARC Architecture 2011 crypto extensions, first
6105 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6106 SHA256/512, MD5, GHASH and modular exponentiation.
6108 *Andy Polyakov, David Miller*
6110 * Accelerated modular exponentiation for Intel processors, a.k.a.
6113 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6115 * Support for new and upcoming Intel processors, including AVX2,
6116 BMI and SHA ISA extensions. This includes additional "stitched"
6117 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6120 This work was sponsored by Intel Corp.
6124 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6125 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6126 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6130 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6131 this fixes a limitation in previous versions of OpenSSL.
6135 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6136 MGF1 digest and OAEP label.
6140 * Add EVP support for key wrapping algorithms, to avoid problems with
6141 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6142 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6143 algorithms and include tests cases.
6147 * Add functions to allocate and set the fields of an ECDSA_METHOD
6150 *Douglas E. Engert, Steve Henson*
6152 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6153 difference in days and seconds between two tm or ASN1_TIME structures.
6157 * Add -rev test option to s_server to just reverse order of characters
6158 received by client and send back to server. Also prints an abbreviated
6159 summary of the connection parameters.
6163 * New option -brief for s_client and s_server to print out a brief summary
6164 of connection parameters.
6168 * Add callbacks for arbitrary TLS extensions.
6170 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6172 * New option -crl_download in several openssl utilities to download CRLs
6173 from CRLDP extension in certificates.
6177 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6181 * New function X509_CRL_diff to generate a delta CRL from the difference
6182 of two full CRLs. Add support to "crl" utility.
6186 * New functions to set lookup_crls function and to retrieve
6187 X509_STORE from X509_STORE_CTX.
6191 * Print out deprecated issuer and subject unique ID fields in
6196 * Extend OCSP I/O functions so they can be used for simple general purpose
6197 HTTP as well as OCSP. New wrapper function which can be used to download
6198 CRLs using the OCSP API.
6202 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6206 * `SSL_CONF*` functions. These provide a common framework for application
6207 configuration using configuration files or command lines.
6211 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6212 message callback and prints the results. Needs compile time option
6213 "enable-ssl-trace". New options to s_client and s_server to enable
6218 * New ctrl and macro to retrieve supported points extensions.
6219 Print out extension in s_server and s_client.
6223 * New functions to retrieve certificate signature and signature
6228 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6233 * New Suite B modes for TLS code. These use and enforce the requirements
6234 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6235 only use Suite B curves. The Suite B modes can be set by using the
6236 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6240 * New chain verification flags for Suite B levels of security. Check
6241 algorithms are acceptable when flags are set in X509_verify_cert.
6245 * Make tls1_check_chain return a set of flags indicating checks passed
6246 by a certificate chain. Add additional tests to handle client
6247 certificates: checks for matching certificate type and issuer name
6252 * If an attempt is made to use a signature algorithm not in the peer
6253 preference list abort the handshake. If client has no suitable
6254 signature algorithms in response to a certificate request do not
6255 use the certificate.
6259 * If server EC tmp key is not in client preference list abort handshake.
6263 * Add support for certificate stores in CERT structure. This makes it
6264 possible to have different stores per SSL structure or one store in
6265 the parent SSL_CTX. Include distinct stores for certificate chain
6266 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6267 to build and store a certificate chain in CERT structure: returning
6268 an error if the chain cannot be built: this will allow applications
6269 to test if a chain is correctly configured.
6271 Note: if the CERT based stores are not set then the parent SSL_CTX
6272 store is used to retain compatibility with existing behaviour.
6276 * New function ssl_set_client_disabled to set a ciphersuite disabled
6277 mask based on the current session, check mask when sending client
6278 hello and checking the requested ciphersuite.
6282 * New ctrls to retrieve and set certificate types in a certificate
6283 request message. Print out received values in s_client. If certificate
6284 types is not set with custom values set sensible values based on
6285 supported signature algorithms.
6289 * Support for distinct client and server supported signature algorithms.
6293 * Add certificate callback. If set this is called whenever a certificate
6294 is required by client or server. An application can decide which
6295 certificate chain to present based on arbitrary criteria: for example
6296 supported signature algorithms. Add very simple example to s_server.
6297 This fixes many of the problems and restrictions of the existing client
6298 certificate callback: for example you can now clear an existing
6299 certificate and specify the whole chain.
6303 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6304 the certificate can be used for (if anything). Set valid_flags field
6305 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6306 to have similar checks in it.
6308 Add new "cert_flags" field to CERT structure and include a "strict mode".
6309 This enforces some TLS certificate requirements (such as only permitting
6310 certificate signature algorithms contained in the supported algorithms
6311 extension) which some implementations ignore: this option should be used
6312 with caution as it could cause interoperability issues.
6316 * Update and tidy signature algorithm extension processing. Work out
6317 shared signature algorithms based on preferences and peer algorithms
6318 and print them out in s_client and s_server. Abort handshake if no
6319 shared signature algorithms.
6323 * Add new functions to allow customised supported signature algorithms
6324 for SSL and SSL_CTX structures. Add options to s_client and s_server
6329 * New function SSL_certs_clear() to delete all references to certificates
6330 from an SSL structure. Before this once a certificate had been added
6331 it couldn't be removed.
6335 * Integrate hostname, email address and IP address checking with certificate
6336 verification. New verify options supporting checking in openssl utility.
6340 * Fixes and wildcard matching support to hostname and email checking
6341 functions. Add manual page.
6343 *Florian Weimer (Red Hat Product Security Team)*
6345 * New functions to check a hostname email or IP address against a
6346 certificate. Add options x509 utility to print results of checks against
6351 * Fix OCSP checking.
6353 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6355 * Initial experimental support for explicitly trusted non-root CAs.
6356 OpenSSL still tries to build a complete chain to a root but if an
6357 intermediate CA has a trust setting included that is used. The first
6358 setting is used: whether to trust (e.g., -addtrust option to the x509
6363 * Add -trusted_first option which attempts to find certificates in the
6364 trusted store even if an untrusted chain is also supplied.
6368 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6369 platform support for Linux and Android.
6373 * Support for linux-x32, ILP32 environment in x86_64 framework.
6377 * Experimental multi-implementation support for FIPS capable OpenSSL.
6378 When in FIPS mode the approved implementations are used as normal,
6379 when not in FIPS mode the internal unapproved versions are used instead.
6380 This means that the FIPS capable OpenSSL isn't forced to use the
6381 (often lower performance) FIPS implementations outside FIPS mode.
6385 * Transparently support X9.42 DH parameters when calling
6386 PEM_read_bio_DHparameters. This means existing applications can handle
6387 the new parameter format automatically.
6391 * Initial experimental support for X9.42 DH parameter format: mainly
6392 to support use of 'q' parameter for RFC5114 parameters.
6396 * Add DH parameters from RFC5114 including test data to dhtest.
6400 * Support for automatic EC temporary key parameter selection. If enabled
6401 the most preferred EC parameters are automatically used instead of
6402 hardcoded fixed parameters. Now a server just has to call:
6403 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6404 support ECDH and use the most appropriate parameters.
6408 * Enhance and tidy EC curve and point format TLS extension code. Use
6409 static structures instead of allocation if default values are used.
6410 New ctrls to set curves we wish to support and to retrieve shared curves.
6411 Print out shared curves in s_server. New options to s_server and s_client
6412 to set list of supported curves.
6416 * New ctrls to retrieve supported signature algorithms and
6417 supported curve values as an array of NIDs. Extend openssl utility
6418 to print out received values.
6422 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6423 between NIDs and the more common NIST names such as "P-256". Enhance
6424 ecparam utility and ECC method to recognise the NIST names for curves.
6428 * Enhance SSL/TLS certificate chain handling to support different
6429 chains for each certificate instead of one chain in the parent SSL_CTX.
6433 * Support for fixed DH ciphersuite client authentication: where both
6434 server and client use DH certificates with common parameters.
6438 * Support for fixed DH ciphersuites: those requiring DH server
6443 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6445 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6446 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6447 X509_CINF_get_signature were reverted post internal team review.
6452 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6454 * OCSP Status Request extension unbounded memory growth
6456 A malicious client can send an excessively large OCSP Status Request
6457 extension. If that client continually requests renegotiation, sending a
6458 large OCSP Status Request extension each time, then there will be unbounded
6459 memory growth on the server. This will eventually lead to a Denial Of
6460 Service attack through memory exhaustion. Servers with a default
6461 configuration are vulnerable even if they do not support OCSP. Builds using
6462 the "no-ocsp" build time option are not affected.
6464 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6469 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6472 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6478 * OOB write in MDC2_Update()
6480 An overflow can occur in MDC2_Update() either if called directly or
6481 through the EVP_DigestUpdate() function using MDC2. If an attacker
6482 is able to supply very large amounts of input data after a previous
6483 call to EVP_EncryptUpdate() with a partial block then a length check
6484 can overflow resulting in a heap corruption.
6486 The amount of data needed is comparable to SIZE_MAX which is impractical
6489 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6494 * Malformed SHA512 ticket DoS
6496 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6497 DoS attack where a malformed ticket will result in an OOB read which will
6500 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6501 a custom server callback and ticket lookup mechanism.
6503 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6508 * OOB write in BN_bn2dec()
6510 The function BN_bn2dec() does not check the return value of BN_div_word().
6511 This can cause an OOB write if an application uses this function with an
6512 overly large BIGNUM. This could be a problem if an overly large certificate
6513 or CRL is printed out from an untrusted source. TLS is not affected because
6514 record limits will reject an oversized certificate before it is parsed.
6516 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6521 * OOB read in TS_OBJ_print_bio()
6523 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6524 the total length the OID text representation would use and not the amount
6525 of data written. This will result in OOB reads when large OIDs are
6528 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6533 * Pointer arithmetic undefined behaviour
6535 Avoid some undefined pointer arithmetic
6537 A common idiom in the codebase is to check limits in the following manner:
6540 Where "p" points to some malloc'd data of SIZE bytes and
6543 "len" here could be from some externally supplied data (e.g. from a TLS
6546 The rules of C pointer arithmetic are such that "p + len" is only well
6547 defined where len <= SIZE. Therefore, the above idiom is actually
6548 undefined behaviour.
6550 For example this could cause problems if some malloc implementation
6551 provides an address for "p" such that "p + len" actually overflows for
6552 values of len that are too big and therefore p + len < limit.
6554 This issue was reported to OpenSSL by Guido Vranken
6559 * Constant time flag not preserved in DSA signing
6561 Operations in the DSA signing algorithm should run in constant time in
6562 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6563 implementation means that a non-constant time codepath is followed for
6564 certain operations. This has been demonstrated through a cache-timing
6565 attack to be sufficient for an attacker to recover the private DSA key.
6567 This issue was reported by César Pereida (Aalto University), Billy Brumley
6568 (Tampere University of Technology), and Yuval Yarom (The University of
6569 Adelaide and NICTA).
6574 * DTLS buffered message DoS
6576 In a DTLS connection where handshake messages are delivered out-of-order
6577 those messages that OpenSSL is not yet ready to process will be buffered
6578 for later use. Under certain circumstances, a flaw in the logic means that
6579 those messages do not get removed from the buffer even though the handshake
6580 has been completed. An attacker could force up to approx. 15 messages to
6581 remain in the buffer when they are no longer required. These messages will
6582 be cleared when the DTLS connection is closed. The default maximum size for
6583 a message is 100k. Therefore, the attacker could force an additional 1500k
6584 to be consumed per connection. By opening many simultaneous connections an
6585 attacker could cause a DoS attack through memory exhaustion.
6587 This issue was reported to OpenSSL by Quan Luo.
6592 * DTLS replay protection DoS
6594 A flaw in the DTLS replay attack protection mechanism means that records
6595 that arrive for future epochs update the replay protection "window" before
6596 the MAC for the record has been validated. This could be exploited by an
6597 attacker by sending a record for the next epoch (which does not have to
6598 decrypt or have a valid MAC), with a very large sequence number. This means
6599 that all subsequent legitimate packets are dropped causing a denial of
6600 service for a specific DTLS connection.
6602 This issue was reported to OpenSSL by the OCAP audit team.
6607 * Certificate message OOB reads
6609 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6610 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6611 theoretical DoS risk but this has not been observed in practice on common
6614 The messages affected are client certificate, client certificate request
6615 and server certificate. As a result the attack can only be performed
6616 against a client or a server which enables client authentication.
6618 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6623 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6625 * Prevent padding oracle in AES-NI CBC MAC check
6627 A MITM attacker can use a padding oracle attack to decrypt traffic
6628 when the connection uses an AES CBC cipher and the server support
6631 This issue was introduced as part of the fix for Lucky 13 padding
6632 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6633 constant time by making sure that always the same bytes are read and
6634 compared against either the MAC or padding bytes. But it no longer
6635 checked that there was enough data to have both the MAC and padding
6638 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6643 * Fix EVP_EncodeUpdate overflow
6645 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6646 Base64 encoding of binary data. If an attacker is able to supply very large
6647 amounts of input data then a length check can overflow resulting in a heap
6650 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6651 the `PEM_write_bio*` family of functions. These are mainly used within the
6652 OpenSSL command line applications, so any application which processes data
6653 from an untrusted source and outputs it as a PEM file should be considered
6654 vulnerable to this issue. User applications that call these APIs directly
6655 with large amounts of untrusted data may also be vulnerable.
6657 This issue was reported by Guido Vranken.
6662 * Fix EVP_EncryptUpdate overflow
6664 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6665 is able to supply very large amounts of input data after a previous call to
6666 EVP_EncryptUpdate() with a partial block then a length check can overflow
6667 resulting in a heap corruption. Following an analysis of all OpenSSL
6668 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6669 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6670 the first called function after an EVP_EncryptInit(), and therefore that
6671 specific call must be safe. The second form is where the length passed to
6672 EVP_EncryptUpdate() can be seen from the code to be some small value and
6673 therefore there is no possibility of an overflow. Since all instances are
6674 one of these two forms, it is believed that there can be no overflows in
6675 internal code due to this problem. It should be noted that
6676 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6677 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6678 of these calls have also been analysed too and it is believed there are no
6679 instances in internal usage where an overflow could occur.
6681 This issue was reported by Guido Vranken.
6686 * Prevent ASN.1 BIO excessive memory allocation
6688 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6689 a short invalid encoding can casuse allocation of large amounts of memory
6690 potentially consuming excessive resources or exhausting memory.
6692 Any application parsing untrusted data through d2i BIO functions is
6693 affected. The memory based functions such as d2i_X509() are *not* affected.
6694 Since the memory based functions are used by the TLS library, TLS
6695 applications are not affected.
6697 This issue was reported by Brian Carpenter.
6704 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6705 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6706 in arbitrary stack data being returned in the buffer.
6708 This issue was reported by Guido Vranken.
6713 * Modify behavior of ALPN to invoke callback after SNI/servername
6714 callback, such that updates to the SSL_CTX affect ALPN.
6718 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6723 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6724 methods are enabled and ssl2 is disabled the methods return NULL.
6728 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6730 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6731 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6732 provide any "EXPORT" or "LOW" strength ciphers.
6736 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6737 is by default disabled at build-time. Builds that are not configured with
6738 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6739 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6740 will need to explicitly call either of:
6742 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6744 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6746 as appropriate. Even if either of those is used, or the application
6747 explicitly uses the version-specific SSLv2_method() or its client and
6748 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6749 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6750 ciphers, and SSLv2 56-bit DES are no longer available.
6755 * Fix a double-free in DSA code
6757 A double free bug was discovered when OpenSSL parses malformed DSA private
6758 keys and could lead to a DoS attack or memory corruption for applications
6759 that receive DSA private keys from untrusted sources. This scenario is
6762 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6768 * Disable SRP fake user seed to address a server memory leak.
6770 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6772 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6773 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6774 was changed to ignore the "fake user" SRP seed, even if the seed
6777 Users should use SRP_VBASE_get1_by_user instead. Note that in
6778 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6779 also that even though configuring the SRP seed attempts to hide
6780 invalid usernames by continuing the handshake with fake
6781 credentials, this behaviour is not constant time and no strong
6782 guarantees are made that the handshake is indistinguishable from
6783 that of a valid user.
6788 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6790 In the BN_hex2bn function the number of hex digits is calculated using an
6791 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6792 large values of `i` this can result in `bn_expand` not allocating any
6793 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6794 field as NULL leading to a subsequent NULL ptr deref. For very large values
6795 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6796 In this case memory is allocated to the internal BIGNUM data field, but it
6797 is insufficiently sized leading to heap corruption. A similar issue exists
6798 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6799 is ever called by user applications with very large untrusted hex/dec data.
6800 This is anticipated to be a rare occurrence.
6802 All OpenSSL internal usage of these functions use data that is not expected
6803 to be untrusted, e.g. config file data or application command line
6804 arguments. If user developed applications generate config file data based
6805 on untrusted data then it is possible that this could also lead to security
6806 consequences. This is also anticipated to be rare.
6808 This issue was reported to OpenSSL by Guido Vranken.
6813 * Fix memory issues in `BIO_*printf` functions
6815 The internal `fmtstr` function used in processing a "%s" format string in
6816 the `BIO_*printf` functions could overflow while calculating the length of a
6817 string and cause an OOB read when printing very long strings.
6819 Additionally the internal `doapr_outch` function can attempt to write to an
6820 OOB memory location (at an offset from the NULL pointer) in the event of a
6821 memory allocation failure. In 1.0.2 and below this could be caused where
6822 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6823 could be in processing a very long "%s" format string. Memory leaks can
6826 The first issue may mask the second issue dependent on compiler behaviour.
6827 These problems could enable attacks where large amounts of untrusted data
6828 is passed to the `BIO_*printf` functions. If applications use these functions
6829 in this way then they could be vulnerable. OpenSSL itself uses these
6830 functions when printing out human-readable dumps of ASN.1 data. Therefore
6831 applications that print this data could be vulnerable if the data is from
6832 untrusted sources. OpenSSL command line applications could also be
6833 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6834 as command line arguments.
6836 Libssl is not considered directly vulnerable. Additionally certificates etc
6837 received via remote connections via libssl are also unlikely to be able to
6838 trigger these issues because of message size limits enforced within libssl.
6840 This issue was reported to OpenSSL Guido Vranken.
6845 * Side channel attack on modular exponentiation
6847 A side-channel attack was found which makes use of cache-bank conflicts on
6848 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6849 of RSA keys. The ability to exploit this issue is limited as it relies on
6850 an attacker who has control of code in a thread running on the same
6851 hyper-threaded core as the victim thread which is performing decryptions.
6853 This issue was reported to OpenSSL by Yuval Yarom, The University of
6854 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6855 Nadia Heninger, University of Pennsylvania with more information at
6856 <http://cachebleed.info>.
6861 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6862 if no keysize is specified with default_bits. This fixes an
6863 omission in an earlier change that changed all RSA/DSA key generation
6864 commands to use 2048 bits by default.
6868 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6870 * Protection for DH small subgroup attacks
6872 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6873 switched on by default and cannot be disabled. This could have some
6878 * SSLv2 doesn't block disabled ciphers
6880 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6881 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6882 been disabled, provided that the SSLv2 protocol was not also disabled via
6885 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6886 and Sebastian Schinzel.
6891 * Reject DH handshakes with parameters shorter than 1024 bits.
6895 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6897 * Certificate verify crash with missing PSS parameter
6899 The signature verification routines will crash with a NULL pointer
6900 dereference if presented with an ASN.1 signature using the RSA PSS
6901 algorithm and absent mask generation function parameter. Since these
6902 routines are used to verify certificate signature algorithms this can be
6903 used to crash any certificate verification operation and exploited in a
6904 DoS attack. Any application which performs certificate verification is
6905 vulnerable including OpenSSL clients and servers which enable client
6908 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6913 * X509_ATTRIBUTE memory leak
6915 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6916 memory. This structure is used by the PKCS#7 and CMS routines so any
6917 application which reads PKCS#7 or CMS data from untrusted sources is
6918 affected. SSL/TLS is not affected.
6920 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6926 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6927 This changes the decoding behaviour for some invalid messages,
6928 though the change is mostly in the more lenient direction, and
6929 legacy behaviour is preserved as much as possible.
6933 * In DSA_generate_parameters_ex, if the provided seed is too short,
6934 use a random seed, as already documented.
6936 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6938 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6940 * Alternate chains certificate forgery
6942 During certificate verfification, OpenSSL will attempt to find an
6943 alternative certificate chain if the first attempt to build such a chain
6944 fails. An error in the implementation of this logic can mean that an
6945 attacker could cause certain checks on untrusted certificates to be
6946 bypassed, such as the CA flag, enabling them to use a valid leaf
6947 certificate to act as a CA and "issue" an invalid certificate.
6949 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6955 * Race condition handling PSK identify hint
6957 If PSK identity hints are received by a multi-threaded client then
6958 the values are wrongly updated in the parent SSL_CTX structure. This can
6959 result in a race condition potentially leading to a double free of the
6965 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6967 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6968 incompatibility in the handling of HMAC. The previous ABI has now been
6971 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6973 * Malformed ECParameters causes infinite loop
6975 When processing an ECParameters structure OpenSSL enters an infinite loop
6976 if the curve specified is over a specially malformed binary polynomial
6979 This can be used to perform denial of service against any
6980 system which processes public keys, certificate requests or
6981 certificates. This includes TLS clients and TLS servers with
6982 client authentication enabled.
6984 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6989 * Exploitable out-of-bounds read in X509_cmp_time
6991 X509_cmp_time does not properly check the length of the ASN1_TIME
6992 string and can read a few bytes out of bounds. In addition,
6993 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6996 An attacker can use this to craft malformed certificates and CRLs of
6997 various sizes and potentially cause a segmentation fault, resulting in
6998 a DoS on applications that verify certificates or CRLs. TLS clients
6999 that verify CRLs are affected. TLS clients and servers with client
7000 authentication enabled may be affected if they use custom verification
7003 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7004 independently by Hanno Böck.
7009 * PKCS7 crash with missing EnvelopedContent
7011 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7012 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7013 with missing content and trigger a NULL pointer dereference on parsing.
7015 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7016 structures from untrusted sources are affected. OpenSSL clients and
7017 servers are not affected.
7019 This issue was reported to OpenSSL by Michal Zalewski (Google).
7024 * CMS verify infinite loop with unknown hash function
7026 When verifying a signedData message the CMS code can enter an infinite loop
7027 if presented with an unknown hash function OID. This can be used to perform
7028 denial of service against any system which verifies signedData messages using
7030 This issue was reported to OpenSSL by Johannes Bauer.
7035 * Race condition handling NewSessionTicket
7037 If a NewSessionTicket is received by a multi-threaded client when attempting to
7038 reuse a previous ticket then a race condition can occur potentially leading to
7039 a double free of the ticket data.
7044 * Reject DH handshakes with parameters shorter than 768 bits.
7046 *Kurt Roeckx and Emilia Kasper*
7048 * dhparam: generate 2048-bit parameters by default.
7050 *Kurt Roeckx and Emilia Kasper*
7052 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7054 * Segmentation fault in ASN1_TYPE_cmp fix
7056 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7057 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7058 certificate signature algorithm consistency this can be used to crash any
7059 certificate verification operation and exploited in a DoS attack. Any
7060 application which performs certificate verification is vulnerable including
7061 OpenSSL clients and servers which enable client authentication.
7066 * ASN.1 structure reuse memory corruption fix
7068 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7069 memory corruption via an invalid write. Such reuse is and has been
7070 strongly discouraged and is believed to be rare.
7072 Applications that parse structures containing CHOICE or ANY DEFINED BY
7073 components may be affected. Certificate parsing (d2i_X509 and related
7074 functions) are however not affected. OpenSSL clients and servers are
7080 * PKCS7 NULL pointer dereferences fix
7082 The PKCS#7 parsing code does not handle missing outer ContentInfo
7083 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7084 missing content and trigger a NULL pointer dereference on parsing.
7086 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7087 otherwise parse PKCS#7 structures from untrusted sources are
7088 affected. OpenSSL clients and servers are not affected.
7090 This issue was reported to OpenSSL by Michal Zalewski (Google).
7095 * DoS via reachable assert in SSLv2 servers fix
7097 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7098 servers that both support SSLv2 and enable export cipher suites by sending
7099 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7101 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7102 (OpenSSL development team).
7107 * Use After Free following d2i_ECPrivatekey error fix
7109 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7110 could cause a use after free condition. This, in turn, could cause a double
7111 free in several private key parsing functions (such as d2i_PrivateKey
7112 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7113 for applications that receive EC private keys from untrusted
7114 sources. This scenario is considered rare.
7116 This issue was discovered by the BoringSSL project and fixed in their
7122 * X509_to_X509_REQ NULL pointer deref fix
7124 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7125 the certificate key is invalid. This function is rarely used in practice.
7127 This issue was discovered by Brian Carpenter.
7132 * Removed the export ciphers from the DEFAULT ciphers
7136 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7138 * Build fixes for the Windows and OpenVMS platforms
7140 *Matt Caswell and Richard Levitte*
7142 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7144 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7145 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7146 dereference. This could lead to a Denial Of Service attack. Thanks to
7147 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7152 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7153 dtls1_buffer_record function under certain conditions. In particular this
7154 could occur if an attacker sent repeated DTLS records with the same
7155 sequence number but for the next epoch. The memory leak could be exploited
7156 by an attacker in a Denial of Service attack through memory exhaustion.
7157 Thanks to Chris Mueller for reporting this issue.
7162 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7163 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7164 method would be set to NULL which could later result in a NULL pointer
7165 dereference. Thanks to Frank Schmirler for reporting this issue.
7170 * Abort handshake if server key exchange message is omitted for ephemeral
7173 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7174 reporting this issue.
7179 * Remove non-export ephemeral RSA code on client and server. This code
7180 violated the TLS standard by allowing the use of temporary RSA keys in
7181 non-export ciphersuites and could be used by a server to effectively
7182 downgrade the RSA key length used to a value smaller than the server
7183 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7184 INRIA or reporting this issue.
7189 * Fixed issue where DH client certificates are accepted without verification.
7190 An OpenSSL server will accept a DH certificate for client authentication
7191 without the certificate verify message. This effectively allows a client to
7192 authenticate without the use of a private key. This only affects servers
7193 which trust a client certificate authority which issues certificates
7194 containing DH keys: these are extremely rare and hardly ever encountered.
7195 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7201 * Ensure that the session ID context of an SSL is updated when its
7202 SSL_CTX is updated via SSL_set_SSL_CTX.
7204 The session ID context is typically set from the parent SSL_CTX,
7205 and can vary with the CTX.
7209 * Fix various certificate fingerprint issues.
7211 By using non-DER or invalid encodings outside the signed portion of a
7212 certificate the fingerprint can be changed without breaking the signature.
7213 Although no details of the signed portion of the certificate can be changed
7214 this can cause problems with some applications: e.g. those using the
7215 certificate fingerprint for blacklists.
7217 1. Reject signatures with non zero unused bits.
7219 If the BIT STRING containing the signature has non zero unused bits reject
7220 the signature. All current signature algorithms require zero unused bits.
7222 2. Check certificate algorithm consistency.
7224 Check the AlgorithmIdentifier inside TBS matches the one in the
7225 certificate signature. NB: this will result in signature failure
7226 errors for some broken certificates.
7228 Thanks to Konrad Kraszewski from Google for reporting this issue.
7230 3. Check DSA/ECDSA signatures use DER.
7232 Re-encode DSA/ECDSA signatures and compare with the original received
7233 signature. Return an error if there is a mismatch.
7235 This will reject various cases including garbage after signature
7236 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7237 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7238 (negative or with leading zeroes).
7240 Further analysis was conducted and fixes were developed by Stephen Henson
7241 of the OpenSSL core team.
7247 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7248 results on some platforms, including x86_64. This bug occurs at random
7249 with a very low probability, and is not known to be exploitable in any
7250 way, though its exact impact is difficult to determine. Thanks to Pieter
7251 Wuille (Blockstream) who reported this issue and also suggested an initial
7252 fix. Further analysis was conducted by the OpenSSL development team and
7253 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7254 the OpenSSL core team.
7259 * Do not resume sessions on the server if the negotiated protocol
7260 version does not match the session's version. Resuming with a different
7261 version, while not strictly forbidden by the RFC, is of questionable
7262 sanity and breaks all known clients.
7264 *David Benjamin, Emilia Käsper*
7266 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7267 early CCS messages during renegotiation. (Note that because
7268 renegotiation is encrypted, this early CCS was not exploitable.)
7272 * Tighten client-side session ticket handling during renegotiation:
7273 ensure that the client only accepts a session ticket if the server sends
7274 the extension anew in the ServerHello. Previously, a TLS client would
7275 reuse the old extension state and thus accept a session ticket if one was
7276 announced in the initial ServerHello.
7278 Similarly, ensure that the client requires a session ticket if one
7279 was advertised in the ServerHello. Previously, a TLS client would
7280 ignore a missing NewSessionTicket message.
7284 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7288 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7289 sends a carefully crafted handshake message, to cause OpenSSL to fail
7290 to free up to 64k of memory causing a memory leak. This could be
7291 exploited in a Denial Of Service attack. This issue affects OpenSSL
7292 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7293 whether SRTP is used or configured. Implementations of OpenSSL that
7294 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7296 The fix was developed by the OpenSSL team.
7301 * Session Ticket Memory Leak.
7303 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7304 integrity of that ticket is first verified. In the event of a session
7305 ticket integrity check failing, OpenSSL will fail to free memory
7306 causing a memory leak. By sending a large number of invalid session
7307 tickets an attacker could exploit this issue in a Denial Of Service
7313 * Build option no-ssl3 is incomplete.
7315 When OpenSSL is configured with "no-ssl3" as a build option, servers
7316 could accept and complete an SSL 3.0 handshake, and clients could be
7317 configured to send them.
7320 *Akamai and the OpenSSL team*
7322 * Add support for TLS_FALLBACK_SCSV.
7323 Client applications doing fallback retries should call
7324 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7327 *Adam Langley, Bodo Moeller*
7329 * Add additional DigestInfo checks.
7331 Re-encode DigestInto in DER and check against the original when
7332 verifying RSA signature: this will reject any improperly encoded
7333 DigestInfo structures.
7335 Note: this is a precautionary measure and no attacks are currently known.
7339 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7341 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7342 SRP code can be overrun an internal buffer. Add sanity check that
7343 g, A, B < N to SRP code.
7345 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7346 Group for discovering this issue.
7351 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7352 TLS 1.0 instead of higher protocol versions when the ClientHello message
7353 is badly fragmented. This allows a man-in-the-middle attacker to force a
7354 downgrade to TLS 1.0 even if both the server and the client support a
7355 higher protocol version, by modifying the client's TLS records.
7357 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7358 researching this issue.
7363 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7364 to a denial of service attack. A malicious server can crash the client
7365 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7366 ciphersuite and sending carefully crafted handshake messages.
7368 Thanks to Felix Gröbert (Google) for discovering and researching this
7374 * By sending carefully crafted DTLS packets an attacker could cause openssl
7375 to leak memory. This can be exploited through a Denial of Service attack.
7376 Thanks to Adam Langley for discovering and researching this issue.
7381 * An attacker can force openssl to consume large amounts of memory whilst
7382 processing DTLS handshake messages. This can be exploited through a
7383 Denial of Service attack.
7384 Thanks to Adam Langley for discovering and researching this issue.
7389 * An attacker can force an error condition which causes openssl to crash
7390 whilst processing DTLS packets due to memory being freed twice. This
7391 can be exploited through a Denial of Service attack.
7392 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7398 * If a multithreaded client connects to a malicious server using a resumed
7399 session and the server sends an ec point format extension it could write
7400 up to 255 bytes to freed memory.
7402 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7408 * A malicious server can crash an OpenSSL client with a null pointer
7409 dereference (read) by specifying an SRP ciphersuite even though it was not
7410 properly negotiated with the client. This can be exploited through a
7411 Denial of Service attack.
7413 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7414 discovering and researching this issue.
7419 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7420 X509_name_oneline, X509_name_print_ex et al. to leak some information
7421 from the stack. Applications may be affected if they echo pretty printing
7422 output to the attacker.
7424 Thanks to Ivan Fratric (Google) for discovering this issue.
7427 *Emilia Käsper, and Steve Henson*
7429 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7430 for corner cases. (Certain input points at infinity could lead to
7431 bogus results, with non-infinity inputs mapped to infinity too.)
7435 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7437 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7438 handshake can force the use of weak keying material in OpenSSL
7439 SSL/TLS clients and servers.
7441 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7442 researching this issue. ([CVE-2014-0224])
7444 *KIKUCHI Masashi, Steve Henson*
7446 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7447 OpenSSL DTLS client the code can be made to recurse eventually crashing
7450 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7453 *Imre Rad, Steve Henson*
7455 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7456 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7457 client or server. This is potentially exploitable to run arbitrary
7458 code on a vulnerable client or server.
7460 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7462 *Jüri Aedla, Steve Henson*
7464 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7465 are subject to a denial of service attack.
7467 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7468 this issue. ([CVE-2014-3470])
7470 *Felix Gröbert, Ivan Fratric, Steve Henson*
7472 * Harmonize version and its documentation. -f flag is used to display
7475 *mancha <mancha1@zoho.com>*
7477 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7478 in i2d_ECPrivateKey.
7480 *mancha <mancha1@zoho.com>*
7482 * Fix some double frees. These are not thought to be exploitable.
7484 *mancha <mancha1@zoho.com>*
7486 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7488 * A missing bounds check in the handling of the TLS heartbeat extension
7489 can be used to reveal up to 64k of memory to a connected client or
7492 Thanks for Neel Mehta of Google Security for discovering this bug and to
7493 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7494 preparing the fix ([CVE-2014-0160])
7496 *Adam Langley, Bodo Moeller*
7498 * Fix for the attack described in the paper "Recovering OpenSSL
7499 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7500 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7501 <http://eprint.iacr.org/2014/140>
7503 Thanks to Yuval Yarom and Naomi Benger for discovering this
7504 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7506 *Yuval Yarom and Naomi Benger*
7508 * TLS pad extension: draft-agl-tls-padding-03
7510 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7511 TLS client Hello record length value would otherwise be > 255 and
7512 less that 512 pad with a dummy extension containing zeroes so it
7513 is at least 512 bytes long.
7515 *Adam Langley, Steve Henson*
7517 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7519 * Fix for TLS record tampering bug. A carefully crafted invalid
7520 handshake could crash OpenSSL with a NULL pointer exception.
7521 Thanks to Anton Johansson for reporting this issues.
7524 * Keep original DTLS digest and encryption contexts in retransmission
7525 structures so we can use the previous session parameters if they need
7526 to be resent. ([CVE-2013-6450])
7530 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7531 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7532 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7533 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7534 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7535 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7537 *Rob Stradling, Adam Langley*
7539 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7541 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7542 supporting platforms or when small records were transferred.
7544 *Andy Polyakov, Steve Henson*
7546 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7548 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7550 This addresses the flaw in CBC record processing discovered by
7551 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7552 at: <http://www.isg.rhul.ac.uk/tls/>
7554 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7555 Security Group at Royal Holloway, University of London
7556 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7557 Emilia Käsper for the initial patch.
7560 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7562 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7563 ciphersuites which can be exploited in a denial of service attack.
7564 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7565 and detecting this bug and to Wolfgang Ettlinger
7566 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7571 * Return an error when checking OCSP signatures when key is NULL.
7572 This fixes a DoS attack. ([CVE-2013-0166])
7576 * Make openssl verify return errors.
7578 *Chris Palmer <palmer@google.com> and Ben Laurie*
7580 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7581 the right response is stapled. Also change SSL_get_certificate()
7582 so it returns the certificate actually sent.
7583 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7585 *Rob Stradling <rob.stradling@comodo.com>*
7587 * Fix possible deadlock when decoding public keys.
7591 * Don't use TLS 1.0 record version number in initial client hello
7596 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7598 * Sanity check record length before skipping explicit IV in TLS
7599 1.2, 1.1 and DTLS to fix DoS attack.
7601 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7602 fuzzing as a service testing platform.
7607 * Initialise tkeylen properly when encrypting CMS messages.
7608 Thanks to Solar Designer of Openwall for reporting this issue.
7612 * In FIPS mode don't try to use composite ciphers as they are not
7617 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7619 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7620 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7621 mean any application compiled against OpenSSL 1.0.0 headers setting
7622 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7623 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7624 0x10000000L Any application which was previously compiled against
7625 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7626 will need to be recompiled as a result. Letting be results in
7627 inability to disable specifically TLS 1.1 and in client context,
7628 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7632 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7633 disable just protocol X, but all protocols above X *if* there are
7634 protocols *below* X still enabled. In more practical terms it means
7635 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7636 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7637 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7642 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7644 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7645 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7646 in CRYPTO_realloc_clean.
7648 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7649 issue and to Adam Langley <agl@chromium.org> for fixing it.
7652 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7654 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7658 * Workarounds for some broken servers that "hang" if a client hello
7659 record length exceeds 255 bytes.
7661 1. Do not use record version number > TLS 1.0 in initial client
7662 hello: some (but not all) hanging servers will now work.
7663 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7664 the number of ciphers sent in the client hello. This should be
7665 set to an even number, such as 50, for example by passing:
7666 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7667 Most broken servers should now work.
7668 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7669 TLS 1.2 client support entirely.
7673 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7677 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7679 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7680 STRING form instead of a DigestInfo.
7684 * The format used for MDC2 RSA signatures is inconsistent between EVP
7685 and the RSA_sign/RSA_verify functions. This was made more apparent when
7686 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7687 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7688 the correct format in RSA_verify so both forms transparently work.
7692 * Some servers which support TLS 1.0 can choke if we initially indicate
7693 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7694 encrypted premaster secret. As a workaround use the maximum permitted
7695 client version in client hello, this should keep such servers happy
7696 and still work with previous versions of OpenSSL.
7700 * Add support for TLS/DTLS heartbeats.
7702 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7704 * Add support for SCTP.
7706 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7708 * Improved PRNG seeding for VOS.
7710 *Paul Green <Paul.Green@stratus.com>*
7712 * Extensive assembler packs updates, most notably:
7714 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7715 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7716 - x86_64: bit-sliced AES implementation;
7717 - ARM: NEON support, contemporary platforms optimizations;
7718 - s390x: z196 support;
7719 - `*`: GHASH and GF(2^m) multiplication implementations;
7723 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7724 (removal of unnecessary code)
7726 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7728 * Add TLS key material exporter from RFC 5705.
7732 * Add DTLS-SRTP negotiation from RFC 5764.
7736 * Add Next Protocol Negotiation,
7737 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7738 disabled with a no-npn flag to config or Configure. Code donated
7741 *Adam Langley <agl@google.com> and Ben Laurie*
7743 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7744 NIST-P256, NIST-P521, with constant-time single point multiplication on
7745 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7746 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7747 Code made available under Apache License version 2.0.
7749 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7750 line to include this in your build of OpenSSL, and run "make depend" (or
7751 "make update"). This enables the following EC_METHODs:
7753 EC_GFp_nistp224_method()
7754 EC_GFp_nistp256_method()
7755 EC_GFp_nistp521_method()
7757 EC_GROUP_new_by_curve_name() will automatically use these (while
7758 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7761 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7763 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7764 all platforms. Move ssize_t definition from e_os.h to the public
7765 header file e_os2.h as it now appears in public header file cms.h
7769 * New -sigopt option to the ca, req and x509 utilities. Additional
7770 signature parameters can be passed using this option and in
7775 * Add RSA PSS signing function. This will generate and set the
7776 appropriate AlgorithmIdentifiers for PSS based on those in the
7777 corresponding EVP_MD_CTX structure. No application support yet.
7781 * Support for companion algorithm specific ASN1 signing routines.
7782 New function ASN1_item_sign_ctx() signs a pre-initialised
7783 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7784 the appropriate parameters.
7788 * Add new algorithm specific ASN1 verification initialisation function
7789 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7790 handling will be the same no matter what EVP_PKEY_METHOD is used.
7791 Add a PSS handler to support verification of PSS signatures: checked
7792 against a number of sample certificates.
7796 * Add signature printing for PSS. Add PSS OIDs.
7798 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7800 * Add algorithm specific signature printing. An individual ASN1 method
7801 can now print out signatures instead of the standard hex dump.
7803 More complex signatures (e.g. PSS) can print out more meaningful
7804 information. Include DSA version that prints out the signature
7809 * Password based recipient info support for CMS library: implementing
7814 * Split password based encryption into PBES2 and PBKDF2 functions. This
7815 neatly separates the code into cipher and PBE sections and is required
7816 for some algorithms that split PBES2 into separate pieces (such as
7817 password based CMS).
7821 * Session-handling fixes:
7822 - Fix handling of connections that are resuming with a session ID,
7823 but also support Session Tickets.
7824 - Fix a bug that suppressed issuing of a new ticket if the client
7825 presented a ticket with an expired session.
7826 - Try to set the ticket lifetime hint to something reasonable.
7827 - Make tickets shorter by excluding irrelevant information.
7828 - On the client side, don't ignore renewed tickets.
7830 *Adam Langley, Bodo Moeller (Google)*
7832 * Fix PSK session representation.
7836 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7838 This work was sponsored by Intel.
7842 * Add GCM support to TLS library. Some custom code is needed to split
7843 the IV between the fixed (from PRF) and explicit (from TLS record)
7844 portions. This adds all GCM ciphersuites supported by RFC5288 and
7845 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7846 add a special AESGCM string for GCM only.
7850 * Expand range of ctrls for AES GCM. Permit setting invocation
7851 field on decrypt and retrieval of invocation field only on encrypt.
7855 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7856 As required by RFC5289 these ciphersuites cannot be used if for
7857 versions of TLS earlier than 1.2.
7861 * For FIPS capable OpenSSL interpret a NULL default public key method
7862 as unset and return the appropriate default but do *not* set the default.
7863 This means we can return the appropriate method in applications that
7864 switch between FIPS and non-FIPS modes.
7868 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7869 ENGINE is used then we cannot handle that in the FIPS module so we
7870 keep original code iff non-FIPS operations are allowed.
7874 * Add -attime option to openssl utilities.
7876 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7878 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7882 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7883 FIPS EC methods unconditionally for now.
7887 * New build option no-ec2m to disable characteristic 2 code.
7891 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7892 all cases can be covered as some introduce binary incompatibilities.
7896 * Redirect RSA operations to FIPS module including keygen,
7897 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7901 * Add similar low-level API blocking to ciphers.
7905 * low-level digest APIs are not approved in FIPS mode: any attempt
7906 to use these will cause a fatal error. Applications that *really* want
7907 to use them can use the `private_*` version instead.
7911 * Redirect cipher operations to FIPS module for FIPS builds.
7915 * Redirect digest operations to FIPS module for FIPS builds.
7919 * Update build system to add "fips" flag which will link in fipscanister.o
7920 for static and shared library builds embedding a signature if needed.
7924 * Output TLS supported curves in preference order instead of numerical
7925 order. This is currently hardcoded for the highest order curves first.
7926 This should be configurable so applications can judge speed vs strength.
7930 * Add TLS v1.2 server support for client authentication.
7934 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7939 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7940 FIPS modules versions.
7944 * Add TLS v1.2 client side support for client authentication. Keep cache
7945 of handshake records longer as we don't know the hash algorithm to use
7946 until after the certificate request message is received.
7950 * Initial TLS v1.2 client support. Add a default signature algorithms
7951 extension including all the algorithms we support. Parse new signature
7952 format in client key exchange. Relax some ECC signing restrictions for
7953 TLS v1.2 as indicated in RFC5246.
7957 * Add server support for TLS v1.2 signature algorithms extension. Switch
7958 to new signature format when needed using client digest preference.
7959 All server ciphersuites should now work correctly in TLS v1.2. No client
7960 support yet and no support for client certificates.
7964 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7965 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7966 ciphersuites. At present only RSA key exchange ciphersuites work with
7967 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7968 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7969 and version checking.
7973 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7974 with this defined it will not be affected by any changes to ssl internal
7975 structures. Add several utility functions to allow openssl application
7976 to work with OPENSSL_NO_SSL_INTERN defined.
7980 * A long standing patch to add support for SRP from EdelWeb (Peter
7981 Sylvester and Christophe Renou) was integrated.
7982 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7983 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7986 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7990 * Permit abbreviated handshakes when renegotiating using the function
7991 SSL_renegotiate_abbreviated().
7993 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7995 * Add call to ENGINE_register_all_complete() to
7996 ENGINE_load_builtin_engines(), so some implementations get used
7997 automatically instead of needing explicit application support.
8001 * Add support for TLS key exporter as described in RFC5705.
8003 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8005 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8006 a few changes are required:
8008 Add SSL_OP_NO_TLSv1_1 flag.
8009 Add TLSv1_1 methods.
8010 Update version checking logic to handle version 1.1.
8011 Add explicit IV handling (ported from DTLS code).
8012 Add command line options to s_client/s_server.
8019 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8021 * X509_ATTRIBUTE memory leak
8023 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8024 memory. This structure is used by the PKCS#7 and CMS routines so any
8025 application which reads PKCS#7 or CMS data from untrusted sources is
8026 affected. SSL/TLS is not affected.
8028 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8034 * Race condition handling PSK identify hint
8036 If PSK identity hints are received by a multi-threaded client then
8037 the values are wrongly updated in the parent SSL_CTX structure. This can
8038 result in a race condition potentially leading to a double free of the
8044 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8046 * Malformed ECParameters causes infinite loop
8048 When processing an ECParameters structure OpenSSL enters an infinite loop
8049 if the curve specified is over a specially malformed binary polynomial
8052 This can be used to perform denial of service against any
8053 system which processes public keys, certificate requests or
8054 certificates. This includes TLS clients and TLS servers with
8055 client authentication enabled.
8057 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8062 * Exploitable out-of-bounds read in X509_cmp_time
8064 X509_cmp_time does not properly check the length of the ASN1_TIME
8065 string and can read a few bytes out of bounds. In addition,
8066 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8069 An attacker can use this to craft malformed certificates and CRLs of
8070 various sizes and potentially cause a segmentation fault, resulting in
8071 a DoS on applications that verify certificates or CRLs. TLS clients
8072 that verify CRLs are affected. TLS clients and servers with client
8073 authentication enabled may be affected if they use custom verification
8076 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8077 independently by Hanno Böck.
8082 * PKCS7 crash with missing EnvelopedContent
8084 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8085 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8086 with missing content and trigger a NULL pointer dereference on parsing.
8088 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8089 structures from untrusted sources are affected. OpenSSL clients and
8090 servers are not affected.
8092 This issue was reported to OpenSSL by Michal Zalewski (Google).
8097 * CMS verify infinite loop with unknown hash function
8099 When verifying a signedData message the CMS code can enter an infinite loop
8100 if presented with an unknown hash function OID. This can be used to perform
8101 denial of service against any system which verifies signedData messages using
8103 This issue was reported to OpenSSL by Johannes Bauer.
8108 * Race condition handling NewSessionTicket
8110 If a NewSessionTicket is received by a multi-threaded client when attempting to
8111 reuse a previous ticket then a race condition can occur potentially leading to
8112 a double free of the ticket data.
8117 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8119 * Segmentation fault in ASN1_TYPE_cmp fix
8121 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8122 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8123 certificate signature algorithm consistency this can be used to crash any
8124 certificate verification operation and exploited in a DoS attack. Any
8125 application which performs certificate verification is vulnerable including
8126 OpenSSL clients and servers which enable client authentication.
8131 * ASN.1 structure reuse memory corruption fix
8133 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8134 memory corruption via an invalid write. Such reuse is and has been
8135 strongly discouraged and is believed to be rare.
8137 Applications that parse structures containing CHOICE or ANY DEFINED BY
8138 components may be affected. Certificate parsing (d2i_X509 and related
8139 functions) are however not affected. OpenSSL clients and servers are
8145 * PKCS7 NULL pointer dereferences fix
8147 The PKCS#7 parsing code does not handle missing outer ContentInfo
8148 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8149 missing content and trigger a NULL pointer dereference on parsing.
8151 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8152 otherwise parse PKCS#7 structures from untrusted sources are
8153 affected. OpenSSL clients and servers are not affected.
8155 This issue was reported to OpenSSL by Michal Zalewski (Google).
8160 * DoS via reachable assert in SSLv2 servers fix
8162 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8163 servers that both support SSLv2 and enable export cipher suites by sending
8164 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8166 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8167 (OpenSSL development team).
8172 * Use After Free following d2i_ECPrivatekey error fix
8174 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8175 could cause a use after free condition. This, in turn, could cause a double
8176 free in several private key parsing functions (such as d2i_PrivateKey
8177 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8178 for applications that receive EC private keys from untrusted
8179 sources. This scenario is considered rare.
8181 This issue was discovered by the BoringSSL project and fixed in their
8187 * X509_to_X509_REQ NULL pointer deref fix
8189 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8190 the certificate key is invalid. This function is rarely used in practice.
8192 This issue was discovered by Brian Carpenter.
8197 * Removed the export ciphers from the DEFAULT ciphers
8201 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8203 * Build fixes for the Windows and OpenVMS platforms
8205 *Matt Caswell and Richard Levitte*
8207 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8209 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8210 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8211 dereference. This could lead to a Denial Of Service attack. Thanks to
8212 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8217 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8218 dtls1_buffer_record function under certain conditions. In particular this
8219 could occur if an attacker sent repeated DTLS records with the same
8220 sequence number but for the next epoch. The memory leak could be exploited
8221 by an attacker in a Denial of Service attack through memory exhaustion.
8222 Thanks to Chris Mueller for reporting this issue.
8227 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8228 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8229 method would be set to NULL which could later result in a NULL pointer
8230 dereference. Thanks to Frank Schmirler for reporting this issue.
8235 * Abort handshake if server key exchange message is omitted for ephemeral
8238 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8239 reporting this issue.
8244 * Remove non-export ephemeral RSA code on client and server. This code
8245 violated the TLS standard by allowing the use of temporary RSA keys in
8246 non-export ciphersuites and could be used by a server to effectively
8247 downgrade the RSA key length used to a value smaller than the server
8248 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8249 INRIA or reporting this issue.
8254 * Fixed issue where DH client certificates are accepted without verification.
8255 An OpenSSL server will accept a DH certificate for client authentication
8256 without the certificate verify message. This effectively allows a client to
8257 authenticate without the use of a private key. This only affects servers
8258 which trust a client certificate authority which issues certificates
8259 containing DH keys: these are extremely rare and hardly ever encountered.
8260 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8266 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8267 results on some platforms, including x86_64. This bug occurs at random
8268 with a very low probability, and is not known to be exploitable in any
8269 way, though its exact impact is difficult to determine. Thanks to Pieter
8270 Wuille (Blockstream) who reported this issue and also suggested an initial
8271 fix. Further analysis was conducted by the OpenSSL development team and
8272 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8273 the OpenSSL core team.
8278 * Fix various certificate fingerprint issues.
8280 By using non-DER or invalid encodings outside the signed portion of a
8281 certificate the fingerprint can be changed without breaking the signature.
8282 Although no details of the signed portion of the certificate can be changed
8283 this can cause problems with some applications: e.g. those using the
8284 certificate fingerprint for blacklists.
8286 1. Reject signatures with non zero unused bits.
8288 If the BIT STRING containing the signature has non zero unused bits reject
8289 the signature. All current signature algorithms require zero unused bits.
8291 2. Check certificate algorithm consistency.
8293 Check the AlgorithmIdentifier inside TBS matches the one in the
8294 certificate signature. NB: this will result in signature failure
8295 errors for some broken certificates.
8297 Thanks to Konrad Kraszewski from Google for reporting this issue.
8299 3. Check DSA/ECDSA signatures use DER.
8301 Reencode DSA/ECDSA signatures and compare with the original received
8302 signature. Return an error if there is a mismatch.
8304 This will reject various cases including garbage after signature
8305 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8306 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8307 (negative or with leading zeroes).
8309 Further analysis was conducted and fixes were developed by Stephen Henson
8310 of the OpenSSL core team.
8316 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8318 * Session Ticket Memory Leak.
8320 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8321 integrity of that ticket is first verified. In the event of a session
8322 ticket integrity check failing, OpenSSL will fail to free memory
8323 causing a memory leak. By sending a large number of invalid session
8324 tickets an attacker could exploit this issue in a Denial Of Service
8330 * Build option no-ssl3 is incomplete.
8332 When OpenSSL is configured with "no-ssl3" as a build option, servers
8333 could accept and complete an SSL 3.0 handshake, and clients could be
8334 configured to send them.
8337 *Akamai and the OpenSSL team*
8339 * Add support for TLS_FALLBACK_SCSV.
8340 Client applications doing fallback retries should call
8341 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8344 *Adam Langley, Bodo Moeller*
8346 * Add additional DigestInfo checks.
8348 Reencode DigestInto in DER and check against the original when
8349 verifying RSA signature: this will reject any improperly encoded
8350 DigestInfo structures.
8352 Note: this is a precautionary measure and no attacks are currently known.
8356 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8358 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8359 to a denial of service attack. A malicious server can crash the client
8360 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8361 ciphersuite and sending carefully crafted handshake messages.
8363 Thanks to Felix Gröbert (Google) for discovering and researching this
8369 * By sending carefully crafted DTLS packets an attacker could cause openssl
8370 to leak memory. This can be exploited through a Denial of Service attack.
8371 Thanks to Adam Langley for discovering and researching this issue.
8376 * An attacker can force openssl to consume large amounts of memory whilst
8377 processing DTLS handshake messages. This can be exploited through a
8378 Denial of Service attack.
8379 Thanks to Adam Langley for discovering and researching this issue.
8384 * An attacker can force an error condition which causes openssl to crash
8385 whilst processing DTLS packets due to memory being freed twice. This
8386 can be exploited through a Denial of Service attack.
8387 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8393 * If a multithreaded client connects to a malicious server using a resumed
8394 session and the server sends an ec point format extension it could write
8395 up to 255 bytes to freed memory.
8397 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8403 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8404 X509_name_oneline, X509_name_print_ex et al. to leak some information
8405 from the stack. Applications may be affected if they echo pretty printing
8406 output to the attacker.
8408 Thanks to Ivan Fratric (Google) for discovering this issue.
8411 *Emilia Käsper, and Steve Henson*
8413 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8414 for corner cases. (Certain input points at infinity could lead to
8415 bogus results, with non-infinity inputs mapped to infinity too.)
8419 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8421 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8422 handshake can force the use of weak keying material in OpenSSL
8423 SSL/TLS clients and servers.
8425 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8426 researching this issue. ([CVE-2014-0224])
8428 *KIKUCHI Masashi, Steve Henson*
8430 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8431 OpenSSL DTLS client the code can be made to recurse eventually crashing
8434 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8437 *Imre Rad, Steve Henson*
8439 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8440 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8441 client or server. This is potentially exploitable to run arbitrary
8442 code on a vulnerable client or server.
8444 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8446 *Jüri Aedla, Steve Henson*
8448 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8449 are subject to a denial of service attack.
8451 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8452 this issue. ([CVE-2014-3470])
8454 *Felix Gröbert, Ivan Fratric, Steve Henson*
8456 * Harmonize version and its documentation. -f flag is used to display
8459 *mancha <mancha1@zoho.com>*
8461 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8462 in i2d_ECPrivateKey.
8464 *mancha <mancha1@zoho.com>*
8466 * Fix some double frees. These are not thought to be exploitable.
8468 *mancha <mancha1@zoho.com>*
8470 * Fix for the attack described in the paper "Recovering OpenSSL
8471 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8472 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8473 <http://eprint.iacr.org/2014/140>
8475 Thanks to Yuval Yarom and Naomi Benger for discovering this
8476 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8478 *Yuval Yarom and Naomi Benger*
8480 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8482 * Keep original DTLS digest and encryption contexts in retransmission
8483 structures so we can use the previous session parameters if they need
8484 to be resent. ([CVE-2013-6450])
8488 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8489 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8490 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8491 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8492 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8493 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8495 *Rob Stradling, Adam Langley*
8497 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8499 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8501 This addresses the flaw in CBC record processing discovered by
8502 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8503 at: <http://www.isg.rhul.ac.uk/tls/>
8505 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8506 Security Group at Royal Holloway, University of London
8507 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8508 Emilia Käsper for the initial patch.
8511 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8513 * Return an error when checking OCSP signatures when key is NULL.
8514 This fixes a DoS attack. ([CVE-2013-0166])
8518 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8519 the right response is stapled. Also change SSL_get_certificate()
8520 so it returns the certificate actually sent.
8521 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8522 (This is a backport)
8524 *Rob Stradling <rob.stradling@comodo.com>*
8526 * Fix possible deadlock when decoding public keys.
8530 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8532 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8535 * Sanity check record length before skipping explicit IV in DTLS
8538 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8539 fuzzing as a service testing platform.
8544 * Initialise tkeylen properly when encrypting CMS messages.
8545 Thanks to Solar Designer of Openwall for reporting this issue.
8549 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8551 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8552 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8553 in CRYPTO_realloc_clean.
8555 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8556 issue and to Adam Langley <agl@chromium.org> for fixing it.
8559 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8561 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8563 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8564 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8565 content decryption and always return the same error. Note: this attack
8566 needs on average 2^20 messages so it only affects automated senders. The
8567 old behaviour can be re-enabled in the CMS code by setting the
8568 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8569 an MMA defence is not necessary.
8570 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8571 this issue. ([CVE-2012-0884])
8575 * Fix CVE-2011-4619: make sure we really are receiving a
8576 client hello before rejecting multiple SGC restarts. Thanks to
8577 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8581 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8583 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8584 Thanks to Antonio Martin, Enterprise Secure Access Research and
8585 Development, Cisco Systems, Inc. for discovering this bug and
8586 preparing a fix. ([CVE-2012-0050])
8590 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8592 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8593 of the Vaudenay padding oracle attack on CBC mode encryption
8594 which enables an efficient plaintext recovery attack against
8595 the OpenSSL implementation of DTLS. Their attack exploits timing
8596 differences arising during decryption processing. A research
8597 paper describing this attack can be found at:
8598 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8599 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8600 Security Group at Royal Holloway, University of London
8601 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8602 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8603 for preparing the fix. ([CVE-2011-4108])
8605 *Robin Seggelmann, Michael Tuexen*
8607 * Clear bytes used for block padding of SSL 3.0 records.
8610 *Adam Langley (Google)*
8612 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8613 Kadianakis <desnacked@gmail.com> for discovering this issue and
8614 Adam Langley for preparing the fix. ([CVE-2011-4619])
8616 *Adam Langley (Google)*
8618 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8620 *Andrey Kulikov <amdeich@gmail.com>*
8622 * Prevent malformed RFC3779 data triggering an assertion failure.
8623 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8624 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8626 *Rob Austein <sra@hactrn.net>*
8628 * Improved PRNG seeding for VOS.
8630 *Paul Green <Paul.Green@stratus.com>*
8632 * Fix ssl_ciph.c set-up race.
8634 *Adam Langley (Google)*
8636 * Fix spurious failures in ecdsatest.c.
8638 *Emilia Käsper (Google)*
8640 * Fix the BIO_f_buffer() implementation (which was mixing different
8641 interpretations of the `..._len` fields).
8643 *Adam Langley (Google)*
8645 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8646 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8647 threads won't reuse the same blinding coefficients.
8649 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8650 lock to call BN_BLINDING_invert_ex, and avoids one use of
8651 BN_BLINDING_update for each BN_BLINDING structure (previously,
8652 the last update always remained unused).
8654 *Emilia Käsper (Google)*
8656 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8658 *Bob Buckholz (Google)*
8660 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8662 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8663 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8665 *Kaspar Brand <ossl@velox.ch>*
8667 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8668 for multi-threaded use of ECDH. ([CVE-2011-3210])
8670 *Adam Langley (Google)*
8672 * Fix x509_name_ex_d2i memory leak on bad inputs.
8676 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8677 signature public key algorithm by using OID xref utilities instead.
8678 Before this you could only use some ECC ciphersuites with SHA1 only.
8682 * Add protection against ECDSA timing attacks as mentioned in the paper
8683 by Billy Bob Brumley and Nicola Tuveri, see:
8684 <http://eprint.iacr.org/2011/232.pdf>
8686 *Billy Bob Brumley and Nicola Tuveri*
8688 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8690 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8692 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8694 * Fix bug in string printing code: if *any* escaping is enabled we must
8695 escape the escape character (backslash) or the resulting string is
8700 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8702 * Disable code workaround for ancient and obsolete Netscape browsers
8703 and servers: an attacker can use it in a ciphersuite downgrade attack.
8704 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8708 * Fixed J-PAKE implementation error, originally discovered by
8709 Sebastien Martini, further info and confirmation from Stefan
8710 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8714 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8716 * Fix extension code to avoid race conditions which can result in a buffer
8717 overrun vulnerability: resumed sessions must not be modified as they can
8718 be shared by multiple threads. CVE-2010-3864
8722 * Fix WIN32 build system to correctly link an ENGINE directory into
8727 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8729 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8732 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8734 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8736 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8737 context. The operation can be customised via the ctrl mechanism in
8738 case ENGINEs want to include additional functionality.
8742 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8746 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8747 output hashes compatible with older versions of OpenSSL.
8749 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8751 * Fix compression algorithm handling: if resuming a session use the
8752 compression algorithm of the resumed session instead of determining
8753 it from client hello again. Don't allow server to change algorithm.
8757 * Add load_crls() function to commands tidying load_certs() too. Add option
8758 to verify utility to allow additional CRLs to be included.
8762 * Update OCSP request code to permit adding custom headers to the request:
8763 some responders need this.
8767 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8770 *Julia Lawall <julia@diku.dk>*
8772 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8773 needlessly dereferenced structures, used obsolete functions and
8774 didn't handle all updated verify codes correctly.
8778 * Disable MD2 in the default configuration.
8782 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8783 indicate the initial BIO being pushed or popped. This makes it possible
8784 to determine whether the BIO is the one explicitly called or as a result
8785 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8786 it handles reference counts correctly and doesn't zero out the I/O bio
8787 when it is not being explicitly popped. WARNING: applications which
8788 included workarounds for the old buggy behaviour will need to be modified
8789 or they could free up already freed BIOs.
8793 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8794 renaming to all platforms (within the 0.9.8 branch, this was
8795 done conditionally on Netware platforms to avoid a name clash).
8797 *Guenter <lists@gknw.net>*
8799 * Add ECDHE and PSK support to DTLS.
8801 *Michael Tuexen <tuexen@fh-muenster.de>*
8803 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8808 * Add "missing" function EVP_MD_flags() (without this the only way to
8809 retrieve a digest flags is by accessing the structure directly. Update
8810 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8811 or cipher is registered as in the "from" argument. Print out all
8812 registered digests in the dgst usage message instead of manually
8813 attempting to work them out.
8817 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8818 this allows the use of compression and extensions. Change default cipher
8819 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8820 by default unless an application cipher string requests it.
8824 * Alter match criteria in PKCS12_parse(). It used to try to use local
8825 key ids to find matching certificates and keys but some PKCS#12 files
8826 don't follow the (somewhat unwritten) rules and this strategy fails.
8827 Now just gather all certificates together and the first private key
8828 then look for the first certificate that matches the key.
8832 * Support use of registered digest and cipher names for dgst and cipher
8833 commands instead of having to add each one as a special case. So now
8840 openssl dgst -sha256 foo
8842 and this works for ENGINE based algorithms too.
8846 * Update Gost ENGINE to support parameter files.
8848 *Victor B. Wagner <vitus@cryptocom.ru>*
8850 * Support GeneralizedTime in ca utility.
8852 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8854 * Enhance the hash format used for certificate directory links. The new
8855 form uses the canonical encoding (meaning equivalent names will work
8856 even if they aren't identical) and uses SHA1 instead of MD5. This form
8857 is incompatible with the older format and as a result c_rehash should
8858 be used to rebuild symbolic links.
8862 * Make PKCS#8 the default write format for private keys, replacing the
8863 traditional format. This form is standardised, more secure and doesn't
8864 include an implicit MD5 dependency.
8868 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8869 committed to OpenSSL should pass this lot as a minimum.
8873 * Add session ticket override functionality for use by EAP-FAST.
8875 *Jouni Malinen <j@w1.fi>*
8877 * Modify HMAC functions to return a value. Since these can be implemented
8878 in an ENGINE errors can occur.
8882 * Type-checked OBJ_bsearch_ex.
8886 * Type-checked OBJ_bsearch. Also some constification necessitated
8887 by type-checking. Still to come: TXT_DB, bsearch(?),
8888 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8893 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8894 seconds to a tm structure directly, instead of going through OS
8895 specific date routines. This avoids any issues with OS routines such
8896 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8897 and X509_time_adj_ex() to cover the extended range. The existing
8898 X509_time_adj() is still usable and will no longer have any date issues.
8902 * Delta CRL support. New use deltas option which will attempt to locate
8903 and search any appropriate delta CRLs available.
8905 This work was sponsored by Google.
8909 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8910 code and add additional score elements. Validate alternate CRL paths
8911 as part of the CRL checking and indicate a new error "CRL path validation
8912 error" in this case. Applications wanting additional details can use
8913 the verify callback and check the new "parent" field. If this is not
8914 NULL CRL path validation is taking place. Existing applications won't
8915 see this because it requires extended CRL support which is off by
8918 This work was sponsored by Google.
8922 * Support for freshest CRL extension.
8924 This work was sponsored by Google.
8928 * Initial indirect CRL support. Currently only supported in the CRLs
8929 passed directly and not via lookup. Process certificate issuer
8930 CRL entry extension and lookup CRL entries by bother issuer name
8931 and serial number. Check and process CRL issuer entry in IDP extension.
8933 This work was sponsored by Google.
8937 * Add support for distinct certificate and CRL paths. The CRL issuer
8938 certificate is validated separately in this case. Only enabled if
8939 an extended CRL support flag is set: this flag will enable additional
8940 CRL functionality in future.
8942 This work was sponsored by Google.
8946 * Add support for policy mappings extension.
8948 This work was sponsored by Google.
8952 * Fixes to pathlength constraint, self issued certificate handling,
8953 policy processing to align with RFC3280 and PKITS tests.
8955 This work was sponsored by Google.
8959 * Support for name constraints certificate extension. DN, email, DNS
8960 and URI types are currently supported.
8962 This work was sponsored by Google.
8966 * To cater for systems that provide a pointer-based thread ID rather
8967 than numeric, deprecate the current numeric thread ID mechanism and
8968 replace it with a structure and associated callback type. This
8969 mechanism allows a numeric "hash" to be extracted from a thread ID in
8970 either case, and on platforms where pointers are larger than 'long',
8971 mixing is done to help ensure the numeric 'hash' is usable even if it
8972 can't be guaranteed unique. The default mechanism is to use "&errno"
8973 as a pointer-based thread ID to distinguish between threads.
8975 Applications that want to provide their own thread IDs should now use
8976 CRYPTO_THREADID_set_callback() to register a callback that will call
8977 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8979 Note that ERR_remove_state() is now deprecated, because it is tied
8980 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8981 to free the current thread's error state should be replaced by
8982 ERR_remove_thread_state(NULL).
8984 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8985 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8986 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8987 application was previously providing a numeric thread callback that
8988 was inappropriate for distinguishing threads, then uniqueness might
8989 have been obtained with &errno that happened immediately in the
8990 intermediate development versions of OpenSSL; this is no longer the
8991 case, the numeric thread callback will now override the automatic use
8994 *Geoff Thorpe, with help from Bodo Moeller*
8996 * Initial support for different CRL issuing certificates. This covers a
8997 simple case where the self issued certificates in the chain exist and
8998 the real CRL issuer is higher in the existing chain.
9000 This work was sponsored by Google.
9004 * Removed effectively defunct crypto/store from the build.
9008 * Revamp of STACK to provide stronger type-checking. Still to come:
9009 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9010 ASN1_STRING, CONF_VALUE.
9014 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9015 RAM on SSL connections. This option can save about 34k per idle SSL.
9019 * Revamp of LHASH to provide stronger type-checking. Still to come:
9020 STACK, TXT_DB, bsearch, qsort.
9024 * Initial support for Cryptographic Message Syntax (aka CMS) based
9025 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9026 support for data, signedData, compressedData, digestedData and
9027 encryptedData, envelopedData types included. Scripts to check against
9028 RFC4134 examples draft and interop and consistency checks of many
9029 content types and variants.
9033 * Add options to enc utility to support use of zlib compression BIO.
9037 * Extend mk1mf to support importing of options and assembly language
9038 files from Configure script, currently only included in VC-WIN32.
9039 The assembly language rules can now optionally generate the source
9040 files from the associated perl scripts.
9044 * Implement remaining functionality needed to support GOST ciphersuites.
9045 Interop testing has been performed using CryptoPro implementations.
9047 *Victor B. Wagner <vitus@cryptocom.ru>*
9049 * s390x assembler pack.
9053 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9058 * Implement Opaque PRF Input TLS extension as specified in
9059 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9060 official specification yet and no extension type assignment by
9061 IANA exists, this extension (for now) will have to be explicitly
9062 enabled when building OpenSSL by providing the extension number
9063 to use. For example, specify an option
9065 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9067 to the "config" or "Configure" script to enable the extension,
9068 assuming extension number 0x9527 (which is a completely arbitrary
9069 and unofficial assignment based on the MD5 hash of the Internet
9070 Draft). Note that by doing so, you potentially lose
9071 interoperability with other TLS implementations since these might
9072 be using the same extension number for other purposes.
9074 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9075 opaque PRF input value to use in the handshake. This will create
9076 an internal copy of the length-'len' string at 'src', and will
9077 return non-zero for success.
9079 To get more control and flexibility, provide a callback function
9082 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9083 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9087 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9090 Callback function 'cb' will be called in handshakes, and is
9091 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9092 Argument 'arg' is for application purposes (the value as given to
9093 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9094 be provided to the callback function). The callback function
9095 has to return non-zero to report success: usually 1 to use opaque
9096 PRF input just if possible, or 2 to enforce use of the opaque PRF
9097 input. In the latter case, the library will abort the handshake
9098 if opaque PRF input is not successfully negotiated.
9100 Arguments 'peerinput' and 'len' given to the callback function
9101 will always be NULL and 0 in the case of a client. A server will
9102 see the client's opaque PRF input through these variables if
9103 available (NULL and 0 otherwise). Note that if the server
9104 provides an opaque PRF input, the length must be the same as the
9105 length of the client's opaque PRF input.
9107 Note that the callback function will only be called when creating
9108 a new session (session resumption can resume whatever was
9109 previously negotiated), and will not be called in SSL 2.0
9110 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9111 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9112 for applications that need to enforce opaque PRF input.
9116 * Update ssl code to support digests other than SHA1+MD5 for handshake
9119 *Victor B. Wagner <vitus@cryptocom.ru>*
9121 * Add RFC4507 support to OpenSSL. This includes the corrections in
9122 RFC4507bis. The encrypted ticket format is an encrypted encoded
9123 SSL_SESSION structure, that way new session features are automatically
9126 If a client application caches session in an SSL_SESSION structure
9127 support is transparent because tickets are now stored in the encoded
9130 The SSL_CTX structure automatically generates keys for ticket
9131 protection in servers so again support should be possible
9132 with no application modification.
9134 If a client or server wishes to disable RFC4507 support then the option
9135 SSL_OP_NO_TICKET can be set.
9137 Add a TLS extension debugging callback to allow the contents of any client
9138 or server extensions to be examined.
9140 This work was sponsored by Google.
9144 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9145 OpenSSL should now compile cleanly on gcc 4.2
9147 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9149 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9150 support including streaming MAC support: this is required for GOST
9151 ciphersuite support.
9153 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9155 * Add option -stream to use PKCS#7 streaming in smime utility. New
9156 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9157 to output in BER and PEM format.
9161 * Experimental support for use of HMAC via EVP_PKEY interface. This
9162 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9163 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9164 ENGINE support for HMAC keys which are unextractable. New -mac and
9165 -macopt options to dgst utility.
9169 * New option -sigopt to dgst utility. Update dgst to use
9170 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9171 alternative signing parameters such as X9.31 or PSS in the dgst
9176 * Change ssl_cipher_apply_rule(), the internal function that does
9177 the work each time a ciphersuite string requests enabling
9178 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9179 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9180 the order of disabled ciphersuites such that those ciphersuites
9181 that most recently went from enabled to disabled not only stay
9182 in order with respect to each other, but also have higher priority
9183 than other disabled ciphersuites the next time ciphersuites are
9186 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9187 the same ciphersuites as with "HIGH" alone, but in a specific
9188 order where the PSK ciphersuites come first (since they are the
9189 most recently disabled ciphersuites when "HIGH" is parsed).
9191 Also, change ssl_create_cipher_list() (using this new
9192 functionality) such that between otherwise identical
9193 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9198 * Change ssl_create_cipher_list() so that it automatically
9199 arranges the ciphersuites in reasonable order before starting
9200 to process the rule string. Thus, the definition for "DEFAULT"
9201 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9202 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9203 This makes it much easier to arrive at a reasonable default order
9204 in applications for which anonymous ciphers are OK (meaning
9205 that you can't actually use DEFAULT).
9207 *Bodo Moeller; suggested by Victor Duchovni*
9209 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9210 processing) into multiple integers instead of setting
9211 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9212 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9213 (These masks as well as the individual bit definitions are hidden
9214 away into the non-exported interface ssl/ssl_locl.h, so this
9215 change to the definition of the SSL_CIPHER structure shouldn't
9216 affect applications.) This give us more bits for each of these
9217 categories, so there is no longer a need to coagulate AES128 and
9218 AES256 into a single algorithm bit, and to coagulate Camellia128
9219 and Camellia256 into a single algorithm bit, which has led to all
9222 Thus, among other things, the kludge introduced in 0.9.7m and
9223 0.9.8e for masking out AES256 independently of AES128 or masking
9224 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9226 With the change, we also introduce new ciphersuite aliases that
9227 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9232 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9233 Use the leftmost N bytes of the signature input if the input is
9234 larger than the prime q (with N being the size in bytes of q).
9238 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9239 it yet and it is largely untested.
9243 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9247 * Initial incomplete changes to avoid need for function casts in OpenSSL
9248 some compilers (gcc 4.2 and later) reject their use. Safestack is
9249 reimplemented. Update ASN1 to avoid use of legacy functions.
9253 * Win32/64 targets are linked with Winsock2.
9257 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9258 to external functions. This can be used to increase CRL handling
9259 efficiency especially when CRLs are very large by (for example) storing
9260 the CRL revoked certificates in a database.
9264 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9265 new CRLs added to a directory can be used. New command line option
9266 -verify_return_error to s_client and s_server. This causes real errors
9267 to be returned by the verify callback instead of carrying on no matter
9268 what. This reflects the way a "real world" verify callback would behave.
9272 * GOST engine, supporting several GOST algorithms and public key formats.
9273 Kindly donated by Cryptocom.
9277 * Partial support for Issuing Distribution Point CRL extension. CRLs
9278 partitioned by DP are handled but no indirect CRL or reason partitioning
9279 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9280 selected via a scoring technique which handles IDP and AKID in CRLs.
9284 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9285 will ultimately be used for all verify operations: this will remove the
9286 X509_STORE dependency on certificate verification and allow alternative
9287 lookup methods. X509_STORE based implementations of these two callbacks.
9291 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9292 Modify get_crl() to find a valid (unexpired) CRL if possible.
9296 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9297 this would be called X509_CRL_cmp() but that name is already used by
9298 a function that just compares CRL issuer names. Cache several CRL
9299 extensions in X509_CRL structure and cache CRLDP in X509.
9303 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9304 this maps equivalent X509_NAME structures into a consistent structure.
9305 Name comparison can then be performed rapidly using memcmp().
9309 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9314 * Allow digests to supply their own micalg string for S/MIME type using
9315 the ctrl EVP_MD_CTRL_MICALG.
9319 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9320 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9321 ctrl. It can then customise the structure before and/or after signing
9326 * New function OBJ_add_sigid() to allow application defined signature OIDs
9327 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9328 to free up any added signature OIDs.
9332 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9333 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9334 digest and cipher tables. New options added to openssl utility:
9335 list-message-digest-algorithms and list-cipher-algorithms.
9339 * Change the array representation of binary polynomials: the list
9340 of degrees of non-zero coefficients is now terminated with -1.
9341 Previously it was terminated with 0, which was also part of the
9342 value; thus, the array representation was not applicable to
9343 polynomials where t^0 has coefficient zero. This change makes
9344 the array representation useful in a more general context.
9348 * Various modifications and fixes to SSL/TLS cipher string
9349 handling. For ECC, the code now distinguishes between fixed ECDH
9350 with RSA certificates on the one hand and with ECDSA certificates
9351 on the other hand, since these are separate ciphersuites. The
9352 unused code for Fortezza ciphersuites has been removed.
9354 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9355 (not "ECDHE"). For consistency with the code for DH
9356 certificates, use of ECDH certificates is now considered ECDH
9357 authentication, not RSA or ECDSA authentication (the latter is
9358 merely the CA's signing algorithm and not actively used in the
9361 The temporary ciphersuite alias "ECCdraft" is no longer
9362 available, and ECC ciphersuites are no longer excluded from "ALL"
9363 and "DEFAULT". The following aliases now exist for RFC 4492
9364 ciphersuites, most of these by analogy with the DH case:
9366 kECDHr - ECDH cert, signed with RSA
9367 kECDHe - ECDH cert, signed with ECDSA
9368 kECDH - ECDH cert (signed with either RSA or ECDSA)
9369 kEECDH - ephemeral ECDH
9370 ECDH - ECDH cert or ephemeral ECDH
9376 AECDH - anonymous ECDH
9377 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9381 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9382 Use correct micalg parameters depending on digest(s) in signed message.
9386 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9387 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9391 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9392 an engine to register a method. Add ENGINE lookups for methods and
9393 functional reference processing.
9397 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9398 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9403 * New -resign option to smime utility. This adds one or more signers
9404 to an existing PKCS#7 signedData structure. Also -md option to use an
9405 alternative message digest algorithm for signing.
9409 * Tidy up PKCS#7 routines and add new functions to make it easier to
9410 create PKCS7 structures containing multiple signers. Update smime
9411 application to support multiple signers.
9415 * New -macalg option to pkcs12 utility to allow setting of an alternative
9420 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9421 Reorganize PBE internals to lookup from a static table using NIDs,
9422 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9423 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9424 PRF which will be automatically used with PBES2.
9428 * Replace the algorithm specific calls to generate keys in "req" with the
9433 * Update PKCS#7 enveloped data routines to use new API. This is now
9434 supported by any public key method supporting the encrypt operation. A
9435 ctrl is added to allow the public key algorithm to examine or modify
9436 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9441 * Add a ctrl to asn1 method to allow a public key algorithm to express
9442 a default digest type to use. In most cases this will be SHA1 but some
9443 algorithms (such as GOST) need to specify an alternative digest. The
9444 return value indicates how strong the preference is 1 means optional and
9445 2 is mandatory (that is it is the only supported type). Modify
9446 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9447 use the default md. Update openssl utilities to use the default digest
9448 type for signing if it is not explicitly indicated.
9452 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9453 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9454 signing method from the key type. This effectively removes the link
9455 between digests and public key types.
9459 * Add an OID cross reference table and utility functions. Its purpose is to
9460 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9461 rsaEncryption. This will allow some of the algorithm specific hackery
9462 needed to use the correct OID to be removed.
9466 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9467 structures for PKCS7_sign(). They are now set up by the relevant public
9472 * Add provisional EC pkey method with support for ECDSA and ECDH.
9476 * Add support for key derivation (agreement) in the API, DH method and
9481 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9482 public and private key formats. As a side effect these add additional
9483 command line functionality not previously available: DSA signatures can be
9484 generated and verified using pkeyutl and DH key support and generation in
9491 *Oliver Tappe <zooey@hirschkaefer.de>*
9493 * New make target "install_html_docs" installs HTML renditions of the
9496 *Oliver Tappe <zooey@hirschkaefer.de>*
9498 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9499 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9500 support key and parameter generation and add initial key generation
9501 functionality for RSA.
9505 * Add functions for main EVP_PKEY_method operations. The undocumented
9506 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9507 `EVP_PKEY_{encrypt,decrypt}_old`.
9511 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9512 key API, doesn't do much yet.
9516 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9517 public key algorithms. New option to openssl utility:
9518 "list-public-key-algorithms" to print out info.
9522 * Implement the Supported Elliptic Curves Extension for
9523 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9527 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9528 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9532 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9533 utilities such as rsa, dsa, dsaparam etc except they process any key
9538 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9539 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9540 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9545 * Initial support for pluggable public key ASN1.
9546 De-spaghettify the public key ASN1 handling. Move public and private
9547 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9548 algorithm specific handling to a single module within the relevant
9549 algorithm directory. Add functions to allow (near) opaque processing
9550 of public and private key structures.
9554 * Implement the Supported Point Formats Extension for
9555 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9559 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9560 for the psk identity [hint] and the psk callback functions to the
9561 SSL_SESSION, SSL and SSL_CTX structure.
9564 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9568 SSL_CTX_use_psk_identity_hint
9569 SSL_get_psk_identity_hint
9570 SSL_get_psk_identity
9571 SSL_use_psk_identity_hint
9573 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9575 * Add RFC 3161 compliant time stamp request creation, response generation
9576 and response verification functionality.
9578 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9580 * Add initial support for TLS extensions, specifically for the server_name
9581 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9582 have new members for a hostname. The SSL data structure has an
9583 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9584 stored in that context to allow for session resumption, even after the
9585 SSL has been switched to a new SSL_CTX in reaction to a client's
9586 server_name extension.
9588 New functions (subject to change):
9590 SSL_get_servername()
9591 SSL_get_servername_type()
9594 New CTRL codes and macros (subject to change):
9596 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9597 - SSL_CTX_set_tlsext_servername_callback()
9598 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9599 - SSL_CTX_set_tlsext_servername_arg()
9600 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9602 openssl s_client has a new '-servername ...' option.
9604 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9605 '-key2 ...', '-servername_fatal' (subject to change). This allows
9606 testing the HostName extension for a specific single hostname ('-cert'
9607 and '-key' remain fallbacks for handshakes without HostName
9608 negotiation). If the unrecognized_name alert has to be sent, this by
9609 default is a warning; it becomes fatal with the '-servername_fatal'
9612 *Peter Sylvester, Remy Allais, Christophe Renou*
9614 * Whirlpool hash implementation is added.
9618 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9619 bn(64,32). Because of instruction set limitations it doesn't have
9620 any negative impact on performance. This was done mostly in order
9621 to make it possible to share assembler modules, such as bn_mul_mont
9622 implementations, between 32- and 64-bit builds without hassle.
9626 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9627 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9632 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9633 dedicated Montgomery multiplication procedure, is introduced.
9634 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9635 "64-bit" performance on certain 32-bit targets.
9639 * New option SSL_OP_NO_COMP to disable use of compression selectively
9640 in SSL structures. New SSL ctrl to set maximum send fragment size.
9641 Save memory by setting the I/O buffer sizes dynamically instead of
9642 using the maximum available value.
9646 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9647 in addition to the text details.
9651 * Very, very preliminary EXPERIMENTAL support for printing of general
9652 ASN1 structures. This currently produces rather ugly output and doesn't
9653 handle several customised structures at all.
9657 * Integrated support for PVK file format and some related formats such
9658 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9659 these in the 'rsa' and 'dsa' utilities.
9663 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9667 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9668 place for the (very old) "NETSCAPE" format certificates which are now
9669 handled using new ASN1 code equivalents.
9673 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9674 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9675 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9679 * Modify CRL distribution points extension code to print out previously
9680 unsupported fields. Enhance extension setting code to allow setting of
9685 * Add print and set support for Issuing Distribution Point CRL extension.
9689 * Change 'Configure' script to enable Camellia by default.
9696 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9698 * When rejecting SSL/TLS records due to an incorrect version number, never
9699 update s->server with a new major version number. As of
9700 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9701 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9702 the previous behavior could result in a read attempt at NULL when
9703 receiving specific incorrect SSL/TLS records once record payload
9704 protection is active. ([CVE-2010-0740])
9706 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9708 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9709 could be crashed if the relevant tables were not present (e.g. chrooted).
9711 *Tomas Hoger <thoger@redhat.com>*
9713 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9715 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9717 *Martin Olsson, Neel Mehta*
9719 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9720 accommodate for stack sorting, always a write lock!).
9724 * On some versions of WIN32 Heap32Next is very slow. This can cause
9725 excessive delays in the RAND_poll(): over a minute. As a workaround
9726 include a time check in the inner Heap32Next loop too.
9730 * The code that handled flushing of data in SSL/TLS originally used the
9731 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9732 the problem outlined in PR#1949. The fix suggested there however can
9733 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9734 of Apache). So instead simplify the code to flush unconditionally.
9735 This should be fine since flushing with no data to flush is a no op.
9739 * Handle TLS versions 2.0 and later properly and correctly use the
9740 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9741 off ancient servers have a habit of sticking around for a while...
9745 * Modify compression code so it frees up structures without using the
9746 ex_data callbacks. This works around a problem where some applications
9747 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9748 restarting) then use compression (e.g. SSL with compression) later.
9749 This results in significant per-connection memory leaks and
9750 has caused some security issues including CVE-2008-1678 and
9755 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9756 change when encrypting or decrypting.
9760 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9761 connect and renegotiate with servers which do not support RI.
9762 Until RI is more widely deployed this option is enabled by default.
9766 * Add "missing" ssl ctrls to clear options and mode.
9770 * If client attempts to renegotiate and doesn't support RI respond with
9771 a no_renegotiation alert as required by RFC5746. Some renegotiating
9772 TLS clients will continue a connection gracefully when they receive
9773 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9774 waiting for a server hello which it will never receive. Now we treat a
9775 received no_renegotiation alert as a fatal error. This is because
9776 applications requesting a renegotiation might well expect it to succeed
9777 and would have no code in place to handle the server denying it so the
9778 only safe thing to do is to terminate the connection.
9782 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9783 peer supports secure renegotiation and 0 otherwise. Print out peer
9784 renegotiation support in s_client/s_server.
9788 * Replace the highly broken and deprecated SPKAC certification method with
9789 the updated NID creation version. This should correctly handle UTF8.
9793 * Implement RFC5746. Re-enable renegotiation but require the extension
9794 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9795 turns out to be a bad idea. It has been replaced by
9796 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9797 SSL_CTX_set_options(). This is really not recommended unless you
9798 know what you are doing.
9800 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9802 * Fixes to stateless session resumption handling. Use initial_ctx when
9803 issuing and attempting to decrypt tickets in case it has changed during
9804 servername handling. Use a non-zero length session ID when attempting
9805 stateless session resumption: this makes it possible to determine if
9806 a resumption has occurred immediately after receiving server hello
9807 (several places in OpenSSL subtly assume this) instead of later in
9812 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9813 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9814 fixes for a few places where the return code is not checked
9817 *Julia Lawall <julia@diku.dk>*
9819 * Add --strict-warnings option to Configure script to include devteam
9820 warnings in other configurations.
9824 * Add support for --libdir option and LIBDIR variable in makefiles. This
9825 makes it possible to install openssl libraries in locations which
9826 have names other than "lib", for example "/usr/lib64" which some
9829 *Steve Henson, based on patch from Jeremy Utley*
9831 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9832 X690 8.9.12 and can produce some misleading textual output of OIDs.
9834 *Steve Henson, reported by Dan Kaminsky*
9836 * Delete MD2 from algorithm tables. This follows the recommendation in
9837 several standards that it is not used in new applications due to
9838 several cryptographic weaknesses. For binary compatibility reasons
9839 the MD2 API is still compiled in by default.
9843 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9848 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9849 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9852 *Guenter <lists@gknw.net>*
9854 * Fix the server certificate chain building code to use X509_verify_cert(),
9855 it used to have an ad-hoc builder which was unable to cope with anything
9856 other than a simple chain.
9858 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9860 * Don't check self signed certificate signatures in X509_verify_cert()
9861 by default (a flag can override this): it just wastes time without
9862 adding any security. As a useful side effect self signed root CAs
9863 with non-FIPS digests are now usable in FIPS mode.
9867 * In dtls1_process_out_of_seq_message() the check if the current message
9868 is already buffered was missing. For every new message was memory
9869 allocated, allowing an attacker to perform an denial of service attack
9870 with sending out of seq handshake messages until there is no memory
9871 left. Additionally every future message was buffered, even if the
9872 sequence number made no sense and would be part of another handshake.
9873 So only messages with sequence numbers less than 10 in advance will be
9874 buffered. ([CVE-2009-1378])
9876 *Robin Seggelmann, discovered by Daniel Mentz*
9878 * Records are buffered if they arrive with a future epoch to be
9879 processed after finishing the corresponding handshake. There is
9880 currently no limitation to this buffer allowing an attacker to perform
9881 a DOS attack with sending records with future epochs until there is no
9882 memory left. This patch adds the pqueue_size() function to determine
9883 the size of a buffer and limits the record buffer to 100 entries.
9886 *Robin Seggelmann, discovered by Daniel Mentz*
9888 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9889 parent structure is freed. ([CVE-2009-1379])
9893 * Handle non-blocking I/O properly in SSL_shutdown() call.
9895 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9897 * Add `2.5.4.*` OIDs
9899 *Ilya O. <vrghost@gmail.com>*
9901 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9903 * Disable renegotiation completely - this fixes a severe security
9904 problem ([CVE-2009-3555]) at the cost of breaking all
9905 renegotiation. Renegotiation can be re-enabled by setting
9906 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9907 run-time. This is really not recommended unless you know what
9912 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9914 * Don't set val to NULL when freeing up structures, it is freed up by
9915 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9916 zeroing past the valid field. ([CVE-2009-0789])
9918 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9920 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9921 checked correctly. This would allow some invalid signed attributes to
9922 appear to verify correctly. ([CVE-2009-0591])
9924 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9926 * Reject UniversalString and BMPString types with invalid lengths. This
9927 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9928 a legal length. ([CVE-2009-0590])
9932 * Set S/MIME signing as the default purpose rather than setting it
9933 unconditionally. This allows applications to override it at the store
9938 * Permit restricted recursion of ASN1 strings. This is needed in practice
9939 to handle some structures.
9943 * Improve efficiency of mem_gets: don't search whole buffer each time
9946 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9948 * New -hex option for openssl rand.
9952 * Print out UTF8String and NumericString when parsing ASN1.
9956 * Support NumericString type for name components.
9960 * Allow CC in the environment to override the automatically chosen
9961 compiler. Note that nothing is done to ensure flags work with the
9966 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9968 * Properly check EVP_VerifyFinal() and similar return values
9971 *Ben Laurie, Bodo Moeller, Google Security Team*
9973 * Enable TLS extensions by default.
9977 * Allow the CHIL engine to be loaded, whether the application is
9978 multithreaded or not. (This does not release the developer from the
9979 obligation to set up the dynamic locking callbacks.)
9981 *Sander Temme <sander@temme.net>*
9983 * Use correct exit code if there is an error in dgst command.
9985 *Steve Henson; problem pointed out by Roland Dirlewanger*
9987 * Tweak Configure so that you need to say "experimental-jpake" to enable
9988 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9992 * Add experimental JPAKE support, including demo authentication in
9993 s_client and s_server.
9997 * Set the comparison function in v3_addr_canonize().
9999 *Rob Austein <sra@hactrn.net>*
10001 * Add support for XMPP STARTTLS in s_client.
10003 *Philip Paeps <philip@freebsd.org>*
10005 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10006 to ensure that even with this option, only ciphersuites in the
10007 server's preference list will be accepted. (Note that the option
10008 applies only when resuming a session, so the earlier behavior was
10009 just about the algorithm choice for symmetric cryptography.)
10013 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10015 * Fix NULL pointer dereference if a DTLS server received
10016 ChangeCipherSpec as first record ([CVE-2009-1386]).
10020 * Fix a state transition in s3_srvr.c and d1_srvr.c
10021 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10023 *Nagendra Modadugu*
10025 * The fix in 0.9.8c that supposedly got rid of unsafe
10026 double-checked locking was incomplete for RSA blinding,
10027 addressing just one layer of what turns out to have been
10028 doubly unsafe triple-checked locking.
10030 So now fix this for real by retiring the MONT_HELPER macro
10031 in crypto/rsa/rsa_eay.c.
10033 *Bodo Moeller; problem pointed out by Marius Schilder*
10035 * Various precautionary measures:
10037 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10039 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10040 (NB: This would require knowledge of the secret session ticket key
10041 to exploit, in which case you'd be SOL either way.)
10043 - Change bn_nist.c so that it will properly handle input BIGNUMs
10044 outside the expected range.
10046 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10049 *Neel Mehta, Bodo Moeller*
10051 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10052 the load fails. Useful for distros.
10054 *Ben Laurie and the FreeBSD team*
10056 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10060 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10064 * Expand ENGINE to support engine supplied SSL client certificate functions.
10066 This work was sponsored by Logica.
10070 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10071 keystores. Support for SSL/TLS client authentication too.
10072 Not compiled unless enable-capieng specified to Configure.
10074 This work was sponsored by Logica.
10078 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10079 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10080 attribute creation routines such as certificate requests and PKCS#12
10085 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10087 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10088 handshake which could lead to a client crash as found using the
10089 Codenomicon TLS test suite ([CVE-2008-1672])
10091 *Steve Henson, Mark Cox*
10093 * Fix double free in TLS server name extensions which could lead to
10094 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10098 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10100 Clear the error queue to ensure that error entries left from
10101 older function calls do not interfere with the correct operation.
10103 *Lutz Jaenicke, Erik de Castro Lopo*
10105 * Remove root CA certificates of commercial CAs:
10107 The OpenSSL project does not recommend any specific CA and does not
10108 have any policy with respect to including or excluding any CA.
10109 Therefore, it does not make any sense to ship an arbitrary selection
10110 of root CA certificates with the OpenSSL software.
10114 * RSA OAEP patches to fix two separate invalid memory reads.
10115 The first one involves inputs when 'lzero' is greater than
10116 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10117 before the beginning of from). The second one involves inputs where
10118 the 'db' section contains nothing but zeroes (there is a one-byte
10119 invalid read after the end of 'db').
10121 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10123 * Partial backport from 0.9.9-dev:
10125 Introduce bn_mul_mont (dedicated Montgomery multiplication
10126 procedure) as a candidate for BIGNUM assembler implementation.
10127 While 0.9.9-dev uses assembler for various architectures, only
10128 x86_64 is available by default here in the 0.9.8 branch, and
10129 32-bit x86 is available through a compile-time setting.
10131 To try the 32-bit x86 assembler implementation, use Configure
10132 option "enable-montasm" (which exists only for this backport).
10134 As "enable-montasm" for 32-bit x86 disclaims code stability
10135 anyway, in this constellation we activate additional code
10136 backported from 0.9.9-dev for further performance improvements,
10137 namely BN_from_montgomery_word. (To enable this otherwise,
10138 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10140 *Andy Polyakov (backport partially by Bodo Moeller)*
10142 * Add TLS session ticket callback. This allows an application to set
10143 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10144 values. This is useful for key rollover for example where several key
10145 sets may exist with different names.
10149 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10150 This was broken until now in 0.9.8 releases, such that the only way
10151 a registered ENGINE could be used (assuming it initialises
10152 successfully on the host) was to explicitly set it as the default
10153 for the relevant algorithms. This is in contradiction with 0.9.7
10154 behaviour and the documentation. With this fix, when an ENGINE is
10155 registered into a given algorithm's table of implementations, the
10156 'uptodate' flag is reset so that auto-discovery will be used next
10157 time a new context for that algorithm attempts to select an
10160 *Ian Lister (tweaked by Geoff Thorpe)*
10162 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10163 implementation in the following ways:
10165 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10168 Lack of BER streaming support means one pass streaming processing is
10169 only supported if data is detached: setting the streaming flag is
10170 ignored for embedded content.
10172 CMS support is disabled by default and must be explicitly enabled
10173 with the enable-cms configuration option.
10177 * Update the GMP engine glue to do direct copies between BIGNUM and
10178 mpz_t when openssl and GMP use the same limb size. Otherwise the
10179 existing "conversion via a text string export" trick is still used.
10181 *Paul Sheer <paulsheer@gmail.com>*
10183 * Zlib compression BIO. This is a filter BIO which compressed and
10184 uncompresses any data passed through it.
10188 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10189 RFC3394 compatible AES key wrapping.
10193 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10194 sets string data without copying. X509_ALGOR_set0() and
10195 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10196 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10197 from an X509_ATTRIBUTE structure optionally checking it occurs only
10198 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10203 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10204 to get the expected BN_FLG_CONSTTIME behavior.
10206 *Bodo Moeller (Google)*
10210 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10211 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10212 - added some more tests to do_tests.pl
10213 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10214 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10215 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10216 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10217 - various changes to netware.pl to enable gcc-cross builds on Win32
10219 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10220 - various changes to fix missing prototype warnings
10221 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10222 - added AES, WHIRLPOOL and CPUID assembler code to build files
10223 - added missing AES assembler make rules to mk1mf.pl
10224 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10226 *Guenter Knauf <eflash@gmx.net>*
10228 * Implement certificate status request TLS extension defined in RFC3546.
10229 A client can set the appropriate parameters and receive the encoded
10230 OCSP response via a callback. A server can query the supplied parameters
10231 and set the encoded OCSP response in the callback. Add simplified examples
10232 to s_client and s_server.
10236 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10238 * Fix various bugs:
10239 + Binary incompatibility of ssl_ctx_st structure
10240 + DTLS interoperation with non-compliant servers
10241 + Don't call get_session_cb() without proposed session
10242 + Fix ia64 assembler code
10244 *Andy Polyakov, Steve Henson*
10246 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10248 * DTLS Handshake overhaul. There were longstanding issues with
10249 OpenSSL DTLS implementation, which were making it impossible for
10250 RFC 4347 compliant client to communicate with OpenSSL server.
10251 Unfortunately just fixing these incompatibilities would "cut off"
10252 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10253 server keeps tolerating non RFC compliant syntax. The opposite is
10254 not true, 0.9.8f client can not communicate with earlier server.
10255 This update even addresses CVE-2007-4995.
10259 * Changes to avoid need for function casts in OpenSSL: some compilers
10260 (gcc 4.2 and later) reject their use.
10261 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10264 * Add RFC4507 support to OpenSSL. This includes the corrections in
10265 RFC4507bis. The encrypted ticket format is an encrypted encoded
10266 SSL_SESSION structure, that way new session features are automatically
10269 If a client application caches session in an SSL_SESSION structure
10270 support is transparent because tickets are now stored in the encoded
10273 The SSL_CTX structure automatically generates keys for ticket
10274 protection in servers so again support should be possible
10275 with no application modification.
10277 If a client or server wishes to disable RFC4507 support then the option
10278 SSL_OP_NO_TICKET can be set.
10280 Add a TLS extension debugging callback to allow the contents of any client
10281 or server extensions to be examined.
10283 This work was sponsored by Google.
10287 * Add initial support for TLS extensions, specifically for the server_name
10288 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10289 have new members for a hostname. The SSL data structure has an
10290 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10291 stored in that context to allow for session resumption, even after the
10292 SSL has been switched to a new SSL_CTX in reaction to a client's
10293 server_name extension.
10295 New functions (subject to change):
10297 SSL_get_servername()
10298 SSL_get_servername_type()
10301 New CTRL codes and macros (subject to change):
10303 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10304 - SSL_CTX_set_tlsext_servername_callback()
10305 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10306 - SSL_CTX_set_tlsext_servername_arg()
10307 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10309 openssl s_client has a new '-servername ...' option.
10311 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10312 '-key2 ...', '-servername_fatal' (subject to change). This allows
10313 testing the HostName extension for a specific single hostname ('-cert'
10314 and '-key' remain fallbacks for handshakes without HostName
10315 negotiation). If the unrecognized_name alert has to be sent, this by
10316 default is a warning; it becomes fatal with the '-servername_fatal'
10319 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10321 * Add AES and SSE2 assembly language support to VC++ build.
10325 * Mitigate attack on final subtraction in Montgomery reduction.
10329 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10330 (which previously caused an internal error).
10334 * Squeeze another 10% out of IGE mode when in != out.
10338 * AES IGE mode speedup.
10340 *Dean Gaudet (Google)*
10342 * Add the Korean symmetric 128-bit cipher SEED (see
10343 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10344 add SEED ciphersuites from RFC 4162:
10346 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10347 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10348 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10349 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10351 To minimize changes between patchlevels in the OpenSSL 0.9.8
10352 series, SEED remains excluded from compilation unless OpenSSL
10353 is configured with 'enable-seed'.
10355 *KISA, Bodo Moeller*
10357 * Mitigate branch prediction attacks, which can be practical if a
10358 single processor is shared, allowing a spy process to extract
10359 information. For detailed background information, see
10360 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10361 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10362 and Necessary Software Countermeasures"). The core of the change
10363 are new versions BN_div_no_branch() and
10364 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10365 respectively, which are slower, but avoid the security-relevant
10366 conditional branches. These are automatically called by BN_div()
10367 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10368 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10369 remove a conditional branch.
10371 BN_FLG_CONSTTIME is the new name for the previous
10372 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10373 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10374 in the exponent causes BN_mod_exp_mont() to use the alternative
10375 implementation in BN_mod_exp_mont_consttime().) The old name
10376 remains as a deprecated alias.
10378 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10379 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10380 constant-time implementations for more than just exponentiation.
10381 Here too the old name is kept as a deprecated alias.
10383 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10384 the BN_BLINDING structure gets an independent copy of the
10385 modulus. This means that the previous `BIGNUM *m` argument to
10386 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10387 essentially becomes `const BIGNUM *m`, although we can't actually
10388 change this in the header file before 0.9.9. It allows
10389 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10390 enable BN_FLG_CONSTTIME.
10392 *Matthew D Wood (Intel Corp)*
10394 * In the SSL/TLS server implementation, be strict about session ID
10395 context matching (which matters if an application uses a single
10396 external cache for different purposes). Previously,
10397 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10398 set. This did ensure strict client verification, but meant that,
10399 with applications using a single external cache for quite
10400 different requirements, clients could circumvent ciphersuite
10401 restrictions for a given session ID context by starting a session
10402 in a different context.
10406 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10407 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10408 authentication-only ciphersuites.
10412 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10413 not complete and could lead to a possible single byte overflow
10414 ([CVE-2007-5135]) [Ben Laurie]
10416 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10418 * Since AES128 and AES256 (and similarly Camellia128 and
10419 Camellia256) share a single mask bit in the logic of
10420 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10421 kludge to work properly if AES128 is available and AES256 isn't
10422 (or if Camellia128 is available and Camellia256 isn't).
10426 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10427 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10428 When a point or a seed is encoded in a BIT STRING, we need to
10429 prevent the removal of trailing zero bits to get the proper DER
10430 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10431 of a NamedBitList, for which trailing 0 bits need to be removed.)
10435 * Have SSL/TLS server implementation tolerate "mismatched" record
10436 protocol version while receiving ClientHello even if the
10437 ClientHello is fragmented. (The server can't insist on the
10438 particular protocol version it has chosen before the ServerHello
10439 message has informed the client about his choice.)
10443 * Add RFC 3779 support.
10445 *Rob Austein for ARIN, Ben Laurie*
10447 * Load error codes if they are not already present instead of using a
10448 static variable. This allows them to be cleanly unloaded and reloaded.
10449 Improve header file function name parsing.
10453 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10454 or CAPABILITY handshake as required by RFCs.
10456 *Goetz Babin-Ebell*
10458 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10460 * Introduce limits to prevent malicious keys being able to
10461 cause a denial of service. ([CVE-2006-2940])
10463 *Steve Henson, Bodo Moeller*
10465 * Fix ASN.1 parsing of certain invalid structures that can result
10466 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10468 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10469 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10471 * Fix SSL client code which could crash if connecting to a
10472 malicious SSLv2 server. ([CVE-2006-4343])
10474 *Tavis Ormandy and Will Drewry, Google Security Team*
10476 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10477 match only those. Before that, "AES256-SHA" would be interpreted
10478 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10479 the same strength classification in 0.9.7h) as we currently only
10480 have a single AES bit in the ciphersuite description bitmap.
10481 That change, however, also applied to ciphersuite strings such as
10482 "RC4-MD5" that intentionally matched multiple ciphersuites --
10483 namely, SSL 2.0 ciphersuites in addition to the more common ones
10484 from SSL 3.0/TLS 1.0.
10486 So we change the selection algorithm again: Naming an explicit
10487 ciphersuite selects this one ciphersuite, and any other similar
10488 ciphersuite (same bitmap) from *other* protocol versions.
10489 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10490 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10492 Since SSL 2.0 does not have any ciphersuites for which the
10493 128/256 bit distinction would be relevant, this works for now.
10494 The proper fix will be to use different bits for AES128 and
10495 AES256, which would have avoided the problems from the beginning;
10496 however, bits are scarce, so we can only do this in a new release
10497 (not just a patchlevel) when we can change the SSL_CIPHER
10498 definition to split the single 'unsigned long mask' bitmap into
10499 multiple values to extend the available space.
10503 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10505 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10506 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10508 * Add AES IGE and biIGE modes.
10512 * Change the Unix randomness entropy gathering to use poll() when
10513 possible instead of select(), since the latter has some
10514 undesirable limitations.
10516 *Darryl Miles via Richard Levitte and Bodo Moeller*
10518 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10519 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10520 cannot be implicitly activated as part of, e.g., the "AES" alias.
10521 However, please upgrade to OpenSSL 0.9.9[-dev] for
10522 non-experimental use of the ECC ciphersuites to get TLS extension
10523 support, which is required for curve and point format negotiation
10524 to avoid potential handshake problems.
10528 * Disable rogue ciphersuites:
10530 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10531 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10532 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10534 The latter two were purportedly from
10535 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10538 Also deactivate the remaining ciphersuites from
10539 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10540 unofficial, and the ID has long expired.
10544 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10545 dual-core machines) and other potential thread-safety issues.
10549 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10550 versions), which is now available for royalty-free use
10551 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10552 Also, add Camellia TLS ciphersuites from RFC 4132.
10554 To minimize changes between patchlevels in the OpenSSL 0.9.8
10555 series, Camellia remains excluded from compilation unless OpenSSL
10556 is configured with 'enable-camellia'.
10560 * Disable the padding bug check when compression is in use. The padding
10561 bug check assumes the first packet is of even length, this is not
10562 necessarily true if compression is enabled and can result in false
10563 positives causing handshake failure. The actual bug test is ancient
10564 code so it is hoped that implementations will either have fixed it by
10565 now or any which still have the bug do not support compression.
10569 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10571 * When applying a cipher rule check to see if string match is an explicit
10572 cipher suite and only match that one cipher suite if it is.
10576 * Link in manifests for VC++ if needed.
10578 *Austin Ziegler <halostatue@gmail.com>*
10580 * Update support for ECC-based TLS ciphersuites according to
10581 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10582 TLS extensions, which are supported starting with the 0.9.9
10583 branch, not in the OpenSSL 0.9.8 branch).
10587 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10588 opaque EVP_CIPHER_CTX handling.
10592 * Fixes and enhancements to zlib compression code. We now only use
10593 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10594 to conform with the standards mentioned here:
10595 <http://www.zlib.net/DLL_FAQ.txt>
10596 Static zlib linking now works on Windows and the new --with-zlib-include
10597 --with-zlib-lib options to Configure can be used to supply the location
10598 of the headers and library. Gracefully handle case where zlib library
10603 * Several fixes and enhancements to the OID generation code. The old code
10604 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10605 handle numbers larger than ULONG_MAX, truncated printing and had a
10606 non standard OBJ_obj2txt() behaviour.
10610 * Add support for building of engines under engine/ as shared libraries
10611 under VC++ build system.
10615 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10616 Hopefully, we will not see any false combination of paths any more.
10620 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10622 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10623 (part of SSL_OP_ALL). This option used to disable the
10624 countermeasure against man-in-the-middle protocol-version
10625 rollback in the SSL 2.0 server implementation, which is a bad
10626 idea. ([CVE-2005-2969])
10628 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10629 for Information Security, National Institute of Advanced Industrial
10630 Science and Technology [AIST], Japan)*
10632 * Add two function to clear and return the verify parameter flags.
10636 * Keep cipherlists sorted in the source instead of sorting them at
10637 runtime, thus removing the need for a lock.
10641 * Avoid some small subgroup attacks in Diffie-Hellman.
10643 *Nick Mathewson and Ben Laurie*
10645 * Add functions for well-known primes.
10649 * Extended Windows CE support.
10651 *Satoshi Nakamura and Andy Polyakov*
10653 * Initialize SSL_METHOD structures at compile time instead of during
10654 runtime, thus removing the need for a lock.
10658 * Make PKCS7_decrypt() work even if no certificate is supplied by
10659 attempting to decrypt each encrypted key in turn. Add support to
10664 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10666 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10669 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10673 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10674 key into the same file any more.
10678 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10682 * Add -utf8 command line and config file option to 'ca'.
10684 *Stefan <stf@udoma.org*
10686 * Removed the macro des_crypt(), as it seems to conflict with some
10687 libraries. Use DES_crypt().
10691 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10692 involves renaming the source and generated shared-libs for
10693 both. The engines will accept the corrected or legacy ids
10694 ('ncipher' and '4758_cca' respectively) when binding. NB,
10695 this only applies when building 'shared'.
10697 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10699 * Add attribute functions to EVP_PKEY structure. Modify
10700 PKCS12_create() to recognize a CSP name attribute and
10701 use it. Make -CSP option work again in pkcs12 utility.
10705 * Add new functionality to the bn blinding code:
10706 - automatic re-creation of the BN_BLINDING parameters after
10707 a fixed number of uses (currently 32)
10708 - add new function for parameter creation
10709 - introduce flags to control the update behaviour of the
10710 BN_BLINDING parameters
10711 - hide BN_BLINDING structure
10712 Add a second BN_BLINDING slot to the RSA structure to improve
10713 performance when a single RSA object is shared among several
10718 * Add support for DTLS.
10720 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10722 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10723 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10727 * Remove buggy and incomplete DH cert support from
10728 ssl/ssl_rsa.c and ssl/s3_both.c
10732 * Use SHA-1 instead of MD5 as the default digest algorithm for
10733 the `apps/openssl` commands.
10737 * Compile clean with "-Wall -Wmissing-prototypes
10738 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10739 DEBUG_SAFESTACK must also be set.
10743 * Change ./Configure so that certain algorithms can be disabled by default.
10744 The new counterpiece to "no-xxx" is "enable-xxx".
10746 The patented RC5 and MDC2 algorithms will now be disabled unless
10747 "enable-rc5" and "enable-mdc2", respectively, are specified.
10749 (IDEA remains enabled despite being patented. This is because IDEA
10750 is frequently required for interoperability, and there is no license
10751 fee for non-commercial use. As before, "no-idea" can be used to
10752 avoid this algorithm.)
10756 * Add processing of proxy certificates (see RFC 3820). This work was
10757 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10758 EGEE (Enabling Grids for E-science in Europe).
10762 * RC4 performance overhaul on modern architectures/implementations, such
10763 as Intel P4, IA-64 and AMD64.
10767 * New utility extract-section.pl. This can be used specify an alternative
10768 section number in a pod file instead of having to treat each file as
10769 a separate case in Makefile. This can be done by adding two lines to the
10772 =for comment openssl_section:XXX
10774 The blank line is mandatory.
10778 * New arguments -certform, -keyform and -pass for s_client and s_server
10779 to allow alternative format key and certificate files and passphrase
10784 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10785 update associated structures and add various utility functions.
10787 Add new policy related verify parameters, include policy checking in
10788 standard verify code. Enhance 'smime' application with extra parameters
10789 to support policy checking and print out.
10793 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10794 Nehemiah processors. These extensions support AES encryption in hardware
10795 as well as RNG (though RNG support is currently disabled).
10797 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10799 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10803 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10805 *Andy Polyakov and a number of other people*
10807 * Improved PowerPC platform support. Most notably BIGNUM assembler
10808 implementation contributed by IBM.
10810 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10812 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10813 exponent rather than 'unsigned long'. There is a corresponding change to
10814 the new 'rsa_keygen' element of the RSA_METHOD structure.
10816 *Jelte Jansen, Geoff Thorpe*
10818 * Functionality for creating the initial serial number file is now
10819 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10821 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10822 number file to 1, which is bound to cause problems. To avoid
10823 the problems while respecting compatibility between different 0.9.7
10824 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10825 CA.pl for serial number initialization. With the new release 0.9.8,
10826 we can fix the problem directly in the 'ca' utility.)
10830 * Reduced header interdependencies by declaring more opaque objects in
10831 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10832 give fewer recursive includes, which could break lazy source code - so
10833 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10834 developers should define this symbol when building and using openssl to
10835 ensure they track the recommended behaviour, interfaces, [etc], but
10836 backwards-compatible behaviour prevails when this isn't defined.
10840 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10844 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10845 This will generate a random key of the appropriate length based on the
10846 cipher context. The EVP_CIPHER can provide its own random key generation
10847 routine to support keys of a specific form. This is used in the des and
10848 3des routines to generate a key of the correct parity. Update S/MIME
10849 code to use new functions and hence generate correct parity DES keys.
10850 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10851 valid (weak or incorrect parity).
10855 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10856 as looking them up. This is useful when the verified structure may contain
10857 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10858 present unless the new PKCS7_NO_CRL flag is asserted.
10862 * Extend ASN1 oid configuration module. It now additionally accepts the
10865 shortName = some long name, 1.2.3.4
10869 * Reimplemented the BN_CTX implementation. There is now no more static
10870 limitation on the number of variables it can handle nor the depth of the
10871 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10872 information can now expand as required, and rather than having a single
10873 static array of bignums, BN_CTX now uses a linked-list of such arrays
10874 allowing it to expand on demand whilst maintaining the usefulness of
10875 BN_CTX's "bundling".
10879 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10880 to allow all RSA operations to function using a single BN_CTX.
10884 * Preliminary support for certificate policy evaluation and checking. This
10885 is initially intended to pass the tests outlined in "Conformance Testing
10886 of Relying Party Client Certificate Path Processing Logic" v1.07.
10890 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10891 remained unused and not that useful. A variety of other little bignum
10892 tweaks and fixes have also been made continuing on from the audit (see
10897 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10898 associated ASN1, EVP and SSL functions and old ASN1 macros.
10902 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10903 and this should never fail. So the return value from the use of
10904 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10905 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10909 * BN_CTX_get() should return zero-valued bignums, providing the same
10910 initialised value as BN_new().
10912 *Geoff Thorpe, suggested by Ulf Möller*
10914 * Support for inhibitAnyPolicy certificate extension.
10918 * An audit of the BIGNUM code is underway, for which debugging code is
10919 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10920 is considered valid when processing BIGNUMs, and causes execution to
10921 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10922 further steps are taken to deliberately pollute unused data in BIGNUM
10923 structures to try and expose faulty code further on. For now, openssl will
10924 (in its default mode of operation) continue to tolerate the inconsistent
10925 forms that it has tolerated in the past, but authors and packagers should
10926 consider trying openssl and their own applications when compiled with
10927 these debugging symbols defined. It will help highlight potential bugs in
10928 their own code, and will improve the test coverage for OpenSSL itself. At
10929 some point, these tighter rules will become openssl's default to improve
10930 maintainability, though the assert()s and other overheads will remain only
10931 in debugging configurations. See bn.h for more details.
10933 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10935 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10936 that can only be obtained through BN_CTX_new() (which implicitly
10937 initialises it). The presence of this function only made it possible
10938 to overwrite an existing structure (and cause memory leaks).
10942 * Because of the callback-based approach for implementing LHASH as a
10943 template type, lh_insert() adds opaque objects to hash-tables and
10944 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10945 to clean up those corresponding objects before destroying the hash table
10946 (and losing the object pointers). So some over-zealous constifications in
10947 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10948 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10949 prototyped to have "const" restrictions on the object pointers they are
10950 given (and so aren't required to cast them away any more).
10954 * The tmdiff.h API was so ugly and minimal that our own timing utility
10955 (speed) prefers to use its own implementation. The two implementations
10956 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10957 its object type properly exposed (MS_TM) instead of casting to/from
10958 `char *`. This may still change yet if someone realises MS_TM and
10960 aren't necessarily the greatest nomenclatures - but this is what was used
10961 internally to the implementation so I've used that for now.
10965 * Ensure that deprecated functions do not get compiled when
10966 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10967 the self-tests were still using deprecated key-generation functions so
10968 these have been updated also.
10972 * Reorganise PKCS#7 code to separate the digest location functionality
10973 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10974 New function PKCS7_set_digest() to set the digest type for PKCS#7
10975 digestedData type. Add additional code to correctly generate the
10976 digestedData type and add support for this type in PKCS7 initialization
10981 * New function PKCS7_set0_type_other() this initializes a PKCS7
10982 structure of type "other".
10986 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10987 sure the loop does correctly stop and breaking ("division by zero")
10988 modulus operations are not performed. The (pre-generated) prime
10989 table crypto/bn/bn_prime.h was already correct, but it could not be
10990 re-generated on some platforms because of the "division by zero"
10991 situation in the script.
10993 *Ralf S. Engelschall*
10995 * Update support for ECC-based TLS ciphersuites according to
10996 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10997 SHA-1 now is only used for "small" curves (where the
10998 representation of a field element takes up to 24 bytes); for
10999 larger curves, the field element resulting from ECDH is directly
11000 used as premaster secret.
11002 *Douglas Stebila (Sun Microsystems Laboratories)*
11004 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11005 curve secp160r1 to the tests.
11007 *Douglas Stebila (Sun Microsystems Laboratories)*
11009 * Add the possibility to load symbols globally with DSO.
11011 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11013 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11014 control of the error stack.
11018 * Add support for STORE in ENGINE.
11022 * Add the STORE type. The intention is to provide a common interface
11023 to certificate and key stores, be they simple file-based stores, or
11024 HSM-type store, or LDAP stores, or...
11025 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11029 * Add a generic structure called OPENSSL_ITEM. This can be used to
11030 pass a list of arguments to any function as well as provide a way
11031 for a function to pass data back to the caller.
11035 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11036 works like BUF_strdup() but can be used to duplicate a portion of
11037 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11042 * Add the function sk_find_ex() which works like sk_find(), but will
11043 return an index to an element even if an exact match couldn't be
11044 found. The index is guaranteed to point at the element where the
11045 searched-for key would be inserted to preserve sorting order.
11049 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11050 takes an extra flags argument for optional functionality. Currently,
11051 the following flags are defined:
11053 OBJ_BSEARCH_VALUE_ON_NOMATCH
11054 This one gets OBJ_bsearch_ex() to return a pointer to the first
11055 element where the comparing function returns a negative or zero
11058 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11059 This one gets OBJ_bsearch_ex() to return a pointer to the first
11060 element where the comparing function returns zero. This is useful
11061 if there are more than one element where the comparing function
11066 * Make it possible to create self-signed certificates with 'openssl ca'
11067 in such a way that the self-signed certificate becomes part of the
11068 CA database and uses the same mechanisms for serial number generation
11069 as all other certificate signing. The new flag '-selfsign' enables
11070 this functionality. Adapt CA.sh and CA.pl.in.
11074 * Add functionality to check the public key of a certificate request
11075 against a given private. This is useful to check that a certificate
11076 request can be signed by that key (self-signing).
11080 * Make it possible to have multiple active certificates with the same
11081 subject in the CA index file. This is done only if the keyword
11082 'unique_subject' is set to 'no' in the main CA section (default
11083 if 'CA_default') of the configuration file. The value is saved
11084 with the database itself in a separate index attribute file,
11085 named like the index file with '.attr' appended to the name.
11089 * Generate multi-valued AVAs using '+' notation in config files for
11094 * Support for nameConstraints certificate extension.
11098 * Support for policyConstraints certificate extension.
11102 * Support for policyMappings certificate extension.
11106 * Make sure the default DSA_METHOD implementation only uses its
11107 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11108 and change its own handlers to be NULL so as to remove unnecessary
11109 indirection. This lets alternative implementations fallback to the
11110 default implementation more easily.
11114 * Support for directoryName in GeneralName related extensions
11119 * Make it possible to link applications using Makefile.shared.
11120 Make that possible even when linking against static libraries!
11124 * Support for single pass processing for S/MIME signing. This now
11125 means that S/MIME signing can be done from a pipe, in addition
11126 cleartext signing (multipart/signed type) is effectively streaming
11127 and the signed data does not need to be all held in memory.
11129 This is done with a new flag PKCS7_STREAM. When this flag is set
11130 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11131 is done after the data is output (and digests calculated) in
11132 SMIME_write_PKCS7().
11136 * Add full support for -rpath/-R, both in shared libraries and
11137 applications, at least on the platforms where it's known how
11142 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11143 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11144 will now compute a table of multiples of the generator that
11145 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11146 faster (notably in the case of a single point multiplication,
11147 scalar * generator).
11149 *Nils Larsch, Bodo Moeller*
11151 * IPv6 support for certificate extensions. The various extensions
11152 which use the IP:a.b.c.d can now take IPv6 addresses using the
11153 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11158 * Added an ENGINE that implements RSA by performing private key
11159 exponentiations with the GMP library. The conversions to and from
11160 GMP's mpz_t format aren't optimised nor are any montgomery forms
11161 cached, and on x86 it appears OpenSSL's own performance has caught up.
11162 However there are likely to be other architectures where GMP could
11163 provide a boost. This ENGINE is not built in by default, but it can be
11164 specified at Configure time and should be accompanied by the necessary
11165 linker additions, eg;
11166 ./config -DOPENSSL_USE_GMP -lgmp
11170 * "openssl engine" will not display ENGINE/DSO load failure errors when
11171 testing availability of engines with "-t" - the old behaviour is
11172 produced by increasing the feature's verbosity with "-tt".
11176 * ECDSA routines: under certain error conditions uninitialized BN objects
11177 could be freed. Solution: make sure initialization is performed early
11178 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11183 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11184 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11185 software implementations. For DSA and DH, parameter generation can
11186 also be overridden by providing the appropriate method callbacks.
11190 * Change the "progress" mechanism used in key-generation and
11191 primality testing to functions that take a new BN_GENCB pointer in
11192 place of callback/argument pairs. The new API functions have `_ex`
11193 postfixes and the older functions are reimplemented as wrappers for
11194 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11195 declarations of the old functions to help (graceful) attempts to
11196 migrate to the new functions. Also, the new key-generation API
11197 functions operate on a caller-supplied key-structure and return
11198 success/failure rather than returning a key or NULL - this is to
11199 help make "keygen" another member function of RSA_METHOD etc.
11201 Example for using the new callback interface:
11203 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11204 void *my_arg = ...;
11207 BN_GENCB_set(&my_cb, my_callback, my_arg);
11209 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11210 /* For the meaning of a, b in calls to my_callback(), see the
11211 * documentation of the function that calls the callback.
11212 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11213 * my_callback should return 1 if it wants BN_is_prime_ex()
11214 * to continue, or 0 to stop.
11219 * Change the ZLIB compression method to be stateful, and make it
11220 available to TLS with the number defined in
11221 draft-ietf-tls-compression-04.txt.
11225 * Add the ASN.1 structures and functions for CertificatePair, which
11226 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11228 CertificatePair ::= SEQUENCE {
11229 forward [0] Certificate OPTIONAL,
11230 reverse [1] Certificate OPTIONAL,
11231 -- at least one of the pair shall be present -- }
11233 Also implement the PEM functions to read and write certificate
11234 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11236 This needed to be defined, mostly for the sake of the LDAP
11237 attribute crossCertificatePair, but may prove useful elsewhere as
11242 * Make it possible to inhibit symlinking of shared libraries in
11243 Makefile.shared, for Cygwin's sake.
11247 * Extend the BIGNUM API by creating a function
11248 void BN_set_negative(BIGNUM *a, int neg);
11249 and a macro that behave like
11250 int BN_is_negative(const BIGNUM *a);
11252 to avoid the need to access 'a->neg' directly in applications.
11256 * Implement fast modular reduction for pseudo-Mersenne primes
11257 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11258 EC_GROUP_new_curve_GFp() will now automatically use this
11261 *Nils Larsch <nla@trustcenter.de>*
11263 * Add new lock type (CRYPTO_LOCK_BN).
11267 * Change the ENGINE framework to automatically load engines
11268 dynamically from specific directories unless they could be
11269 found to already be built in or loaded. Move all the
11270 current engines except for the cryptodev one to a new
11271 directory engines/.
11272 The engines in engines/ are built as shared libraries if
11273 the "shared" options was given to ./Configure or ./config.
11274 Otherwise, they are inserted in libcrypto.a.
11275 /usr/local/ssl/engines is the default directory for dynamic
11276 engines, but that can be overridden at configure time through
11277 the usual use of --prefix and/or --openssldir, and at run
11278 time with the environment variable OPENSSL_ENGINES.
11280 *Geoff Thorpe and Richard Levitte*
11282 * Add Makefile.shared, a helper makefile to build shared
11283 libraries. Adapt Makefile.org.
11287 * Add version info to Win32 DLLs.
11289 *Peter 'Luna' Runestig" <peter@runestig.com>*
11291 * Add new 'medium level' PKCS#12 API. Certificates and keys
11292 can be added using this API to created arbitrary PKCS#12
11293 files while avoiding the low-level API.
11295 New options to PKCS12_create(), key or cert can be NULL and
11296 will then be omitted from the output file. The encryption
11297 algorithm NIDs can be set to -1 for no encryption, the mac
11298 iteration count can be set to 0 to omit the mac.
11300 Enhance pkcs12 utility by making the -nokeys and -nocerts
11301 options work when creating a PKCS#12 file. New option -nomac
11302 to omit the mac, NONE can be set for an encryption algorithm.
11303 New code is modified to use the enhanced PKCS12_create()
11304 instead of the low-level API.
11308 * Extend ASN1 encoder to support indefinite length constructed
11309 encoding. This can output sequences tags and octet strings in
11310 this form. Modify pk7_asn1.c to support indefinite length
11311 encoding. This is experimental and needs additional code to
11312 be useful, such as an ASN1 bio and some enhanced streaming
11315 Extend template encode functionality so that tagging is passed
11316 down to the template encoder.
11320 * Let 'openssl req' fail if an argument to '-newkey' is not
11321 recognized instead of using RSA as a default.
11325 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11326 As these are not official, they are not included in "ALL";
11327 the "ECCdraft" ciphersuite group alias can be used to select them.
11329 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11331 * Add ECDH engine support.
11333 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11335 * Add ECDH in new directory crypto/ecdh/.
11337 *Douglas Stebila (Sun Microsystems Laboratories)*
11339 * Let BN_rand_range() abort with an error after 100 iterations
11340 without success (which indicates a broken PRNG).
11344 * Change BN_mod_sqrt() so that it verifies that the input value
11345 is really the square of the return value. (Previously,
11346 BN_mod_sqrt would show GIGO behaviour.)
11350 * Add named elliptic curves over binary fields from X9.62, SECG,
11351 and WAP/WTLS; add OIDs that were still missing.
11353 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11355 * Extend the EC library for elliptic curves over binary fields
11356 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11359 EC_GF2m_simple_method
11363 EC_GROUP_new_curve_GF2m
11364 EC_GROUP_set_curve_GF2m
11365 EC_GROUP_get_curve_GF2m
11366 EC_POINT_set_affine_coordinates_GF2m
11367 EC_POINT_get_affine_coordinates_GF2m
11368 EC_POINT_set_compressed_coordinates_GF2m
11370 Point compression for binary fields is disabled by default for
11371 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11374 As binary polynomials are represented as BIGNUMs, various members
11375 of the EC_GROUP and EC_POINT data structures can be shared
11376 between the implementations for prime fields and binary fields;
11377 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11378 are essentially identical to their `..._GFp` counterparts.
11379 (For simplicity, the `..._GFp` prefix has been dropped from
11380 various internal method names.)
11382 An internal 'field_div' method (similar to 'field_mul' and
11383 'field_sqr') has been added; this is used only for binary fields.
11385 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11387 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11388 through methods ('mul', 'precompute_mult').
11390 The generic implementations (now internally called 'ec_wNAF_mul'
11391 and 'ec_wNAF_precomputed_mult') remain the default if these
11392 methods are undefined.
11394 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11396 * New function EC_GROUP_get_degree, which is defined through
11397 EC_METHOD. For curves over prime fields, this returns the bit
11398 length of the modulus.
11400 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11402 * New functions EC_GROUP_dup, EC_POINT_dup.
11403 (These simply call ..._new and ..._copy).
11405 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11407 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11408 Polynomials are represented as BIGNUMs (where the sign bit is not
11409 used) in the following functions [macros]:
11412 BN_GF2m_sub [= BN_GF2m_add]
11413 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11414 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11415 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11417 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11418 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11419 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11420 BN_GF2m_cmp [= BN_ucmp]
11422 (Note that only the 'mod' functions are actually for fields GF(2^m).
11423 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11425 For some functions, an the irreducible polynomial defining a
11426 field can be given as an 'unsigned int[]' with strictly
11427 decreasing elements giving the indices of those bits that are set;
11428 i.e., p[] represents the polynomial
11429 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11431 p[0] > p[1] > ... > p[k] = 0.
11432 This applies to the following functions:
11435 BN_GF2m_mod_mul_arr
11436 BN_GF2m_mod_sqr_arr
11437 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11438 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11439 BN_GF2m_mod_exp_arr
11440 BN_GF2m_mod_sqrt_arr
11441 BN_GF2m_mod_solve_quad_arr
11445 Conversion can be performed by the following functions:
11450 bntest.c has additional tests for binary polynomial arithmetic.
11452 Two implementations for BN_GF2m_mod_div() are available.
11453 The default algorithm simply uses BN_GF2m_mod_inv() and
11454 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11455 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11456 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11458 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11460 * Add new error code 'ERR_R_DISABLED' that can be used when some
11461 functionality is disabled at compile-time.
11463 *Douglas Stebila <douglas.stebila@sun.com>*
11465 * Change default behaviour of 'openssl asn1parse' so that more
11466 information is visible when viewing, e.g., a certificate:
11468 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11469 mode the content of non-printable OCTET STRINGs is output in a
11470 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11471 avoid the appearance of a printable string.
11473 *Nils Larsch <nla@trustcenter.de>*
11475 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11477 EC_GROUP_set_asn1_flag()
11478 EC_GROUP_get_asn1_flag()
11479 EC_GROUP_set_point_conversion_form()
11480 EC_GROUP_get_point_conversion_form()
11481 These control ASN1 encoding details:
11482 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11483 has been set to OPENSSL_EC_NAMED_CURVE.
11484 - Points are encoded in uncompressed form by default; options for
11485 asn1_for are as for point2oct, namely
11486 POINT_CONVERSION_COMPRESSED
11487 POINT_CONVERSION_UNCOMPRESSED
11488 POINT_CONVERSION_HYBRID
11490 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11492 EC_GROUP_set_seed()
11493 EC_GROUP_get0_seed()
11494 EC_GROUP_get_seed_len()
11495 This is used only for ASN1 purposes (so far).
11497 *Nils Larsch <nla@trustcenter.de>*
11499 * Add 'field_type' member to EC_METHOD, which holds the NID
11500 of the appropriate field type OID. The new function
11501 EC_METHOD_get_field_type() returns this value.
11503 *Nils Larsch <nla@trustcenter.de>*
11506 EC_POINT_point2bn()
11507 EC_POINT_bn2point()
11508 EC_POINT_point2hex()
11509 EC_POINT_hex2point()
11510 providing useful interfaces to EC_POINT_point2oct() and
11511 EC_POINT_oct2point().
11513 *Nils Larsch <nla@trustcenter.de>*
11515 * Change internals of the EC library so that the functions
11516 EC_GROUP_set_generator()
11517 EC_GROUP_get_generator()
11518 EC_GROUP_get_order()
11519 EC_GROUP_get_cofactor()
11520 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11521 to methods, which would lead to unnecessary code duplication when
11522 adding different types of curves.
11524 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11526 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11527 arithmetic, and such that modified wNAFs are generated
11528 (which avoid length expansion in many cases).
11532 * Add a function EC_GROUP_check_discriminant() (defined via
11533 EC_METHOD) that verifies that the curve discriminant is non-zero.
11535 Add a function EC_GROUP_check() that makes some sanity tests
11536 on a EC_GROUP, its generator and order. This includes
11537 EC_GROUP_check_discriminant().
11539 *Nils Larsch <nla@trustcenter.de>*
11541 * Add ECDSA in new directory crypto/ecdsa/.
11543 Add applications 'openssl ecparam' and 'openssl ecdsa'
11544 (these are based on 'openssl dsaparam' and 'openssl dsa').
11546 ECDSA support is also included in various other files across the
11547 library. Most notably,
11548 - 'openssl req' now has a '-newkey ecdsa:file' option;
11549 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11550 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11551 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11552 them suitable for ECDSA where domain parameters must be
11553 extracted before the specific public key;
11554 - ECDSA engine support has been added.
11556 *Nils Larsch <nla@trustcenter.de>*
11558 * Include some named elliptic curves, and add OIDs from X9.62,
11559 SECG, and WAP/WTLS. Each curve can be obtained from the new
11561 EC_GROUP_new_by_curve_name(),
11562 and the list of available named curves can be obtained with
11563 EC_get_builtin_curves().
11564 Also add a 'curve_name' member to EC_GROUP objects, which can be
11566 EC_GROUP_set_curve_name()
11567 EC_GROUP_get_curve_name()
11569 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11571 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11572 was actually never needed) and in BN_mul(). The removal in BN_mul()
11573 required a small change in bn_mul_part_recursive() and the addition
11574 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11575 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11576 bn_sub_words() and bn_add_words() except they take arrays with
11581 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11583 * Cleanse PEM buffers before freeing them since they may contain
11586 *Benjamin Bennett <ben@psc.edu>*
11588 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11589 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11590 authentication-only ciphersuites.
11594 * Since AES128 and AES256 share a single mask bit in the logic of
11595 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11596 kludge to work properly if AES128 is available and AES256 isn't.
11600 * Expand security boundary to match 1.1.1 module.
11604 * Remove redundant features: hash file source, editing of test vectors
11605 modify fipsld to use external fips_premain.c signature.
11609 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11610 run algorithm test programs.
11614 * Make algorithm test programs more tolerant of whitespace.
11618 * Have SSL/TLS server implementation tolerate "mismatched" record
11619 protocol version while receiving ClientHello even if the
11620 ClientHello is fragmented. (The server can't insist on the
11621 particular protocol version it has chosen before the ServerHello
11622 message has informed the client about his choice.)
11626 * Load error codes if they are not already present instead of using a
11627 static variable. This allows them to be cleanly unloaded and reloaded.
11631 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11633 * Introduce limits to prevent malicious keys being able to
11634 cause a denial of service. ([CVE-2006-2940])
11636 *Steve Henson, Bodo Moeller*
11638 * Fix ASN.1 parsing of certain invalid structures that can result
11639 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11641 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11642 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11644 * Fix SSL client code which could crash if connecting to a
11645 malicious SSLv2 server. ([CVE-2006-4343])
11647 *Tavis Ormandy and Will Drewry, Google Security Team*
11649 * Change ciphersuite string processing so that an explicit
11650 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11651 will no longer include "AES128-SHA"), and any other similar
11652 ciphersuite (same bitmap) from *other* protocol versions (so that
11653 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11654 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11655 changes from 0.9.8b and 0.9.8d.
11659 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11661 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11662 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11664 * Change the Unix randomness entropy gathering to use poll() when
11665 possible instead of select(), since the latter has some
11666 undesirable limitations.
11668 *Darryl Miles via Richard Levitte and Bodo Moeller*
11670 * Disable rogue ciphersuites:
11672 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11673 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11674 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11676 The latter two were purportedly from
11677 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11680 Also deactivate the remaining ciphersuites from
11681 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11682 unofficial, and the ID has long expired.
11686 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11687 dual-core machines) and other potential thread-safety issues.
11691 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11693 * Adapt fipsld and the build system to link against the validated FIPS
11694 module in FIPS mode.
11698 * Fixes for VC++ 2005 build under Windows.
11702 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11703 from a Windows bash shell such as MSYS. It is autodetected from the
11704 "config" script when run from a VC++ environment. Modify standard VC++
11705 build to use fipscanister.o from the GNU make build.
11709 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11711 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11712 The value now differs depending on if you build for FIPS or not.
11713 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11714 safely run with a non-FIPSed libcrypto, as it may crash because of
11715 the difference induced by this change.
11719 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11721 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11722 (part of SSL_OP_ALL). This option used to disable the
11723 countermeasure against man-in-the-middle protocol-version
11724 rollback in the SSL 2.0 server implementation, which is a bad
11725 idea. ([CVE-2005-2969])
11727 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11728 for Information Security, National Institute of Advanced Industrial
11729 Science and Technology [AIST, Japan)]*
11731 * Minimal support for X9.31 signatures and PSS padding modes. This is
11732 mainly for FIPS compliance and not fully integrated at this stage.
11736 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11737 the exponentiation using a fixed-length exponent. (Otherwise,
11738 the information leaked through timing could expose the secret key
11739 after many signatures; cf. Bleichenbacher's attack on DSA with
11744 * Make a new fixed-window mod_exp implementation the default for
11745 RSA, DSA, and DH private-key operations so that the sequence of
11746 squares and multiplies and the memory access pattern are
11747 independent of the particular secret key. This will mitigate
11748 cache-timing and potential related attacks.
11750 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11751 and this is automatically used by BN_mod_exp_mont() if the new flag
11752 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11753 will use this BN flag for private exponents unless the flag
11754 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11755 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11757 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11759 * Change the client implementation for SSLv23_method() and
11760 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11761 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11762 (Previously, the SSL 2.0 backwards compatible Client Hello
11763 message format would be used even with SSL_OP_NO_SSLv2.)
11767 * Add support for smime-type MIME parameter in S/MIME messages which some
11772 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11773 a threadsafe manner. Modify rsa code to use new function and add calls
11774 to dsa and dh code (which had race conditions before).
11778 * Include the fixed error library code in the C error file definitions
11779 instead of fixing them up at runtime. This keeps the error code
11780 structures constant.
11784 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11786 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11789 * Fixes for newer kerberos headers. NB: the casts are needed because
11790 the 'length' field is signed on one version and unsigned on another
11791 with no (?) obvious way to tell the difference, without these VC++
11792 complains. Also the "definition" of FAR (blank) is no longer included
11793 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11794 some needed definitions.
11798 * Undo Cygwin change.
11802 * Added support for proxy certificates according to RFC 3820.
11803 Because they may be a security thread to unaware applications,
11804 they must be explicitly allowed in run-time. See
11805 docs/HOWTO/proxy_certificates.txt for further information.
11809 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11811 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11812 server and client random values. Previously
11813 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11814 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11816 This change has negligible security impact because:
11818 1. Server and client random values still have 24 bytes of pseudo random
11821 2. Server and client random values are sent in the clear in the initial
11824 3. The master secret is derived using the premaster secret (48 bytes in
11825 size for static RSA ciphersuites) as well as client server and random
11828 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11831 *Stephen Henson, reported by UK NISCC*
11833 * Use Windows randomness collection on Cygwin.
11837 * Fix hang in EGD/PRNGD query when communication socket is closed
11838 prematurely by EGD/PRNGD.
11840 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11842 * Prompt for pass phrases when appropriate for PKCS12 input format.
11846 * Back-port of selected performance improvements from development
11847 branch, as well as improved support for PowerPC platforms.
11851 * Add lots of checks for memory allocation failure, error codes to indicate
11852 failure and freeing up memory if a failure occurs.
11854 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11856 * Add new -passin argument to dgst.
11860 * Perform some character comparisons of different types in X509_NAME_cmp:
11861 this is needed for some certificates that re-encode DNs into UTF8Strings
11862 (in violation of RFC3280) and can't or won't issue name rollover
11867 * Make an explicit check during certificate validation to see that
11868 the CA setting in each certificate on the chain is correct. As a
11869 side effect always do the following basic checks on extensions,
11870 not just when there's an associated purpose to the check:
11872 - if there is an unhandled critical extension (unless the user
11873 has chosen to ignore this fault)
11874 - if the path length has been exceeded (if one is set at all)
11875 - that certain extensions fit the associated purpose (if one has
11880 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11882 * Avoid a race condition when CRLs are checked in a multi threaded
11883 environment. This would happen due to the reordering of the revoked
11884 entries during signature checking and serial number lookup. Now the
11885 encoding is cached and the serial number sort performed under a lock.
11886 Add new STACK function sk_is_sorted().
11890 * Add Delta CRL to the extension code.
11894 * Various fixes to s3_pkt.c so alerts are sent properly.
11896 *David Holmes <d.holmes@f5.com>*
11898 * Reduce the chances of duplicate issuer name and serial numbers (in
11899 violation of RFC3280) using the OpenSSL certificate creation utilities.
11900 This is done by creating a random 64 bit value for the initial serial
11901 number when a serial number file is created or when a self signed
11902 certificate is created using 'openssl req -x509'. The initial serial
11903 number file is created using 'openssl x509 -next_serial' in CA.pl
11904 rather than being initialized to 1.
11908 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11910 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11911 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11913 *Joe Orton, Steve Henson*
11915 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11918 *Joe Orton, Steve Henson*
11920 * Make it possible to have multiple active certificates with the same
11921 subject in the CA index file. This is done only if the keyword
11922 'unique_subject' is set to 'no' in the main CA section (default
11923 if 'CA_default') of the configuration file. The value is saved
11924 with the database itself in a separate index attribute file,
11925 named like the index file with '.attr' appended to the name.
11929 * X509 verify fixes. Disable broken certificate workarounds when
11930 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11931 keyUsage extension present. Don't accept CRLs with unhandled critical
11932 extensions: since verify currently doesn't process CRL extensions this
11933 rejects a CRL with *any* critical extensions. Add new verify error codes
11938 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11939 A clarification of RFC2560 will require the use of OCTET STRINGs and
11940 some implementations cannot handle the current raw format. Since OpenSSL
11941 copies and compares OCSP nonces as opaque blobs without any attempt at
11942 parsing them this should not create any compatibility issues.
11946 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11947 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11948 this HMAC (and other) operations are several times slower than OpenSSL
11953 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11955 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11957 * Use the correct content when signing type "other".
11961 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11963 * Fix various bugs revealed by running the NISCC test suite:
11965 Stop out of bounds reads in the ASN1 code when presented with
11966 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11968 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11970 If verify callback ignores invalid public key errors don't try to check
11971 certificate signature with the NULL public key.
11975 * New -ignore_err option in ocsp application to stop the server
11976 exiting on the first error in a request.
11980 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11981 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11986 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11987 extra data after the compression methods not only for TLS 1.0
11988 but also for SSL 3.0 (as required by the specification).
11990 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11992 * Change X509_certificate_type() to mark the key as exported/exportable
11993 when it's 512 *bits* long, not 512 bytes.
11997 * Change AES_cbc_encrypt() so it outputs exact multiple of
11998 blocks during encryption.
12002 * Various fixes to base64 BIO and non blocking I/O. On write
12003 flushes were not handled properly if the BIO retried. On read
12004 data was not being buffered properly and had various logic bugs.
12005 This also affects blocking I/O when the data being decoded is a
12010 * Various S/MIME bugfixes and compatibility changes:
12011 output correct application/pkcs7 MIME type if
12012 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12013 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12014 of files as .eml work). Correctly handle very long lines in MIME
12019 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12021 * Countermeasure against the Klima-Pokorny-Rosa extension of
12022 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12023 a protocol version number mismatch like a decryption error
12024 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12028 * Turn on RSA blinding by default in the default implementation
12029 to avoid a timing attack. Applications that don't want it can call
12030 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12031 They would be ill-advised to do so in most cases.
12033 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12035 * Change RSA blinding code so that it works when the PRNG is not
12036 seeded (in this case, the secret RSA exponent is abused as
12037 an unpredictable seed -- if it is not unpredictable, there
12038 is no point in blinding anyway). Make RSA blinding thread-safe
12039 by remembering the creator's thread ID in rsa->blinding and
12040 having all other threads use local one-time blinding factors
12041 (this requires more computation than sharing rsa->blinding, but
12042 avoids excessive locking; and if an RSA object is not shared
12043 between threads, blinding will still be very fast).
12047 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12048 ENGINE as defaults for all supported algorithms irrespective of
12049 the 'flags' parameter. 'flags' is now honoured, so applications
12050 should make sure they are passing it correctly.
12054 * Target "mingw" now allows native Windows code to be generated in
12055 the Cygwin environment as well as with the MinGW compiler.
12059 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12061 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12062 via timing by performing a MAC computation even if incorrect
12063 block cipher padding has been found. This is a countermeasure
12064 against active attacks where the attacker has to distinguish
12065 between bad padding and a MAC verification error. ([CVE-2003-0078])
12067 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12068 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12069 Martin Vuagnoux (EPFL, Ilion)*
12071 * Make the no-err option work as intended. The intention with no-err
12072 is not to have the whole error stack handling routines removed from
12073 libcrypto, it's only intended to remove all the function name and
12074 reason texts, thereby removing some of the footprint that may not
12075 be interesting if those errors aren't displayed anyway.
12077 NOTE: it's still possible for any application or module to have its
12078 own set of error texts inserted. The routines are there, just not
12079 used by default when no-err is given.
12083 * Add support for FreeBSD on IA64.
12085 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12087 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12088 Kerberos function mit_des_cbc_cksum(). Before this change,
12089 the value returned by DES_cbc_cksum() was like the one from
12090 mit_des_cbc_cksum(), except the bytes were swapped.
12092 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12094 * Allow an application to disable the automatic SSL chain building.
12095 Before this a rather primitive chain build was always performed in
12096 ssl3_output_cert_chain(): an application had no way to send the
12097 correct chain if the automatic operation produced an incorrect result.
12099 Now the chain builder is disabled if either:
12101 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12103 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12105 The reasoning behind this is that an application would not want the
12106 auto chain building to take place if extra chain certificates are
12107 present and it might also want a means of sending no additional
12108 certificates (for example the chain has two certificates and the
12113 * Add the possibility to build without the ENGINE framework.
12115 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12117 * Under Win32 gmtime() can return NULL: check return value in
12118 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12122 * DSA routines: under certain error conditions uninitialized BN objects
12123 could be freed. Solution: make sure initialization is performed early
12124 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12125 Nils Larsch <nla@trustcenter.de> via PR#459)
12129 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12130 checked on reconnect on the client side, therefore session resumption
12131 could still fail with a "ssl session id is different" error. This
12132 behaviour is masked when SSL_OP_ALL is used due to
12133 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12134 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12135 followup to PR #377.
12139 * IA-32 assembler support enhancements: unified ELF targets, support
12140 for SCO/Caldera platforms, fix for Cygwin shared build.
12144 * Add support for FreeBSD on sparc64. As a consequence, support for
12145 FreeBSD on non-x86 processors is separate from x86 processors on
12146 the config script, much like the NetBSD support.
12148 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12150 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12152 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12155 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12156 code (06) was taken as the first octet of the session ID and the last
12157 octet was ignored consequently. As a result SSLv2 client side session
12158 caching could not have worked due to the session ID mismatch between
12160 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12165 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12166 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12171 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12172 seems that in spite of existing for more than a year, many application
12173 author have done nothing to provide the necessary callbacks, which
12174 means that this particular engine will not work properly anywhere.
12175 This is a very unfortunate situation which forces us, in the name
12176 of usability, to give the hw_ncipher.c a static lock, which is part
12178 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12179 appear in 0.9.8 or later. We EXPECT application authors to have
12180 dealt properly with this when 0.9.8 is released (unless we actually
12181 make such changes in the libcrypto locking code that changes will
12182 have to be made anyway).
12186 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12187 octets have been read, EOF or an error occurs. Without this change
12188 some truncated ASN1 structures will not produce an error.
12192 * Disable Heimdal support, since it hasn't been fully implemented.
12193 Still give the possibility to force the use of Heimdal, but with
12194 warnings and a request that patches get sent to openssl-dev.
12198 * Add the VC-CE target, introduce the WINCE sysname, and add
12199 INSTALL.WCE and appropriate conditionals to make it build.
12201 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12203 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12204 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12205 edit numbers of the version.
12207 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12209 * Introduce safe string copy and catenation functions
12210 (BUF_strlcpy() and BUF_strlcat()).
12212 *Ben Laurie (CHATS) and Richard Levitte*
12214 * Avoid using fixed-size buffers for one-line DNs.
12216 *Ben Laurie (CHATS)*
12218 * Add BUF_MEM_grow_clean() to avoid information leakage when
12219 resizing buffers containing secrets, and use where appropriate.
12221 *Ben Laurie (CHATS)*
12223 * Avoid using fixed size buffers for configuration file location.
12225 *Ben Laurie (CHATS)*
12227 * Avoid filename truncation for various CA files.
12229 *Ben Laurie (CHATS)*
12231 * Use sizeof in preference to magic numbers.
12233 *Ben Laurie (CHATS)*
12235 * Avoid filename truncation in cert requests.
12237 *Ben Laurie (CHATS)*
12239 * Add assertions to check for (supposedly impossible) buffer
12242 *Ben Laurie (CHATS)*
12244 * Don't cache truncated DNS entries in the local cache (this could
12245 potentially lead to a spoofing attack).
12247 *Ben Laurie (CHATS)*
12249 * Fix various buffers to be large enough for hex/decimal
12250 representations in a platform independent manner.
12252 *Ben Laurie (CHATS)*
12254 * Add CRYPTO_realloc_clean() to avoid information leakage when
12255 resizing buffers containing secrets, and use where appropriate.
12257 *Ben Laurie (CHATS)*
12259 * Add BIO_indent() to avoid much slightly worrying code to do
12262 *Ben Laurie (CHATS)*
12264 * Convert sprintf()/BIO_puts() to BIO_printf().
12266 *Ben Laurie (CHATS)*
12268 * buffer_gets() could terminate with the buffer only half
12271 *Ben Laurie (CHATS)*
12273 * Add assertions to prevent user-supplied crypto functions from
12274 overflowing internal buffers by having large block sizes, etc.
12276 *Ben Laurie (CHATS)*
12278 * New OPENSSL_assert() macro (similar to assert(), but enabled
12281 *Ben Laurie (CHATS)*
12283 * Eliminate unused copy of key in RC4.
12285 *Ben Laurie (CHATS)*
12287 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12289 *Ben Laurie (CHATS)*
12291 * Fix off-by-one error in EGD path.
12293 *Ben Laurie (CHATS)*
12295 * If RANDFILE path is too long, ignore instead of truncating.
12297 *Ben Laurie (CHATS)*
12299 * Eliminate unused and incorrectly sized X.509 structure
12302 *Ben Laurie (CHATS)*
12304 * Eliminate unused and dangerous function knumber().
12306 *Ben Laurie (CHATS)*
12308 * Eliminate unused and dangerous structure, KSSL_ERR.
12310 *Ben Laurie (CHATS)*
12312 * Protect against overlong session ID context length in an encoded
12313 session object. Since these are local, this does not appear to be
12316 *Ben Laurie (CHATS)*
12318 * Change from security patch (see 0.9.6e below) that did not affect
12319 the 0.9.6 release series:
12321 Remote buffer overflow in SSL3 protocol - an attacker could
12322 supply an oversized master key in Kerberos-enabled versions.
12325 *Ben Laurie (CHATS)*
12327 * Change the SSL kerb5 codes to match RFC 2712.
12331 * Make -nameopt work fully for req and add -reqopt switch.
12333 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12335 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12337 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12339 * Make sure tests can be performed even if the corresponding algorithms
12340 have been removed entirely. This was also the last step to make
12341 OpenSSL compilable with DJGPP under all reasonable conditions.
12343 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12345 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12346 to allow version independent disabling of normally unselected ciphers,
12347 which may be activated as a side-effect of selecting a single cipher.
12349 (E.g., cipher list string "RSA" enables ciphersuites that are left
12350 out of "ALL" because they do not provide symmetric encryption.
12351 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12353 *Lutz Jaenicke, Bodo Moeller*
12355 * Add appropriate support for separate platform-dependent build
12356 directories. The recommended way to make a platform-dependent
12357 build directory is the following (tested on Linux), maybe with
12360 # Place yourself outside of the OpenSSL source tree. In
12361 # this example, the environment variable OPENSSL_SOURCE
12362 # is assumed to contain the absolute OpenSSL source directory.
12363 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12364 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12365 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12366 mkdir -p `dirname $F`
12367 ln -s $OPENSSL_SOURCE/$F $F
12370 To be absolutely sure not to disturb the source tree, a "make clean"
12371 is a good thing. If it isn't successful, don't worry about it,
12372 it probably means the source directory is very clean.
12376 * Make sure any ENGINE control commands make local copies of string
12377 pointers passed to them whenever necessary. Otherwise it is possible
12378 the caller may have overwritten (or deallocated) the original string
12379 data when a later ENGINE operation tries to use the stored values.
12381 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12383 * Improve diagnostics in file reading and command-line digests.
12385 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12387 * Add AES modes CFB and OFB to the object database. Correct an
12388 error in AES-CFB decryption.
12392 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12393 allows existing EVP_CIPHER_CTX structures to be reused after
12394 calling `EVP_*Final()`. This behaviour is used by encryption
12395 BIOs and some applications. This has the side effect that
12396 applications must explicitly clean up cipher contexts with
12397 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12401 * Check the values of dna and dnb in bn_mul_recursive before calling
12402 bn_mul_comba (a non zero value means the a or b arrays do not contain
12403 n2 elements) and fallback to bn_mul_normal if either is not zero.
12407 * Fix escaping of non-ASCII characters when using the -subj option
12408 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12412 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12413 form for "surname", serialNumber has no short form.
12414 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12415 therefore remove "mail" short name for "internet 7".
12416 The OID for unique identifiers in X509 certificates is
12417 x500UniqueIdentifier, not uniqueIdentifier.
12418 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12422 * Add an "init" command to the ENGINE config module and auto initialize
12423 ENGINEs. Without any "init" command the ENGINE will be initialized
12424 after all ctrl commands have been executed on it. If init=1 the
12425 ENGINE is initialized at that point (ctrls before that point are run
12426 on the uninitialized ENGINE and after on the initialized one). If
12427 init=0 then the ENGINE will not be initialized at all.
12431 * Fix the 'app_verify_callback' interface so that the user-defined
12432 argument is actually passed to the callback: In the
12433 SSL_CTX_set_cert_verify_callback() prototype, the callback
12434 declaration has been changed from
12437 int (*cb)(X509_STORE_CTX *,void *);
12438 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12439 i=s->ctx->app_verify_callback(&ctx)
12440 has been changed into
12441 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12443 To update applications using SSL_CTX_set_cert_verify_callback(),
12444 a dummy argument can be added to their callback functions.
12446 *D. K. Smetters <smetters@parc.xerox.com>*
12448 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12450 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12452 * Add and OPENSSL_LOAD_CONF define which will cause
12453 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12454 This allows older applications to transparently support certain
12455 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12456 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12457 load the config file and OPENSSL_add_all_algorithms_conf() which will
12458 always load it have also been added.
12462 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12463 Adjust NIDs and EVP layer.
12465 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12467 * Config modules support in openssl utility.
12469 Most commands now load modules from the config file,
12470 though in a few (such as version) this isn't done
12471 because it couldn't be used for anything.
12473 In the case of ca and req the config file used is
12474 the same as the utility itself: that is the -config
12475 command line option can be used to specify an
12480 * Move default behaviour from OPENSSL_config(). If appname is NULL
12481 use "openssl_conf" if filename is NULL use default openssl config file.
12485 * Add an argument to OPENSSL_config() to allow the use of an alternative
12486 config section name. Add a new flag to tolerate a missing config file
12487 and move code to CONF_modules_load_file().
12491 * Support for crypto accelerator cards from Accelerated Encryption
12492 Processing, www.aep.ie. (Use engine 'aep')
12493 The support was copied from 0.9.6c [engine] and adapted/corrected
12494 to work with the new engine framework.
12496 *AEP Inc. and Richard Levitte*
12498 * Support for SureWare crypto accelerator cards from Baltimore
12499 Technologies. (Use engine 'sureware')
12500 The support was copied from 0.9.6c [engine] and adapted
12501 to work with the new engine framework.
12505 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12506 make the newer ENGINE framework commands for the CHIL engine work.
12508 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12510 * Make it possible to produce shared libraries on ReliantUNIX.
12512 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12514 * Add the configuration target debug-linux-ppro.
12515 Make 'openssl rsa' use the general key loading routines
12516 implemented in `apps.c`, and make those routines able to
12517 handle the key format FORMAT_NETSCAPE and the variant
12520 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12522 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12524 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12526 * Add -keyform to rsautl, and document -engine.
12528 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12530 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12531 BIO_R_NO_SUCH_FILE error code rather than the generic
12532 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12536 * Add new functions
12537 ERR_peek_last_error
12538 ERR_peek_last_error_line
12539 ERR_peek_last_error_line_data.
12540 These are similar to
12542 ERR_peek_error_line
12543 ERR_peek_error_line_data,
12544 but report on the latest error recorded rather than the first one
12545 still in the error queue.
12547 *Ben Laurie, Bodo Moeller*
12549 * default_algorithms option in ENGINE config module. This allows things
12551 default_algorithms = ALL
12552 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12556 * Preliminary ENGINE config module.
12560 * New experimental application configuration code.
12564 * Change the AES code to follow the same name structure as all other
12565 symmetric ciphers, and behave the same way. Move everything to
12566 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12568 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12570 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12572 *Ben Laurie and Theo de Raadt*
12574 * Add option to output public keys in req command.
12576 *Massimiliano Pala madwolf@openca.org*
12578 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12579 (up to about 10% better than before for P-192 and P-224).
12583 * New functions/macros
12585 SSL_CTX_set_msg_callback(ctx, cb)
12586 SSL_CTX_set_msg_callback_arg(ctx, arg)
12587 SSL_set_msg_callback(ssl, cb)
12588 SSL_set_msg_callback_arg(ssl, arg)
12590 to request calling a callback function
12592 void cb(int write_p, int version, int content_type,
12593 const void *buf, size_t len, SSL *ssl, void *arg)
12595 whenever a protocol message has been completely received
12596 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12597 protocol version according to which the SSL library interprets
12598 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12599 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12600 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12601 specification (change_cipher_spec(20), alert(21), handshake(22)).
12602 'buf' and 'len' point to the actual message, 'ssl' to the
12603 SSL object, and 'arg' is the application-defined value set by
12604 SSL[_CTX]_set_msg_callback_arg().
12606 'openssl s_client' and 'openssl s_server' have new '-msg' options
12607 to enable a callback that displays all protocol messages.
12611 * Change the shared library support so shared libraries are built as
12612 soon as the corresponding static library is finished, and thereby get
12613 openssl and the test programs linked against the shared library.
12614 This still only happens when the keyword "shard" has been given to
12615 the configuration scripts.
12617 NOTE: shared library support is still an experimental thing, and
12618 backward binary compatibility is still not guaranteed.
12620 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12622 * Add support for Subject Information Access extension.
12624 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12626 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12627 additional bytes when new memory had to be allocated, not just
12628 when reusing an existing buffer.
12632 * New command line and configuration option 'utf8' for the req command.
12633 This allows field values to be specified as UTF8 strings.
12637 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12638 runs for the former and machine-readable output for the latter.
12642 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12643 of the e-mail address in the DN (i.e., it will go into a certificate
12644 extension only). The new configuration file option 'email_in_dn = no'
12645 has the same effect.
12647 *Massimiliano Pala madwolf@openca.org*
12649 * Change all functions with names starting with `des_` to be starting
12650 with `DES_` instead. Add wrappers that are compatible with libdes,
12651 but are named `_ossl_old_des_*`. Finally, add macros that map the
12652 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12653 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12654 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12657 Since we provide two compatibility mappings, the user needs to
12658 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12659 compatibility is desired. The default (i.e., when that macro
12660 isn't defined) is OpenSSL 0.9.6c compatibility.
12662 There are also macros that enable and disable the support of old
12663 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12664 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12665 are defined, the default will apply: to support the old des routines.
12667 In either case, one must include openssl/des.h to get the correct
12668 definitions. Do not try to just include openssl/des_old.h, that
12671 NOTE: This is a major break of an old API into a new one. Software
12672 authors are encouraged to switch to the `DES_` style functions. Some
12673 time in the future, des_old.h and the libdes compatibility functions
12674 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12675 default), and then completely removed.
12679 * Test for certificates which contain unsupported critical extensions.
12680 If such a certificate is found during a verify operation it is
12681 rejected by default: this behaviour can be overridden by either
12682 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12683 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12684 X509_supported_extension() has also been added which returns 1 if a
12685 particular extension is supported.
12689 * Modify the behaviour of EVP cipher functions in similar way to digests
12690 to retain compatibility with existing code.
12694 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12695 compatibility with existing code. In particular the 'ctx' parameter does
12696 not have to be to be initialized before the call to EVP_DigestInit() and
12697 it is tidied up after a call to EVP_DigestFinal(). New function
12698 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12699 EVP_MD_CTX_copy() changed to not require the destination to be
12700 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12701 requires the destination to be valid.
12703 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12704 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12708 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12709 so that complete 'Handshake' protocol structures are kept in memory
12710 instead of overwriting 'msg_type' and 'length' with 'body' data.
12714 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12716 *Massimo Santin via Richard Levitte*
12718 * Major restructuring to the underlying ENGINE code. This includes
12719 reduction of linker bloat, separation of pure "ENGINE" manipulation
12720 (initialisation, etc) from functionality dealing with implementations
12721 of specific crypto interfaces. This change also introduces integrated
12722 support for symmetric ciphers and digest implementations - so ENGINEs
12723 can now accelerate these by providing EVP_CIPHER and EVP_MD
12724 implementations of their own. This is detailed in
12725 [crypto/engine/README.md](crypto/engine/README.md)
12726 as it couldn't be adequately described here. However, there are a few
12727 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12728 were changed in the original introduction of ENGINE code have now
12729 reverted back - the hooking from this code to ENGINE is now a good
12730 deal more passive and at run-time, operations deal directly with
12731 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12732 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12733 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12734 they were not being used by the framework as there is no concept of a
12735 BIGNUM_METHOD and they could not be generalised to the new
12736 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12737 ENGINE_cpy() has been removed as it cannot be consistently defined in
12742 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12746 * Change mkdef.pl to sort symbols that get the same entry number,
12747 and make sure the automatically generated functions `ERR_load_*`
12748 become part of libeay.num as well.
12752 * New function SSL_renegotiate_pending(). This returns true once
12753 renegotiation has been requested (either SSL_renegotiate() call
12754 or HelloRequest/ClientHello received from the peer) and becomes
12755 false once a handshake has been completed.
12756 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12757 sends a HelloRequest, but does not ensure that a handshake takes
12758 place. SSL_renegotiate_pending() is useful for checking if the
12759 client has followed the request.)
12763 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12764 By default, clients may request session resumption even during
12765 renegotiation (if session ID contexts permit); with this option,
12766 session resumption is possible only in the first handshake.
12768 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12769 more bits available for options that should not be part of
12770 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12774 * Add some demos for certificate and certificate request creation.
12778 * Make maximum certificate chain size accepted from the peer application
12779 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12780 "Douglas E. Engert" <deengert@anl.gov>.
12784 * Add support for shared libraries for Unixware-7
12785 (Boyd Lynn Gerber <gerberb@zenez.com>).
12789 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12790 be done prior to destruction. Use this to unload error strings from
12791 ENGINEs that load their own error strings. NB: This adds two new API
12792 functions to "get" and "set" this destroy handler in an ENGINE.
12796 * Alter all existing ENGINE implementations (except "openssl" and
12797 "openbsd") to dynamically instantiate their own error strings. This
12798 makes them more flexible to be built both as statically-linked ENGINEs
12799 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12800 Also, add stub code to each that makes building them as self-contained
12801 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12805 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12806 implementations into applications that are completely implemented in
12807 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12808 commands that can be used to configure what shared-library to load and
12809 to control aspects of the way it is handled. Also, made an update to
12810 the [README-Engine.md](README-Engine.md) file
12811 that brings its information up-to-date and
12812 provides some information and instructions on the "dynamic" ENGINE
12813 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12817 * Make it possible to unload ranges of ERR strings with a new
12818 "ERR_unload_strings" function.
12822 * Add a copy() function to EVP_MD.
12826 * Make EVP_MD routines take a context pointer instead of just the
12827 md_data void pointer.
12831 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12832 that the digest can only process a single chunk of data
12833 (typically because it is provided by a piece of
12834 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12835 is only going to provide a single chunk of data, and hence the
12836 framework needn't accumulate the data for oneshot drivers.
12840 * As with "ERR", make it possible to replace the underlying "ex_data"
12841 functions. This change also alters the storage and management of global
12842 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12843 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12844 index counters. The API functions that use this state have been changed
12845 to take a "class_index" rather than pointers to the class's local STACK
12846 and counter, and there is now an API function to dynamically create new
12847 classes. This centralisation allows us to (a) plug a lot of the
12848 thread-safety problems that existed, and (b) makes it possible to clean
12849 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12850 such data would previously have always leaked in application code and
12851 workarounds were in place to make the memory debugging turn a blind eye
12852 to it. Application code that doesn't use this new function will still
12853 leak as before, but their memory debugging output will announce it now
12854 rather than letting it slide.
12856 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12857 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12858 has a return value to indicate success or failure.
12862 * Make it possible to replace the underlying "ERR" functions such that the
12863 global state (2 LHASH tables and 2 locks) is only used by the "default"
12864 implementation. This change also adds two functions to "get" and "set"
12865 the implementation prior to it being automatically set the first time
12866 any other ERR function takes place. Ie. an application can call "get",
12867 pass the return value to a module it has just loaded, and that module
12868 can call its own "set" function using that value. This means the
12869 module's "ERR" operations will use (and modify) the error state in the
12870 application and not in its own statically linked copy of OpenSSL code.
12874 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12875 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12876 the operation, and provides a more encapsulated way for external code
12877 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12878 to use these functions rather than manually incrementing the counts.
12880 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12884 * Add EVP test program.
12888 * Add symmetric cipher support to ENGINE. Expect the API to change!
12892 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12893 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12894 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12895 These allow a CRL to be built without having to access X509_CRL fields
12896 directly. Modify 'ca' application to use new functions.
12900 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12901 bug workarounds. Rollback attack detection is a security feature.
12902 The problem will only arise on OpenSSL servers when TLSv1 is not
12903 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12904 Software authors not wanting to support TLSv1 will have special reasons
12905 for their choice and can explicitly enable this option.
12907 *Bodo Moeller, Lutz Jaenicke*
12909 * Rationalise EVP so it can be extended: don't include a union of
12910 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12911 (similar to those existing for EVP_CIPHER_CTX).
12916 EVP_MD_CTX_init(&md); /* new function call */
12917 EVP_DigestInit(&md, EVP_sha1());
12918 EVP_DigestUpdate(&md, in, len);
12919 EVP_DigestFinal(&md, out, NULL);
12920 EVP_MD_CTX_cleanup(&md); /* new function call */
12924 * Make DES key schedule conform to the usual scheme, as well as
12925 correcting its structure. This means that calls to DES functions
12926 now have to pass a pointer to a des_key_schedule instead of a
12927 plain des_key_schedule (which was actually always a pointer
12930 des_key_schedule ks;
12932 des_set_key_checked(..., &ks);
12933 des_ncbc_encrypt(..., &ks, ...);
12935 (Note that a later change renames 'des_...' into 'DES_...'.)
12939 * Initial reduction of linker bloat: the use of some functions, such as
12940 PEM causes large amounts of unused functions to be linked in due to
12941 poor organisation. For example pem_all.c contains every PEM function
12942 which has a knock on effect of linking in large amounts of (unused)
12943 ASN1 code. Grouping together similar functions and splitting unrelated
12944 functions prevents this.
12948 * Cleanup of EVP macros.
12952 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12953 correct `_ecb suffix`.
12957 * Add initial OCSP responder support to ocsp application. The
12958 revocation information is handled using the text based index
12959 use by the ca application. The responder can either handle
12960 requests generated internally, supplied in files (for example
12961 via a CGI script) or using an internal minimal server.
12965 * Add configuration choices to get zlib compression for TLS.
12969 * Changes to Kerberos SSL for RFC 2712 compliance:
12970 1. Implemented real KerberosWrapper, instead of just using
12971 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12972 2. Implemented optional authenticator field of KerberosWrapper.
12974 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12975 and authenticator structs; see crypto/krb5/.
12977 Generalized Kerberos calls to support multiple Kerberos libraries.
12978 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12979 via Richard Levitte*
12981 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12982 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12983 values for each of the key sizes rather than having just
12984 parameters (and 'speed' generating keys each time).
12988 * Speed up EVP routines.
12991 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12992 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12993 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12994 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12996 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12997 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12998 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13001 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13003 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13007 * Added the OS2-EMX target.
13009 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13011 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13012 New functions to support `NCONF` routines in extension code.
13013 New function `CONF_set_nconf()`
13014 to allow functions which take an `NCONF` to also handle the old `LHASH`
13015 structure: this means that the old `CONF` compatible routines can be
13016 retained (in particular w.rt. extensions) without having to duplicate the
13017 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13021 * Enhance the general user interface with mechanisms for inner control
13022 and with possibilities to have yes/no kind of prompts.
13026 * Change all calls to low-level digest routines in the library and
13027 applications to use EVP. Add missing calls to HMAC_cleanup() and
13028 don't assume HMAC_CTX can be copied using memcpy().
13030 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13032 * Add the possibility to control engines through control names but with
13033 arbitrary arguments instead of just a string.
13034 Change the key loaders to take a UI_METHOD instead of a callback
13035 function pointer. NOTE: this breaks binary compatibility with earlier
13036 versions of OpenSSL [engine].
13037 Adapt the nCipher code for these new conditions and add a card insertion
13042 * Enhance the general user interface with mechanisms to better support
13043 dialog box interfaces, application-defined prompts, the possibility
13044 to use defaults (for example default passwords from somewhere else)
13045 and interrupts/cancellations.
13049 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13050 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13054 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13055 tidy up some unnecessarily weird code in 'sk_new()').
13057 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13059 * Change the key loading routines for ENGINEs to use the same kind
13060 callback (pem_password_cb) as all other routines that need this
13065 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13066 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13067 than this minimum value is recommended.
13071 * New random seeder for OpenVMS, using the system process statistics
13072 that are easily reachable.
13076 * Windows apparently can't transparently handle global
13077 variables defined in DLLs. Initialisations such as:
13079 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13081 won't compile. This is used by the any applications that need to
13082 declare their own ASN1 modules. This was fixed by adding the option
13083 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13084 needed for static libraries under Win32.
13088 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13089 setting of purpose and trust fields. New X509_STORE trust and
13090 purpose functions and tidy up setting in other SSL functions.
13094 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13095 structure. These are inherited by X509_STORE_CTX when it is
13096 initialised. This allows various defaults to be set in the
13097 X509_STORE structure (such as flags for CRL checking and custom
13098 purpose or trust settings) for functions which only use X509_STORE_CTX
13099 internally such as S/MIME.
13101 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13102 trust settings if they are not set in X509_STORE. This allows X509_STORE
13103 purposes and trust (in S/MIME for example) to override any set by default.
13105 Add command line options for CRL checking to smime, s_client and s_server
13110 * Initial CRL based revocation checking. If the CRL checking flag(s)
13111 are set then the CRL is looked up in the X509_STORE structure and
13112 its validity and signature checked, then if the certificate is found
13113 in the CRL the verify fails with a revoked error.
13115 Various new CRL related callbacks added to X509_STORE_CTX structure.
13117 Command line options added to 'verify' application to support this.
13119 This needs some additional work, such as being able to handle multiple
13120 CRLs with different times, extension based lookup (rather than just
13121 by subject name) and ultimately more complete V2 CRL extension
13126 * Add a general user interface API (crypto/ui/). This is designed
13127 to replace things like des_read_password and friends (backward
13128 compatibility functions using this new API are provided).
13129 The purpose is to remove prompting functions from the DES code
13130 section as well as provide for prompting through dialog boxes in
13131 a window system and the like.
13135 * Add "ex_data" support to ENGINE so implementations can add state at a
13136 per-structure level rather than having to store it globally.
13140 * Make it possible for ENGINE structures to be copied when retrieved by
13141 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13142 This causes the "original" ENGINE structure to act like a template,
13143 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13144 operational state can be localised to each ENGINE structure, despite the
13145 fact they all share the same "methods". New ENGINE structures returned in
13146 this case have no functional references and the return value is the single
13147 structural reference. This matches the single structural reference returned
13148 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13153 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13154 needs to match any other type at all we need to manually clear the
13159 * Changes to the "openssl engine" utility to include;
13160 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13161 about an ENGINE's available control commands.
13162 - executing control commands from command line arguments using the
13163 '-pre' and '-post' switches. '-post' is only used if '-t' is
13164 specified and the ENGINE is successfully initialised. The syntax for
13165 the individual commands are colon-separated, for example;
13166 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13170 * New dynamic control command support for ENGINEs. ENGINEs can now
13171 declare their own commands (numbers), names (strings), descriptions,
13172 and input types for run-time discovery by calling applications. A
13173 subset of these commands are implicitly classed as "executable"
13174 depending on their input type, and only these can be invoked through
13175 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13176 can be based on user input, config files, etc). The distinction is
13177 that "executable" commands cannot return anything other than a boolean
13178 result and can only support numeric or string input, whereas some
13179 discoverable commands may only be for direct use through
13180 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13181 pointers, or other custom uses. The "executable" commands are to
13182 support parameterisations of ENGINE behaviour that can be
13183 unambiguously defined by ENGINEs and used consistently across any
13184 OpenSSL-based application. Commands have been added to all the
13185 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13186 control over shared-library paths without source code alterations.
13190 * Changed all ENGINE implementations to dynamically allocate their
13191 ENGINEs rather than declaring them statically. Apart from this being
13192 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13193 this also allows the implementations to compile without using the
13194 internal engine_int.h header.
13198 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13199 'const' value. Any code that should be able to modify a RAND_METHOD
13200 should already have non-const pointers to it (ie. they should only
13201 modify their own ones).
13205 * Made a variety of little tweaks to the ENGINE code.
13206 - "atalla" and "ubsec" string definitions were moved from header files
13207 to C code. "nuron" string definitions were placed in variables
13208 rather than hard-coded - allowing parameterisation of these values
13209 later on via ctrl() commands.
13210 - Removed unused "#if 0"'d code.
13211 - Fixed engine list iteration code so it uses ENGINE_free() to release
13212 structural references.
13213 - Constified the RAND_METHOD element of ENGINE structures.
13214 - Constified various get/set functions as appropriate and added
13215 missing functions (including a catch-all ENGINE_cpy that duplicates
13216 all ENGINE values onto a new ENGINE except reference counts/state).
13217 - Removed NULL parameter checks in get/set functions. Setting a method
13218 or function to NULL is a way of cancelling out a previously set
13219 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13220 and doesn't justify the extra error symbols and code.
13221 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13222 flags from engine_int.h to engine.h.
13223 - Changed prototypes for ENGINE handler functions (init(), finish(),
13224 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13228 * Implement binary inversion algorithm for BN_mod_inverse in addition
13229 to the algorithm using long division. The binary algorithm can be
13230 used only if the modulus is odd. On 32-bit systems, it is faster
13231 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13232 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13233 up to 450 bits. In 64-bit environments, the binary algorithm
13234 appears to be advantageous for much longer moduli; here we use it
13235 for moduli up to 2048 bits.
13239 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13240 could not support the combine flag in choice fields.
13244 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13245 extensions from a certificate request to the certificate.
13249 * Allow multiple 'certopt' and 'nameopt' options to be separated
13250 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13251 file: this allows the display of the certificate about to be
13252 signed to be customised, to allow certain fields to be included
13253 or excluded and extension details. The old system didn't display
13254 multicharacter strings properly, omitted fields not in the policy
13255 and couldn't display additional details such as extensions.
13259 * Function EC_POINTs_mul for multiple scalar multiplication
13260 of an arbitrary number of elliptic curve points
13261 \sum scalars[i]*points[i],
13262 optionally including the generator defined for the EC_GROUP:
13263 scalar*generator + \sum scalars[i]*points[i].
13265 EC_POINT_mul is a simple wrapper function for the typical case
13266 that the point list has just one item (besides the optional
13271 * First EC_METHODs for curves over GF(p):
13273 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13274 operations and provides various method functions that can also
13275 operate with faster implementations of modular arithmetic.
13277 EC_GFp_mont_method() reuses most functions that are part of
13278 EC_GFp_simple_method, but uses Montgomery arithmetic.
13280 *Bodo Moeller; point addition and point doubling
13281 implementation directly derived from source code provided by
13282 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13284 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13285 crypto/ec/ec_lib.c):
13287 Curves are EC_GROUP objects (with an optional group generator)
13288 based on EC_METHODs that are built into the library.
13290 Points are EC_POINT objects based on EC_GROUP objects.
13292 Most of the framework would be able to handle curves over arbitrary
13293 finite fields, but as there are no obvious types for fields other
13294 than GF(p), some functions are limited to that for now.
13298 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13299 that the file contains a complete HTTP response.
13303 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13304 change the def and num file printf format specifier from "%-40sXXX"
13305 to "%-39s XXX". The latter will always guarantee a space after the
13306 field while the former will cause them to run together if the field
13307 is 40 of more characters long.
13311 * Constify the cipher and digest 'method' functions and structures
13312 and modify related functions to take constant EVP_MD and EVP_CIPHER
13317 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13318 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13322 * Modify `EVP_Digest*()` routines so they now return values. Although the
13323 internal software routines can never fail additional hardware versions
13328 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13330 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13331 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13334 ERR_R_NESTED_ASN1_ERROR
13336 ERR_R_MISSING_ASN1_EOS
13337 were 4 .. 9, conflicting with
13338 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13340 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13341 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13343 Add new error code 'ERR_R_INTERNAL_ERROR'.
13347 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13352 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13353 sets the subject name for a new request or supersedes the
13354 subject name in a given request. Formats that can be parsed are
13355 'CN=Some Name, OU=myOU, C=IT'
13357 'CN=Some Name/OU=myOU/C=IT'.
13359 Add options '-batch' and '-verbose' to 'openssl req'.
13361 *Massimiliano Pala <madwolf@hackmasters.net>*
13363 * Introduce the possibility to access global variables through
13364 functions on platform were that's the best way to handle exporting
13365 global variables in shared libraries. To enable this functionality,
13366 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13367 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13368 is normally done by Configure or something similar).
13370 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13371 in the source file (foo.c) like this:
13373 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13374 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13376 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13377 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13379 OPENSSL_DECLARE_GLOBAL(int,foo);
13380 #define foo OPENSSL_GLOBAL_REF(foo)
13381 OPENSSL_DECLARE_GLOBAL(double,bar);
13382 #define bar OPENSSL_GLOBAL_REF(bar)
13384 The #defines are very important, and therefore so is including the
13385 header file everywhere where the defined globals are used.
13387 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13388 of ASN.1 items, but that structure is a bit different.
13390 The largest change is in util/mkdef.pl which has been enhanced with
13391 better and easier to understand logic to choose which symbols should
13392 go into the Windows .def files as well as a number of fixes and code
13393 cleanup (among others, algorithm keywords are now sorted
13394 lexicographically to avoid constant rewrites).
13398 * In BN_div() keep a copy of the sign of 'num' before writing the
13399 result to 'rm' because if rm==num the value will be overwritten
13400 and produce the wrong result if 'num' is negative: this caused
13401 problems with BN_mod() and BN_nnmod().
13405 * Function OCSP_request_verify(). This checks the signature on an
13406 OCSP request and verifies the signer certificate. The signer
13407 certificate is just checked for a generic purpose and OCSP request
13412 * Add OCSP_check_validity() function to check the validity of OCSP
13413 responses. OCSP responses are prepared in real time and may only
13414 be a few seconds old. Simply checking that the current time lies
13415 between thisUpdate and nextUpdate max reject otherwise valid responses
13416 caused by either OCSP responder or client clock inaccuracy. Instead
13417 we allow thisUpdate and nextUpdate to fall within a certain period of
13418 the current time. The age of the response can also optionally be
13419 checked. Two new options -validity_period and -status_age added to
13424 * If signature or public key algorithm is unrecognized print out its
13425 OID rather that just UNKNOWN.
13429 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13430 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13431 ID to be generated from the issuer certificate alone which can then be
13432 passed to OCSP_id_issuer_cmp().
13436 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13437 ASN1 modules to export functions returning ASN1_ITEM pointers
13438 instead of the ASN1_ITEM structures themselves. This adds several
13439 new macros which allow the underlying ASN1 function/structure to
13440 be accessed transparently. As a result code should not use ASN1_ITEM
13441 references directly (such as &X509_it) but instead use the relevant
13442 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13443 use of the new ASN1 code on platforms where exporting structures
13444 is problematical (for example in shared libraries) but exporting
13445 functions returning pointers to structures is not.
13449 * Add support for overriding the generation of SSL/TLS session IDs.
13450 These callbacks can be registered either in an SSL_CTX or per SSL.
13451 The purpose of this is to allow applications to control, if they wish,
13452 the arbitrary values chosen for use as session IDs, particularly as it
13453 can be useful for session caching in multiple-server environments. A
13454 command-line switch for testing this (and any client code that wishes
13455 to use such a feature) has been added to "s_server".
13457 *Geoff Thorpe, Lutz Jaenicke*
13459 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13460 of the form `#if defined(...) || defined(...) || ...` and
13461 `#if !defined(...) && !defined(...) && ...`. This also avoids
13462 the growing number of special cases it was previously handling.
13466 * Make all configuration macros available for application by making
13467 sure they are available in opensslconf.h, by giving them names starting
13468 with `OPENSSL_` to avoid conflicts with other packages and by making
13469 sure e_os2.h will cover all platform-specific cases together with
13471 Additionally, it is now possible to define configuration/platform-
13472 specific names (called "system identities"). In the C code, these
13473 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13474 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13475 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13480 * New option -set_serial to 'req' and 'x509' this allows the serial
13481 number to use to be specified on the command line. Previously self
13482 signed certificates were hard coded with serial number 0 and the
13483 CA options of 'x509' had to use a serial number in a file which was
13488 * New options to 'ca' utility to support V2 CRL entry extensions.
13489 Currently CRL reason, invalidity date and hold instruction are
13490 supported. Add new CRL extensions to V3 code and some new objects.
13494 * New function EVP_CIPHER_CTX_set_padding() this is used to
13495 disable standard block padding (aka PKCS#5 padding) in the EVP
13496 API, which was previously mandatory. This means that the data is
13497 not padded in any way and so the total length much be a multiple
13498 of the block size, otherwise an error occurs.
13502 * Initial (incomplete) OCSP SSL support.
13506 * New function OCSP_parse_url(). This splits up a URL into its host,
13507 port and path components: primarily to parse OCSP URLs. New -url
13508 option to ocsp utility.
13512 * New nonce behavior. The return value of OCSP_check_nonce() now
13513 reflects the various checks performed. Applications can decide
13514 whether to tolerate certain situations such as an absent nonce
13515 in a response when one was present in a request: the ocsp application
13516 just prints out a warning. New function OCSP_add1_basic_nonce()
13517 this is to allow responders to include a nonce in a response even if
13518 the request is nonce-less.
13522 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13523 skipped when using openssl x509 multiple times on a single input file,
13524 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13528 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13529 set string type: to handle setting ASN1_TIME structures. Fix ca
13530 utility to correctly initialize revocation date of CRLs.
13534 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13535 the clients preferred ciphersuites and rather use its own preferences.
13536 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13537 Internet Explorer by ensuring unchanged hash method during stepup.
13538 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13542 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13543 to aes and add a new 'exist' option to print out symbols that don't
13548 * Additional options to ocsp utility to allow flags to be set and
13549 additional certificates supplied.
13553 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13554 OCSP client a number of certificate to only verify the response
13559 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13560 handle the new API. Currently only ECB, CBC modes supported. Add new
13563 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13564 Encryption Standard (AES) Ciphersuites for Transport Layer
13565 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13566 not enabled by default and were not part of the "ALL" ciphersuite
13567 alias because they were not yet official; they could be
13568 explicitly requested by specifying the "AESdraft" ciphersuite
13569 group alias. In the final release of OpenSSL 0.9.7, the group
13570 alias is called "AES" and is part of "ALL".)
13572 *Ben Laurie, Steve Henson, Bodo Moeller*
13574 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13575 request to response.
13579 * Functions for OCSP responders. OCSP_request_onereq_count(),
13580 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13581 extract information from a certificate request. OCSP_response_create()
13582 creates a response and optionally adds a basic response structure.
13583 OCSP_basic_add1_status() adds a complete single response to a basic
13584 response and returns the OCSP_SINGLERESP structure just added (to allow
13585 extensions to be included for example). OCSP_basic_add1_cert() adds a
13586 certificate to a basic response and OCSP_basic_sign() signs a basic
13587 response with various flags. New helper functions ASN1_TIME_check()
13588 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13589 (converts ASN1_TIME to GeneralizedTime).
13593 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13594 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13595 structure from a certificate. X509_pubkey_digest() digests the public_key
13596 contents: this is used in various key identifiers.
13600 * Make sk_sort() tolerate a NULL argument.
13602 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13604 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13605 passed by the function are trusted implicitly. If any of them signed the
13606 response then it is assumed to be valid and is not verified.
13610 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13611 to data. This was previously part of the PKCS7 ASN1 code. This
13612 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13613 *Steve Henson, reported by Kenneth R. Robinette
13614 <support@securenetterm.com>*
13616 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13617 routines: without these tracing memory leaks is very painful.
13618 Fix leaks in PKCS12 and PKCS7 routines.
13622 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13623 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13624 effectively meant GeneralizedTime would never be used. Now it
13625 is initialised to -1 but X509_time_adj() now has to check the value
13626 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13627 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13628 *Steve Henson, reported by Kenneth R. Robinette
13629 <support@securenetterm.com>*
13631 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13632 result in a zero length in the ASN1_INTEGER structure which was
13633 not consistent with the structure when d2i_ASN1_INTEGER() was used
13634 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13635 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13636 where it did not print out a minus for negative ASN1_INTEGER.
13640 * Add summary printout to ocsp utility. The various functions which
13641 convert status values to strings have been renamed to:
13642 OCSP_response_status_str(), OCSP_cert_status_str() and
13643 OCSP_crl_reason_str() and are no longer static. New options
13644 to verify nonce values and to disable verification. OCSP response
13645 printout format cleaned up.
13649 * Add additional OCSP certificate checks. These are those specified
13650 in RFC2560. This consists of two separate checks: the CA of the
13651 certificate being checked must either be the OCSP signer certificate
13652 or the issuer of the OCSP signer certificate. In the latter case the
13653 OCSP signer certificate must contain the OCSP signing extended key
13654 usage. This check is performed by attempting to match the OCSP
13655 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13656 in the OCSP_CERTID structures of the response.
13660 * Initial OCSP certificate verification added to OCSP_basic_verify()
13661 and related routines. This uses the standard OpenSSL certificate
13662 verify routines to perform initial checks (just CA validity) and
13663 to obtain the certificate chain. Then additional checks will be
13664 performed on the chain. Currently the root CA is checked to see
13665 if it is explicitly trusted for OCSP signing. This is used to set
13666 a root CA as a global signing root: that is any certificate that
13667 chains to that CA is an acceptable OCSP signing certificate.
13671 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13672 extensions from a separate configuration file.
13673 As when reading extensions from the main configuration file,
13674 the '-extensions ...' option may be used for specifying the
13677 *Massimiliano Pala <madwolf@comune.modena.it>*
13679 * New OCSP utility. Allows OCSP requests to be generated or
13680 read. The request can be sent to a responder and the output
13681 parsed, outputted or printed in text form. Not complete yet:
13682 still needs to check the OCSP response validity.
13686 * New subcommands for 'openssl ca':
13687 `openssl ca -status <serial>` prints the status of the cert with
13688 the given serial number (according to the index file).
13689 `openssl ca -updatedb` updates the expiry status of certificates
13692 *Massimiliano Pala <madwolf@comune.modena.it>*
13694 * New '-newreq-nodes' command option to CA.pl. This is like
13695 '-newreq', but calls 'openssl req' with the '-nodes' option
13696 so that the resulting key is not encrypted.
13698 *Damien Miller <djm@mindrot.org>*
13700 * New configuration for the GNU Hurd.
13702 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13704 * Initial code to implement OCSP basic response verify. This
13705 is currently incomplete. Currently just finds the signer's
13706 certificate and verifies the signature on the response.
13710 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13711 value of OPENSSLDIR. This is available via the new '-d' option
13712 to 'openssl version', and is also included in 'openssl version -a'.
13716 * Allowing defining memory allocation callbacks that will be given
13717 file name and line number information in additional arguments
13718 (a `const char*` and an int). The basic functionality remains, as
13719 well as the original possibility to just replace malloc(),
13720 realloc() and free() by functions that do not know about these
13721 additional arguments. To register and find out the current
13722 settings for extended allocation functions, the following
13723 functions are provided:
13725 CRYPTO_set_mem_ex_functions
13726 CRYPTO_set_locked_mem_ex_functions
13727 CRYPTO_get_mem_ex_functions
13728 CRYPTO_get_locked_mem_ex_functions
13730 These work the same way as CRYPTO_set_mem_functions and friends.
13731 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13732 extended allocation function is enabled.
13733 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13734 a conventional allocation function is enabled.
13736 *Richard Levitte, Bodo Moeller*
13738 * Finish off removing the remaining LHASH function pointer casts.
13739 There should no longer be any prototype-casting required when using
13740 the LHASH abstraction, and any casts that remain are "bugs". See
13741 the callback types and macros at the head of lhash.h for details
13742 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13746 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13747 If /dev/[u]random devices are not available or do not return enough
13748 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13750 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13751 /etc/entropy will be queried once each in this sequence, querying stops
13752 when enough entropy was collected without querying more sockets.
13756 * Change the Unix RAND_poll() variant to be able to poll several
13757 random devices, as specified by DEVRANDOM, until a sufficient amount
13758 of data has been collected. We spend at most 10 ms on each file
13759 (select timeout) and read in non-blocking mode. DEVRANDOM now
13760 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13761 (previously it was just the string "/dev/urandom"), so on typical
13762 platforms the 10 ms delay will never occur.
13763 Also separate out the Unix variant to its own file, rand_unix.c.
13764 For VMS, there's a currently-empty rand_vms.c.
13768 * Move OCSP client related routines to ocsp_cl.c. These
13769 provide utility functions which an application needing
13770 to issue a request to an OCSP responder and analyse the
13771 response will typically need: as opposed to those which an
13772 OCSP responder itself would need which will be added later.
13774 OCSP_request_sign() signs an OCSP request with an API similar
13775 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13776 response. OCSP_response_get1_basic() extracts basic response
13777 from response. OCSP_resp_find_status(): finds and extracts status
13778 information from an OCSP_CERTID structure (which will be created
13779 when the request structure is built). These are built from lower
13780 level functions which work on OCSP_SINGLERESP structures but
13781 won't normally be used unless the application wishes to examine
13782 extensions in the OCSP response for example.
13784 Replace nonce routines with a pair of functions.
13785 OCSP_request_add1_nonce() adds a nonce value and optionally
13786 generates a random value. OCSP_check_nonce() checks the
13787 validity of the nonce in an OCSP response.
13791 * Change function OCSP_request_add() to OCSP_request_add0_id().
13792 This doesn't copy the supplied OCSP_CERTID and avoids the
13793 need to free up the newly created id. Change return type
13794 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13795 This can then be used to add extensions to the request.
13796 Deleted OCSP_request_new(), since most of its functionality
13797 is now in OCSP_REQUEST_new() (and the case insensitive name
13798 clash) apart from the ability to set the request name which
13799 will be added elsewhere.
13803 * Update OCSP API. Remove obsolete extensions argument from
13804 various functions. Extensions are now handled using the new
13805 OCSP extension code. New simple OCSP HTTP function which
13806 can be used to send requests and parse the response.
13810 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13811 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13812 uses the special reorder version of SET OF to sort the attributes
13813 and reorder them to match the encoded order. This resolves a long
13814 standing problem: a verify on a PKCS7 structure just after signing
13815 it used to fail because the attribute order did not match the
13816 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13817 it uses the received order. This is necessary to tolerate some broken
13818 software that does not order SET OF. This is handled by encoding
13819 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13820 to produce the required SET OF.
13824 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13825 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13826 files to get correct declarations of the ASN.1 item variables.
13830 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13831 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13832 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13833 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13834 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13835 ASN1_ITEM and no wrapper functions.
13839 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13840 replace the old function pointer based I/O routines. Change most of
13841 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13845 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13846 lines, recognize more "algorithms" that can be deselected, and make
13847 it complain about algorithm deselection that isn't recognised.
13851 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13852 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13853 to use new functions. Add NO_ASN1_OLD which can be set to remove
13854 some old style ASN1 functions: this can be used to determine if old
13855 code will still work when these eventually go away.
13859 * New extension functions for OCSP structures, these follow the
13860 same conventions as certificates and CRLs.
13864 * New function X509V3_add1_i2d(). This automatically encodes and
13865 adds an extension. Its behaviour can be customised with various
13866 flags to append, replace or delete. Various wrappers added for
13867 certificates and CRLs.
13871 * Fix to avoid calling the underlying ASN1 print routine when
13872 an extension cannot be parsed. Correct a typo in the
13873 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13877 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13878 entries for variables.
13882 * Add functionality to `apps/openssl.c` for detecting locking
13883 problems: As the program is single-threaded, all we have
13884 to do is register a locking callback using an array for
13885 storing which locks are currently held by the program.
13889 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13890 SSL_get_ex_data_X509_STORE_idx(), which is used in
13891 ssl_verify_cert_chain() and thus can be called at any time
13892 during TLS/SSL handshakes so that thread-safety is essential.
13893 Unfortunately, the ex_data design is not at all suited
13894 for multi-threaded use, so it probably should be abolished.
13898 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13900 *Broadcom, tweaked and integrated by Geoff Thorpe*
13902 * Move common extension printing code to new function
13903 X509V3_print_extensions(). Reorganise OCSP print routines and
13904 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13908 * New function X509_signature_print() to remove duplication in some
13913 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13914 set (this was treated exactly the same as SET OF previously). This
13915 is used to reorder the STACK representing the structure to match the
13916 encoding. This will be used to get round a problem where a PKCS7
13917 structure which was signed could not be verified because the STACK
13918 order did not reflect the encoded order.
13922 * Reimplement the OCSP ASN1 module using the new code.
13926 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13927 for its ASN1 operations. The old style function pointers still exist
13928 for now but they will eventually go away.
13932 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13933 completely replaces the old ASN1 functionality with a table driven
13934 encoder and decoder which interprets an ASN1_ITEM structure describing
13935 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13936 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13937 has also been converted to the new form.
13941 * Change BN_mod_exp_recp so that negative moduli are tolerated
13942 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13943 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13944 for negative moduli.
13948 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13949 of not touching the result's sign bit.
13953 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13958 * Changed the LHASH code to use prototypes for callbacks, and created
13959 macros to declare and implement thin (optionally static) functions
13960 that provide type-safety and avoid function pointer casting for the
13961 type-specific callbacks.
13965 * Added Kerberos Cipher Suites to be used with TLS, as written in
13967 *Veers Staats <staatsvr@asc.hpc.mil>,
13968 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13970 * Reformat the FAQ so the different questions and answers can be divided
13971 in sections depending on the subject.
13975 * Have the zlib compression code load ZLIB.DLL dynamically under
13980 * New function BN_mod_sqrt for computing square roots modulo a prime
13981 (using the probabilistic Tonelli-Shanks algorithm unless
13982 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13983 be handled deterministically).
13985 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13987 * Make BN_mod_inverse faster by explicitly handling small quotients
13988 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13989 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13993 * New function BN_kronecker.
13997 * Fix BN_gcd so that it works on negative inputs; the result is
13998 positive unless both parameters are zero.
13999 Previously something reasonably close to an infinite loop was
14000 possible because numbers could be growing instead of shrinking
14001 in the implementation of Euclid's algorithm.
14005 * Fix BN_is_word() and BN_is_one() macros to take into account the
14006 sign of the number in question.
14008 Fix BN_is_word(a,w) to work correctly for w == 0.
14010 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14011 because its test if the absolute value of 'a' equals 'w'.
14012 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14013 it exists mostly for use in the implementations of BN_is_zero(),
14014 BN_is_one(), and BN_is_word().
14018 * New function BN_swap.
14022 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14023 the exponentiation functions are more likely to produce reasonable
14024 results on negative inputs.
14028 * Change BN_mod_mul so that the result is always non-negative.
14029 Previously, it could be negative if one of the factors was negative;
14030 I don't think anyone really wanted that behaviour.
14034 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14035 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14036 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14037 and add new functions:
14046 BN_mod_lshift1_quick
14048 BN_mod_lshift_quick
14050 These functions always generate non-negative results.
14052 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14053 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14055 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14056 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14057 be reduced modulo `m`.
14059 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14062 The following entry accidentally appeared in the CHANGES file
14063 distributed with OpenSSL 0.9.7. The modifications described in
14064 it do *not* apply to OpenSSL 0.9.7.
14066 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14067 was actually never needed) and in BN_mul(). The removal in BN_mul()
14068 required a small change in bn_mul_part_recursive() and the addition
14069 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14070 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14071 bn_sub_words() and bn_add_words() except they take arrays with
14077 * In 'openssl passwd', verify passwords read from the terminal
14078 unless the '-salt' option is used (which usually means that
14079 verification would just waste user's time since the resulting
14080 hash is going to be compared with some given password hash)
14081 or the new '-noverify' option is used.
14083 This is an incompatible change, but it does not affect
14084 non-interactive use of 'openssl passwd' (passwords on the command
14085 line, '-stdin' option, '-in ...' option) and thus should not
14086 cause any problems.
14090 * Remove all references to RSAref, since there's no more need for it.
14094 * Make DSO load along a path given through an environment variable
14095 (SHLIB_PATH) with shl_load().
14099 * Constify the ENGINE code as a result of BIGNUM constification.
14100 Also constify the RSA code and most things related to it. In a
14101 few places, most notable in the depth of the ASN.1 code, ugly
14102 casts back to non-const were required (to be solved at a later
14107 * Make it so the openssl application has all engines loaded by default.
14111 * Constify the BIGNUM routines a little more.
14115 * Add the following functions:
14117 ENGINE_load_cswift()
14119 ENGINE_load_atalla()
14120 ENGINE_load_nuron()
14121 ENGINE_load_builtin_engines()
14123 That way, an application can itself choose if external engines that
14124 are built-in in OpenSSL shall ever be used or not. The benefit is
14125 that applications won't have to be linked with libdl or other dso
14126 libraries unless it's really needed.
14128 Changed 'openssl engine' to load all engines on demand.
14129 Changed the engine header files to avoid the duplication of some
14130 declarations (they differed!).
14134 * 'openssl engine' can now list capabilities.
14138 * Better error reporting in 'openssl engine'.
14142 * Never call load_dh_param(NULL) in s_server.
14146 * Add engine application. It can currently list engines by name and
14147 identity, and test if they are actually available.
14151 * Improve RPM specification file by forcing symbolic linking and making
14152 sure the installed documentation is also owned by root.root.
14154 *Damien Miller <djm@mindrot.org>*
14156 * Give the OpenSSL applications more possibilities to make use of
14157 keys (public as well as private) handled by engines.
14161 * Add OCSP code that comes from CertCo.
14165 * Add VMS support for the Rijndael code.
14169 * Added untested support for Nuron crypto accelerator.
14173 * Add support for external cryptographic devices. This code was
14174 previously distributed separately as the "engine" branch.
14176 *Geoff Thorpe, Richard Levitte*
14178 * Rework the filename-translation in the DSO code. It is now possible to
14179 have far greater control over how a "name" is turned into a filename
14180 depending on the operating environment and any oddities about the
14181 different shared library filenames on each system.
14185 * Support threads on FreeBSD-elf in Configure.
14189 * Fix for SHA1 assembly problem with MASM: it produces
14190 warnings about corrupt line number information when assembling
14191 with debugging information. This is caused by the overlapping
14194 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14197 NCONF_get_number() has no error checking at all. As a replacement,
14198 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14199 promoted strongly. The old NCONF_get_number is kept around for
14200 binary backward compatibility.
14201 Make it possible for methods to load from something other than a BIO,
14202 by providing a function pointer that is given a name instead of a BIO.
14203 For example, this could be used to load configuration data from an
14208 * Fix for non blocking accept BIOs. Added new I/O special reason
14209 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14210 with non blocking I/O was not possible because no retry code was
14211 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14216 * Added the beginnings of Rijndael support.
14220 * Fix for bug in DirectoryString mask setting. Add support for
14221 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14222 to allow certificate printing to more controllable, additional
14223 'certopt' option to 'x509' to allow new printing options to be
14228 * Clean old EAY MD5 hack from e_os.h.
14232 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14234 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14235 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14237 *Joe Orton, Steve Henson*
14239 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14241 * Fix additional bug revealed by the NISCC test suite:
14243 Stop bug triggering large recursion when presented with
14244 certain ASN.1 tags ([CVE-2003-0851])
14248 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14250 * Fix various bugs revealed by running the NISCC test suite:
14252 Stop out of bounds reads in the ASN1 code when presented with
14253 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14255 If verify callback ignores invalid public key errors don't try to check
14256 certificate signature with the NULL public key.
14260 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14261 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14266 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14267 extra data after the compression methods not only for TLS 1.0
14268 but also for SSL 3.0 (as required by the specification).
14270 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14272 * Change X509_certificate_type() to mark the key as exported/exportable
14273 when it's 512 *bits* long, not 512 bytes.
14277 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14279 * Countermeasure against the Klima-Pokorny-Rosa extension of
14280 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14281 a protocol version number mismatch like a decryption error
14282 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14286 * Turn on RSA blinding by default in the default implementation
14287 to avoid a timing attack. Applications that don't want it can call
14288 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14289 They would be ill-advised to do so in most cases.
14291 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14293 * Change RSA blinding code so that it works when the PRNG is not
14294 seeded (in this case, the secret RSA exponent is abused as
14295 an unpredictable seed -- if it is not unpredictable, there
14296 is no point in blinding anyway). Make RSA blinding thread-safe
14297 by remembering the creator's thread ID in rsa->blinding and
14298 having all other threads use local one-time blinding factors
14299 (this requires more computation than sharing rsa->blinding, but
14300 avoids excessive locking; and if an RSA object is not shared
14301 between threads, blinding will still be very fast).
14305 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14307 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14308 via timing by performing a MAC computation even if incorrect
14309 block cipher padding has been found. This is a countermeasure
14310 against active attacks where the attacker has to distinguish
14311 between bad padding and a MAC verification error. ([CVE-2003-0078])
14313 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14314 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14315 Martin Vuagnoux (EPFL, Ilion)*
14317 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14319 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14320 memory from its contents. This is done with a counter that will
14321 place alternating values in each byte. This can be used to solve
14322 two issues: 1) the removal of calls to memset() by highly optimizing
14323 compilers, and 2) cleansing with other values than 0, since those can
14324 be read through on certain media, for example a swap space on disk.
14328 * Bugfix: client side session caching did not work with external caching,
14329 because the session->cipher setting was not restored when reloading
14330 from the external cache. This problem was masked, when
14331 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14332 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14336 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14337 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14339 *Zeev Lieber <zeev-l@yahoo.com>*
14341 * Undo an undocumented change introduced in 0.9.6e which caused
14342 repeated calls to OpenSSL_add_all_ciphers() and
14343 OpenSSL_add_all_digests() to be ignored, even after calling
14348 * Change the default configuration reader to deal with last line not
14349 being properly terminated.
14353 * Change X509_NAME_cmp() so it applies the special rules on handling
14354 DN values that are of type PrintableString, as well as RDNs of type
14355 emailAddress where the value has the type ia5String.
14357 *stefank@valicert.com via Richard Levitte*
14359 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14360 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14361 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14362 the bitwise-OR of the two for use by the majority of applications
14363 wanting this behaviour, and update the docs. The documented
14364 behaviour and actual behaviour were inconsistent and had been
14365 changing anyway, so this is more a bug-fix than a behavioural
14368 *Geoff Thorpe, diagnosed by Nadav Har'El*
14370 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14371 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14375 * Fix initialization code race conditions in
14376 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14377 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14378 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14379 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14380 ssl2_get_cipher_by_char(),
14381 ssl3_get_cipher_by_char().
14383 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14385 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14386 the cached sessions are flushed, as the remove_cb() might use ex_data
14387 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14388 (see [openssl.org #212]).
14390 *Geoff Thorpe, Lutz Jaenicke*
14392 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14393 length, instead of the encoding length to d2i_ASN1_OBJECT.
14397 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14399 * [In 0.9.6g-engine release:]
14400 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14402 *Lynn Gazis <lgazis@rainbow.com>*
14404 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14406 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14407 and get fix the header length calculation.
14408 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14409 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14411 * Use proper error handling instead of 'assertions' in buffer
14412 overflow checks added in 0.9.6e. This prevents DoS (the
14413 assertions could call abort()).
14415 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14417 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14419 * Add various sanity checks to asn1_get_length() to reject
14420 the ASN1 length bytes if they exceed sizeof(long), will appear
14421 negative or the content length exceeds the length of the
14424 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14426 * Fix cipher selection routines: ciphers without encryption had no flags
14427 for the cipher strength set and where therefore not handled correctly
14428 by the selection routines (PR #130).
14432 * Fix EVP_dsa_sha macro.
14437 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14438 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14439 that was added in OpenSSL 0.9.6d.
14441 As the countermeasure turned out to be incompatible with some
14442 broken SSL implementations, the new option is part of SSL_OP_ALL.
14443 SSL_OP_ALL is usually employed when compatibility with weird SSL
14444 implementations is desired (e.g. '-bugs' option to 's_client' and
14445 's_server'), so the new option is automatically set in many
14450 * Changes in security patch:
14452 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14453 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14454 Air Force Materiel Command, USAF, under agreement number
14457 * Add various sanity checks to asn1_get_length() to reject
14458 the ASN1 length bytes if they exceed sizeof(long), will appear
14459 negative or the content length exceeds the length of the
14460 supplied buffer. ([CVE-2002-0659])
14462 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14464 * Assertions for various potential buffer overflows, not known to
14465 happen in practice.
14467 *Ben Laurie (CHATS)*
14469 * Various temporary buffers to hold ASCII versions of integers were
14470 too small for 64 bit platforms. ([CVE-2002-0655])
14471 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14473 * Remote buffer overflow in SSL3 protocol - an attacker could
14474 supply an oversized session ID to a client. ([CVE-2002-0656])
14476 *Ben Laurie (CHATS)*
14478 * Remote buffer overflow in SSL2 protocol - an attacker could
14479 supply an oversized client master key. ([CVE-2002-0656])
14481 *Ben Laurie (CHATS)*
14483 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14485 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14486 encoded as NULL) with id-dsa-with-sha1.
14488 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14490 * Check various `X509_...()` return values in `apps/req.c`.
14492 *Nils Larsch <nla@trustcenter.de>*
14494 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14495 an end-of-file condition would erroneously be flagged, when the CRLF
14496 was just at the end of a processed block. The bug was discovered when
14497 processing data through a buffering memory BIO handing the data to a
14498 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14499 <ptsekov@syntrex.com> and Nedelcho Stanev.
14503 * Implement a countermeasure against a vulnerability recently found
14504 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14505 before application data chunks to avoid the use of known IVs
14506 with data potentially chosen by the attacker.
14510 * Fix length checks in ssl3_get_client_hello().
14514 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14515 to prevent ssl3_read_internal() from incorrectly assuming that
14516 ssl3_read_bytes() found application data while handshake
14517 processing was enabled when in fact s->s3->in_read_app_data was
14518 merely automatically cleared during the initial handshake.
14520 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14522 * Fix object definitions for Private and Enterprise: they were not
14523 recognized in their shortname (=lowercase) representation. Extend
14524 obj_dat.pl to issue an error when using undefined keywords instead
14525 of silently ignoring the problem (Svenning Sorensen
14526 <sss@sss.dnsalias.net>).
14530 * Fix DH_generate_parameters() so that it works for 'non-standard'
14531 generators, i.e. generators other than 2 and 5. (Previously, the
14532 code did not properly initialise the 'add' and 'rem' values to
14533 BN_generate_prime().)
14535 In the new general case, we do not insist that 'generator' is
14536 actually a primitive root: This requirement is rather pointless;
14537 a generator of the order-q subgroup is just as good, if not
14542 * Map new X509 verification errors to alerts. Discovered and submitted by
14543 Tom Wu <tom@arcot.com>.
14547 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14548 returning non-zero before the data has been completely received
14549 when using non-blocking I/O.
14551 *Bodo Moeller; problem pointed out by John Hughes*
14553 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14555 *Ben Laurie, Lutz Jaenicke*
14557 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14558 Yoram Zahavi <YoramZ@gilian.com>).
14562 * Add information about CygWin 1.3 and on, and preserve proper
14563 configuration for the versions before that.
14565 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14567 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14568 check whether we deal with a copy of a session and do not delete from
14569 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14570 <izhar@checkpoint.com>.
14574 * Do not store session data into the internal session cache, if it
14575 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14576 flag is set). Proposed by Aslam <aslam@funk.com>.
14580 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14585 * [In 0.9.6d-engine release:]
14586 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14588 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14590 * Add the configuration target linux-s390x.
14592 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14594 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14595 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14596 variable as an indication that a ClientHello message has been
14597 received. As the flag value will be lost between multiple
14598 invocations of ssl3_accept when using non-blocking I/O, the
14599 function may not be aware that a handshake has actually taken
14600 place, thus preventing a new session from being added to the
14603 To avoid this problem, we now set s->new_session to 2 instead of
14604 using a local variable.
14606 *Lutz Jaenicke, Bodo Moeller*
14608 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14609 if the SSL_R_LENGTH_MISMATCH error is detected.
14611 *Geoff Thorpe, Bodo Moeller*
14613 * New 'shared_ldflag' column in Configure platform table.
14617 * Fix EVP_CIPHER_mode macro.
14619 *"Dan S. Camper" <dan@bti.net>*
14621 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14622 type, we must throw them away by setting rr->length to 0.
14624 *D P Chang <dpc@qualys.com>*
14626 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14628 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14629 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14630 worked incorrectly for those cases where range = `10..._2` and
14631 `3*range` is two bits longer than range.)
14635 * Only add signing time to PKCS7 structures if it is not already
14640 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14641 OBJ_ld_ce should be OBJ_id_ce.
14642 Also some ip-pda OIDs in crypto/objects/objects.txt were
14643 incorrect (cf. RFC 3039).
14645 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14647 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14648 returns early because it has nothing to do.
14650 *Andy Schneider <andy.schneider@bjss.co.uk>*
14652 * [In 0.9.6c-engine release:]
14653 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14655 *Andy Schneider <andy.schneider@bjss.co.uk>*
14657 * [In 0.9.6c-engine release:]
14658 Add support for Cryptographic Appliance's keyserver technology.
14659 (Use engine 'keyclient')
14661 *Cryptographic Appliances and Geoff Thorpe*
14663 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14664 is called via tools/c89.sh because arguments have to be
14665 rearranged (all '-L' options must appear before the first object
14668 *Richard Shapiro <rshapiro@abinitio.com>*
14670 * [In 0.9.6c-engine release:]
14671 Add support for Broadcom crypto accelerator cards, backported
14674 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14676 * [In 0.9.6c-engine release:]
14677 Add support for SureWare crypto accelerator cards from
14678 Baltimore Technologies. (Use engine 'sureware')
14680 *Baltimore Technologies and Mark Cox*
14682 * [In 0.9.6c-engine release:]
14683 Add support for crypto accelerator cards from Accelerated
14684 Encryption Processing, www.aep.ie. (Use engine 'aep')
14686 *AEP Inc. and Mark Cox*
14688 * Add a configuration entry for gcc on UnixWare.
14690 *Gary Benson <gbenson@redhat.com>*
14692 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14693 messages are stored in a single piece (fixed-length part and
14694 variable-length part combined) and fix various bugs found on the way.
14698 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14699 instead. BIO_gethostbyname() does not know what timeouts are
14700 appropriate, so entries would stay in cache even when they have
14702 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14704 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14705 faced with a pathologically small ClientHello fragment that does
14706 not contain client_version: Instead of aborting with an error,
14707 simply choose the highest available protocol version (i.e.,
14708 TLS 1.0 unless it is disabled). In practice, ClientHello
14709 messages are never sent like this, but this change gives us
14710 strictly correct behaviour at least for TLS.
14714 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14715 never resets s->method to s->ctx->method when called from within
14716 one of the SSL handshake functions.
14718 *Bodo Moeller; problem pointed out by Niko Baric*
14720 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14721 (sent using the client's version number) if client_version is
14722 smaller than the protocol version in use. Also change
14723 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14724 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14725 the client will at least see that alert.
14729 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14734 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14735 client receives HelloRequest while in a handshake.
14737 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14739 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14740 should end in 'break', not 'goto end' which circumvents various
14741 cleanups done in state SSL_ST_OK. But session related stuff
14742 must be disabled for SSL_ST_OK in the case that we just sent a
14745 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14746 before just sending a HelloRequest.
14748 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14750 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14751 reveal whether illegal block cipher padding was found or a MAC
14752 verification error occurred. (Neither SSLerr() codes nor alerts
14753 are directly visible to potential attackers, but the information
14754 may leak via logfiles.)
14756 Similar changes are not required for the SSL 2.0 implementation
14757 because the number of padding bytes is sent in clear for SSL 2.0,
14758 and the extra bytes are just ignored. However ssl/s2_pkt.c
14759 failed to verify that the purported number of padding bytes is in
14764 * Add OpenUNIX-8 support including shared libraries
14765 (Boyd Lynn Gerber <gerberb@zenez.com>).
14769 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14770 'wristwatch attack' using huge encoding parameters (cf.
14771 James H. Manger's CRYPTO 2001 paper). Note that the
14772 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14773 encoding parameters and hence was not vulnerable.
14777 * BN_sqr() bug fix.
14779 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14781 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14782 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14783 followed by modular reduction.
14785 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14787 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14788 equivalent based on BN_pseudo_rand() instead of BN_rand().
14792 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14793 This function was broken, as the check for a new client hello message
14794 to handle SGC did not allow these large messages.
14795 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14799 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14803 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14804 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14808 * Rework the configuration and shared library support for Tru64 Unix.
14809 The configuration part makes use of modern compiler features and
14810 still retains old compiler behavior for those that run older versions
14811 of the OS. The shared library support part includes a variant that
14812 uses the RPATH feature, and is available through the special
14813 configuration target "alpha-cc-rpath", which will never be selected
14816 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14818 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14819 with the same message size as in ssl3_get_certificate_request().
14820 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14821 messages might inadvertently be reject as too long.
14823 *Petr Lampa <lampa@fee.vutbr.cz>*
14825 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14829 * Modified SSL library such that the verify_callback that has been set
14830 specifically for an SSL object with SSL_set_verify() is actually being
14831 used. Before the change, a verify_callback set with this function was
14832 ignored and the verify_callback() set in the SSL_CTX at the time of
14833 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14834 to allow the necessary settings.
14838 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14839 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14840 done automatically (in contradiction to the requirements of the C
14841 standard). This made problems when used from OpenSSH.
14845 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14846 dh->length and always used
14848 BN_rand_range(priv_key, dh->p).
14850 BN_rand_range() is not necessary for Diffie-Hellman, and this
14851 specific range makes Diffie-Hellman unnecessarily inefficient if
14852 dh->length (recommended exponent length) is much smaller than the
14853 length of dh->p. We could use BN_rand_range() if the order of
14854 the subgroup was stored in the DH structure, but we only have
14859 BN_rand(priv_key, l, ...)
14861 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14868 RSA_eay_public_encrypt
14869 RSA_eay_private_decrypt
14870 RSA_eay_private_encrypt (signing)
14871 RSA_eay_public_decrypt (signature verification)
14873 (default implementations for RSA_public_encrypt,
14874 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14875 always reject numbers >= n.
14879 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14880 to synchronize access to 'locking_thread'. This is necessary on
14881 systems where access to 'locking_thread' (an 'unsigned long'
14882 variable) is not atomic.
14886 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14887 *before* setting the 'crypto_lock_rand' flag. The previous code had
14888 a race condition if 0 is a valid thread ID.
14890 *Travis Vitek <vitek@roguewave.com>*
14892 * Add support for shared libraries under Irix.
14894 *Albert Chin-A-Young <china@thewrittenword.com>*
14896 * Add configuration option to build on Linux on both big-endian and
14897 little-endian MIPS.
14899 *Ralf Baechle <ralf@uni-koblenz.de>*
14901 * Add the possibility to create shared libraries on HP-UX.
14905 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14907 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14908 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14909 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14910 PRNG state recovery was possible based on the output of
14911 one PRNG request appropriately sized to gain knowledge on
14912 'md' followed by enough consecutive 1-byte PRNG requests
14913 to traverse all of 'state'.
14915 1. When updating 'md_local' (the current thread's copy of 'md')
14916 during PRNG output generation, hash all of the previous
14917 'md_local' value, not just the half used for PRNG output.
14919 2. Make the number of bytes from 'state' included into the hash
14920 independent from the number of PRNG bytes requested.
14922 The first measure alone would be sufficient to avoid
14923 Markku-Juhani's attack. (Actually it had never occurred
14924 to me that the half of 'md_local' used for chaining was the
14925 half from which PRNG output bytes were taken -- I had always
14926 assumed that the secret half would be used.) The second
14927 measure makes sure that additional data from 'state' is never
14928 mixed into 'md_local' in small portions; this heuristically
14929 further strengthens the PRNG.
14933 * Fix crypto/bn/asm/mips3.s.
14937 * When only the key is given to "enc", the IV is undefined. Print out
14938 an error message in this case.
14942 * Handle special case when X509_NAME is empty in X509 printing routines.
14946 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14947 positive and less than q.
14951 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14952 used: it isn't thread safe and the add_lock_callback should handle
14955 *Paul Rose <Paul.Rose@bridge.com>*
14957 * Verify that incoming data obeys the block size in
14958 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14964 *Ulf Möller, Bodo Möller*
14966 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14967 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14968 when fixing the server behaviour for backwards-compatible 'client
14969 hello' messages. (Note that the attack is impractical against
14970 SSL 3.0 and TLS 1.0 anyway because length and version checking
14971 means that the probability of guessing a valid ciphertext is
14972 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14975 Before 0.9.5, the countermeasure (hide the error by generating a
14976 random 'decryption result') did not work properly because
14977 ERR_clear_error() was missing, meaning that SSL_get_error() would
14978 detect the supposedly ignored error.
14980 Both problems are now fixed.
14984 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14985 (previously it was 1024).
14989 * Fix for compatibility mode trust settings: ignore trust settings
14990 unless some valid trust or reject settings are present.
14994 * Fix for blowfish EVP: its a variable length cipher.
14998 * Fix various bugs related to DSA S/MIME verification. Handle missing
14999 parameters in DSA public key structures and return an error in the
15000 DSA routines if parameters are absent.
15004 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15005 in the current directory if neither $RANDFILE nor $HOME was set.
15006 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15007 caused some confusion to Windows users who haven't defined $HOME.
15008 Thus RAND_file_name() is changed again: e_os.h can define a
15009 DEFAULT_HOME, which will be used if $HOME is not set.
15010 For Windows, we use "C:"; on other platforms, we still require
15011 environment variables.
15013 * Move 'if (!initialized) RAND_poll()' into regions protected by
15014 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15015 having multiple threads call RAND_poll() concurrently.
15019 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15020 combination of a flag and a thread ID variable.
15021 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15022 flag), *other* threads can enter ssleay_add_bytes without obeying
15023 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15024 that they do not hold after the first thread unsets add_do_not_lock).
15028 * Change bctest again: '-x' expressions are not available in all
15029 versions of 'test'.
15033 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15035 * Fix a couple of memory leaks in PKCS7_dataDecode()
15037 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15039 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15040 the default extension for executables, if any. Also, make the perl
15041 scripts that use symlink() to test if it really exists and use "cp"
15042 if it doesn't. All this made OpenSSL compilable and installable in
15047 * Fix for asn1_GetSequence() for indefinite length constructed data.
15048 If SEQUENCE is length is indefinite just set c->slen to the total
15049 amount of data available.
15051 *Steve Henson, reported by shige@FreeBSD.org*
15053 *This change does not apply to 0.9.7.*
15055 * Change bctest to avoid here-documents inside command substitution
15056 (workaround for FreeBSD /bin/sh bug).
15057 For compatibility with Ultrix, avoid shell functions (introduced
15058 in the bctest version that searches along $PATH).
15062 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15063 with des_encrypt() defined on some operating systems, like Solaris
15068 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15069 On the Importance of Eliminating Errors in Cryptographic
15070 Computations, J. Cryptology 14 (2001) 2, 101-119,
15071 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15075 * MIPS assembler BIGNUM division bug fix.
15079 * Disabled incorrect Alpha assembler code.
15083 * Fix PKCS#7 decode routines so they correctly update the length
15084 after reading an EOC for the EXPLICIT tag.
15088 *This change does not apply to 0.9.7.*
15090 * Fix bug in PKCS#12 key generation routines. This was triggered
15091 if a 3DES key was generated with a 0 initial byte. Include
15092 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15093 (but broken) behaviour.
15097 * Enhance bctest to search for a working bc along $PATH and print
15100 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15102 * Fix memory leaks in err.c: free err_data string if necessary;
15103 don't write to the wrong index in ERR_set_error_data.
15107 * Implement ssl23_peek (analogous to ssl23_read), which previously
15112 * Replace rdtsc with `_emit` statements for VC++ version 5.
15114 *Jeremy Cooper <jeremy@baymoo.org>*
15116 * Make it possible to reuse SSLv2 sessions.
15120 * In copy_email() check for >= 0 as a return value for
15121 X509_NAME_get_index_by_NID() since 0 is a valid index.
15123 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15125 * Avoid coredump with unsupported or invalid public keys by checking if
15126 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15127 PKCS7_verify() fails with non detached data.
15131 * Don't use getenv in library functions when run as setuid/setgid.
15132 New function OPENSSL_issetugid().
15136 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15137 due to incorrect handling of multi-threading:
15139 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15141 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15143 3. Count how many times MemCheck_off() has been called so that
15144 nested use can be treated correctly. This also avoids
15145 inband-signalling in the previous code (which relied on the
15146 assumption that thread ID 0 is impossible).
15150 * Add "-rand" option also to s_client and s_server.
15154 * Fix CPU detection on Irix 6.x.
15155 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15156 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15158 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15163 *This change does not apply to 0.9.7.*
15165 * Use the cached encoding of an X509_NAME structure rather than
15166 copying it. This is apparently the reason for the libsafe "errors"
15167 but the code is actually correct.
15171 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15172 Bleichenbacher's DSA attack.
15173 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15174 to be set and top=0 forces the highest bit to be set; top=-1 is new
15175 and leaves the highest bit random.
15177 *Ulf Moeller, Bodo Moeller*
15179 * In the `NCONF_...`-based implementations for `CONF_...` queries
15180 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15181 a temporary CONF structure with the data component set to NULL
15182 (which gives segmentation faults in lh_retrieve).
15183 Instead, use NULL for the CONF pointer in CONF_get_string and
15184 CONF_get_number (which may use environment variables) and directly
15185 return NULL from CONF_get_section.
15189 * Fix potential buffer overrun for EBCDIC.
15193 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15194 keyUsage if basicConstraints absent for a CA.
15198 * Make SMIME_write_PKCS7() write mail header values with a format that
15199 is more generally accepted (no spaces before the semicolon), since
15200 some programs can't parse those values properly otherwise. Also make
15201 sure BIO's that break lines after each write do not create invalid
15206 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15207 macros previously used would not encode an empty SEQUENCE OF
15208 and break the signature.
15212 *This change does not apply to 0.9.7.*
15214 * Zero the premaster secret after deriving the master secret in
15219 * Add some EVP_add_digest_alias registrations (as found in
15220 OpenSSL_add_all_digests()) to SSL_library_init()
15221 aka OpenSSL_add_ssl_algorithms(). This provides improved
15222 compatibility with peers using X.509 certificates
15223 with unconventional AlgorithmIdentifier OIDs.
15227 * Fix for Irix with NO_ASM.
15229 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15231 * ./config script fixes.
15233 *Ulf Moeller, Richard Levitte*
15235 * Fix 'openssl passwd -1'.
15239 * Change PKCS12_key_gen_asc() so it can cope with non null
15240 terminated strings whose length is passed in the passlen
15241 parameter, for example from PEM callbacks. This was done
15242 by adding an extra length parameter to asc2uni().
15244 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15246 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15247 call failed, free the DSA structure.
15251 * Fix to uni2asc() to cope with zero length Unicode strings.
15252 These are present in some PKCS#12 files.
15256 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15257 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15258 when writing a 32767 byte record.
15260 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15262 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15263 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15265 (RSA objects have a reference count access to which is protected
15266 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15267 so they are meant to be shared between threads.)
15268 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15269 "Reddie, Steven" <Steven.Reddie@ca.com>*
15271 * Fix a deadlock in CRYPTO_mem_leaks().
15275 * Use better test patterns in bntest.
15279 * rand_win.c fix for Borland C.
15283 * BN_rshift bugfix for n == 0.
15287 * Add a 'bctest' script that checks for some known 'bc' bugs
15288 so that 'make test' does not abort just because 'bc' is broken.
15292 * Store verify_result within SSL_SESSION also for client side to
15293 avoid potential security hole. (Re-used sessions on the client side
15294 always resulted in verify_result==X509_V_OK, not using the original
15295 result of the server certificate verification.)
15299 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15300 SSL3_RT_APPLICATION_DATA, return 0.
15301 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15306 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15307 releases, have been re-implemented by renaming the previous
15308 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15309 and ssl3_read_internal, respectively, and adding 'peek' parameters
15310 to them. The new ssl[23]_{read,peek} functions are calls to
15311 ssl[23]_read_internal with the 'peek' flag set appropriately.
15312 A 'peek' parameter has also been added to ssl3_read_bytes, which
15313 does the actual work for ssl3_read_internal.
15317 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15318 the method-specific "init()" handler. Also clean up ex_data after
15319 calling the method-specific "finish()" handler. Previously, this was
15320 happening the other way round.
15324 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15325 The previous value, 12, was not always sufficient for BN_mod_exp().
15329 * Make sure that shared libraries get the internal name engine with
15330 the full version number and not just 0. This should mark the
15331 shared libraries as not backward compatible. Of course, this should
15332 be changed again when we can guarantee backward binary compatibility.
15336 * Fix typo in get_cert_by_subject() in by_dir.c
15338 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15340 * Rework the system to generate shared libraries:
15342 - Make note of the expected extension for the shared libraries and
15343 if there is a need for symbolic links from for example libcrypto.so.0
15344 to libcrypto.so.0.9.7. There is extended info in Configure for
15347 - Make as few rebuilds of the shared libraries as possible.
15349 - Still avoid linking the OpenSSL programs with the shared libraries.
15351 - When installing, install the shared libraries separately from the
15356 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15358 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15359 and not in SSL_clear because the latter is also used by the
15360 accept/connect functions; previously, the settings made by
15361 SSL_set_read_ahead would be lost during the handshake.
15363 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15365 * Correct util/mkdef.pl to be selective about disabled algorithms.
15366 Previously, it would create entries for disabled algorithms no
15371 * Added several new manual pages for SSL_* function.
15375 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15377 * In ssl23_get_client_hello, generate an error message when faced
15378 with an initial SSL 3.0/TLS record that is too small to contain the
15379 first two bytes of the ClientHello message, i.e. client_version.
15380 (Note that this is a pathologic case that probably has never happened
15381 in real life.) The previous approach was to use the version number
15382 from the record header as a substitute; but our protocol choice
15383 should not depend on that one because it is not authenticated
15384 by the Finished messages.
15388 * More robust randomness gathering functions for Windows.
15390 *Jeffrey Altman <jaltman@columbia.edu>*
15392 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15393 not set then we don't setup the error code for issuer check errors
15394 to avoid possibly overwriting other errors which the callback does
15395 handle. If an application does set the flag then we assume it knows
15396 what it is doing and can handle the new informational codes
15401 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15402 a general "ANY" type, as such it should be able to decode anything
15403 including tagged types. However it didn't check the class so it would
15404 wrongly interpret tagged types in the same way as their universal
15405 counterpart and unknown types were just rejected. Changed so that the
15406 tagged and unknown types are handled in the same way as a SEQUENCE:
15407 that is the encoding is stored intact. There is also a new type
15408 "V_ASN1_OTHER" which is used when the class is not universal, in this
15409 case we have no idea what the actual type is so we just lump them all
15414 * On VMS, stdout may very well lead to a file that is written to
15415 in a record-oriented fashion. That means that every write() will
15416 write a separate record, which will be read separately by the
15417 programs trying to read from it. This can be very confusing.
15419 The solution is to put a BIO filter in the way that will buffer
15420 text until a linefeed is reached, and then write everything a
15421 line at a time, so every record written will be an actual line,
15422 not chunks of lines and not (usually doesn't happen, but I've
15423 seen it once) several lines in one record. BIO_f_linebuffer() is
15426 Currently, it's a VMS-only method, because that's where it has
15427 been tested well enough.
15431 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15432 it can return incorrect results.
15433 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15434 but it was in 0.9.6-beta[12].)
15438 * Disable the check for content being present when verifying detached
15439 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15440 include zero length content when signing messages.
15444 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15445 BIO_ctrl (for BIO pairs).
15449 * Add DSO method for VMS.
15453 * Bug fix: Montgomery multiplication could produce results with the
15458 * Add RPM specification openssl.spec and modify it to build three
15459 packages. The default package contains applications, application
15460 documentation and run-time libraries. The devel package contains
15461 include files, static libraries and function documentation. The
15462 doc package contains the contents of the doc directory. The original
15463 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15467 * Add a large number of documentation files for many SSL routines.
15469 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15471 * Add a configuration entry for Sony News 4.
15473 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15475 * Don't set the two most significant bits to one when generating a
15476 random number < q in the DSA library.
15480 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15481 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15482 the underlying transport is blocking) if a handshake took place.
15483 (The default behaviour is needed by applications such as s_client
15484 and s_server that use select() to determine when to use SSL_read;
15485 but for applications that know in advance when to expect data, it
15486 just makes things more complicated.)
15490 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15495 * Add a few more EBCDIC conditionals that make `req` and `x509`
15496 work better on such systems.
15498 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15500 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15501 Update PKCS12_parse() so it copies the friendlyName and the
15502 keyid to the certificates aux info.
15506 * Fix bug in PKCS7_verify() which caused an infinite loop
15507 if there was more than one signature.
15509 *Sven Uszpelkat <su@celocom.de>*
15511 * Major change in util/mkdef.pl to include extra information
15512 about each symbol, as well as presenting variables as well
15513 as functions. This change means that there's n more need
15514 to rebuild the .num files when some algorithms are excluded.
15518 * Allow the verify time to be set by an application,
15519 rather than always using the current time.
15523 * Phase 2 verify code reorganisation. The certificate
15524 verify code now looks up an issuer certificate by a
15525 number of criteria: subject name, authority key id
15526 and key usage. It also verifies self signed certificates
15527 by the same criteria. The main comparison function is
15528 X509_check_issued() which performs these checks.
15530 Lot of changes were necessary in order to support this
15531 without completely rewriting the lookup code.
15533 Authority and subject key identifier are now cached.
15535 The LHASH 'certs' is X509_STORE has now been replaced
15536 by a STACK_OF(X509_OBJECT). This is mainly because an
15537 LHASH can't store or retrieve multiple objects with
15538 the same hash value.
15540 As a result various functions (which were all internal
15541 use only) have changed to handle the new X509_STORE
15542 structure. This will break anything that messed round
15543 with X509_STORE internally.
15545 The functions X509_STORE_add_cert() now checks for an
15546 exact match, rather than just subject name.
15548 The X509_STORE API doesn't directly support the retrieval
15549 of multiple certificates matching a given criteria, however
15550 this can be worked round by performing a lookup first
15551 (which will fill the cache with candidate certificates)
15552 and then examining the cache for matches. This is probably
15553 the best we can do without throwing out X509_LOOKUP
15554 entirely (maybe later...).
15556 The X509_VERIFY_CTX structure has been enhanced considerably.
15558 All certificate lookup operations now go via a get_issuer()
15559 callback. Although this currently uses an X509_STORE it
15560 can be replaced by custom lookups. This is a simple way
15561 to bypass the X509_STORE hackery necessary to make this
15562 work and makes it possible to use more efficient techniques
15563 in future. A very simple version which uses a simple
15564 STACK for its trusted certificate store is also provided
15565 using X509_STORE_CTX_trusted_stack().
15567 The verify_cb() and verify() callbacks now have equivalents
15568 in the X509_STORE_CTX structure.
15570 X509_STORE_CTX also has a 'flags' field which can be used
15571 to customise the verify behaviour.
15575 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15576 excludes S/MIME capabilities.
15580 * When a certificate request is read in keep a copy of the
15581 original encoding of the signed data and use it when outputting
15582 again. Signatures then use the original encoding rather than
15583 a decoded, encoded version which may cause problems if the
15584 request is improperly encoded.
15588 * For consistency with other BIO_puts implementations, call
15589 buffer_write(b, ...) directly in buffer_puts instead of calling
15592 In BIO_puts, increment b->num_write as in BIO_write.
15594 *Peter.Sylvester@EdelWeb.fr*
15596 * Fix BN_mul_word for the case where the word is 0. (We have to use
15597 BN_zero, we may not return a BIGNUM with an array consisting of
15598 words set to zero.)
15602 * Avoid calling abort() from within the library when problems are
15603 detected, except if preprocessor symbols have been defined
15604 (such as REF_CHECK, BN_DEBUG etc.).
15608 * New openssl application 'rsautl'. This utility can be
15609 used for low-level RSA operations. DER public key
15610 BIO/fp routines also added.
15614 * New Configure entry and patches for compiling on QNX 4.
15616 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15618 * A demo state-machine implementation was sponsored by
15619 Nuron (<http://www.nuron.com/>) and is now available in
15620 demos/state_machine.
15624 * New options added to the 'dgst' utility for signature
15625 generation and verification.
15629 * Unrecognized PKCS#7 content types are now handled via a
15630 catch all ASN1_TYPE structure. This allows unsupported
15631 types to be stored as a "blob" and an application can
15632 encode and decode it manually.
15636 * Fix various signed/unsigned issues to make a_strex.c
15637 compile under VC++.
15639 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15641 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15642 length if passed a buffer. ASN1_INTEGER_to_BN failed
15643 if passed a NULL BN and its argument was negative.
15645 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15647 * Modification to PKCS#7 encoding routines to output definite
15648 length encoding. Since currently the whole structures are in
15649 memory there's not real point in using indefinite length
15650 constructed encoding. However if OpenSSL is compiled with
15651 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15655 * Added BIO_vprintf() and BIO_vsnprintf().
15659 * Added more prefixes to parse for in the strings written
15660 through a logging bio, to cover all the levels that are available
15661 through syslog. The prefixes are now:
15663 PANIC, EMERG, EMR => LOG_EMERG
15664 ALERT, ALR => LOG_ALERT
15665 CRIT, CRI => LOG_CRIT
15666 ERROR, ERR => LOG_ERR
15667 WARNING, WARN, WAR => LOG_WARNING
15668 NOTICE, NOTE, NOT => LOG_NOTICE
15669 INFO, INF => LOG_INFO
15670 DEBUG, DBG => LOG_DEBUG
15672 and as before, if none of those prefixes are present at the
15673 beginning of the string, LOG_ERR is chosen.
15675 On Win32, the `LOG_*` levels are mapped according to this:
15677 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15678 LOG_WARNING => EVENTLOG_WARNING_TYPE
15679 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15683 * Made it possible to reconfigure with just the configuration
15684 argument "reconf" or "reconfigure". The command line arguments
15685 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15686 and are retrieved from there when reconfiguring.
15692 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15694 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15698 * The obj_dat.pl script was messing up the sorting of object
15699 names. The reason was that it compared the quoted version
15700 of strings as a result "OCSP" > "OCSP Signing" because
15701 " > SPACE. Changed script to store unquoted versions of
15702 names and add quotes on output. It was also omitting some
15703 names from the lookup table if they were given a default
15704 value (that is if SN is missing it is given the same
15705 value as LN and vice versa), these are now added on the
15706 grounds that if an object has a name we should be able to
15707 look it up. Finally added warning output when duplicate
15708 short or long names are found.
15712 * Changes needed for Tandem NSK.
15714 *Scott Uroff <scott@xypro.com>*
15716 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15717 RSA_padding_check_SSLv23(), special padding was never detected
15718 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15719 version rollback attacks was not effective.
15721 In s23_clnt.c, don't use special rollback-attack detection padding
15722 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15723 client; similarly, in s23_srvr.c, don't do the rollback check if
15724 SSL 2.0 is the only protocol enabled in the server.
15728 * Make it possible to get hexdumps of unprintable data with 'openssl
15729 asn1parse'. By implication, the functions ASN1_parse_dump() and
15730 BIO_dump_indent() are added.
15734 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15735 these print out strings and name structures based on various
15736 flags including RFC2253 support and proper handling of
15737 multibyte characters. Added options to the 'x509' utility
15738 to allow the various flags to be set.
15742 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15743 Also change the functions X509_cmp_current_time() and
15744 X509_gmtime_adj() work with an ASN1_TIME structure,
15745 this will enable certificates using GeneralizedTime in validity
15746 dates to be checked.
15750 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15751 negative public key encodings) on by default,
15752 NO_NEG_PUBKEY_BUG can be set to disable it.
15756 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15757 content octets. An i2c_ASN1_OBJECT is unnecessary because
15758 the encoding can be trivially obtained from the structure.
15762 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15763 not read locks (`CRYPTO_r_[un]lock`).
15767 * A first attempt at creating official support for shared
15768 libraries through configuration. I've kept it so the
15769 default is static libraries only, and the OpenSSL programs
15770 are always statically linked for now, but there are
15771 preparations for dynamic linking in place.
15772 This has been tested on Linux and Tru64.
15776 * Randomness polling function for Win9x, as described in:
15777 Peter Gutmann, Software Generation of Practically Strong
15782 * Fix so PRNG is seeded in req if using an already existing
15787 * New options to smime application. -inform and -outform
15788 allow alternative formats for the S/MIME message including
15789 PEM and DER. The -content option allows the content to be
15790 specified separately. This should allow things like Netscape
15791 form signing output easier to verify.
15795 * Fix the ASN1 encoding of tags using the 'long form'.
15799 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15800 STRING types. These convert content octets to and from the
15801 underlying type. The actual tag and length octets are
15802 already assumed to have been read in and checked. These
15803 are needed because all other string types have virtually
15804 identical handling apart from the tag. By having versions
15805 of the ASN1 functions that just operate on content octets
15806 IMPLICIT tagging can be handled properly. It also allows
15807 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15808 and ASN1_INTEGER are identical apart from the tag.
15812 * Change the handling of OID objects as follows:
15814 - New object identifiers are inserted in objects.txt, following
15815 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15816 - objects.pl is used to process obj_mac.num and create a new
15818 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15821 This is currently kind of a hack, and the perl code in objects.pl
15822 isn't very elegant, but it works as I intended. The simplest way
15823 to check that it worked correctly is to look in obj_dat.h and
15824 check the array nid_objs and make sure the objects haven't moved
15825 around (this is important!). Additions are OK, as well as
15826 consistent name changes.
15830 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15834 * Addition of the command line parameter '-rand file' to 'openssl req'.
15835 The given file adds to whatever has already been seeded into the
15836 random pool through the RANDFILE configuration file option or
15837 environment variable, or the default random state file.
15841 * mkstack.pl now sorts each macro group into lexical order.
15842 Previously the output order depended on the order the files
15843 appeared in the directory, resulting in needless rewriting
15848 * Patches to make OpenSSL compile under Win32 again. Mostly
15849 work arounds for the VC++ problem that it treats func() as
15850 func(void). Also stripped out the parts of mkdef.pl that
15851 added extra typesafe functions: these no longer exist.
15855 * Reorganisation of the stack code. The macros are now all
15856 collected in safestack.h . Each macro is defined in terms of
15857 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15858 DEBUG_SAFESTACK is now handled in terms of function casts,
15859 this has the advantage of retaining type safety without the
15860 use of additional functions. If DEBUG_SAFESTACK is not defined
15861 then the non typesafe macros are used instead. Also modified the
15862 mkstack.pl script to handle the new form. Needs testing to see
15863 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15864 the default if no major problems. Similar behaviour for ASN1_SET_OF
15865 and PKCS12_STACK_OF.
15869 * When some versions of IIS use the 'NET' form of private key the
15870 key derivation algorithm is different. Normally MD5(password) is
15871 used as a 128 bit RC4 key. In the modified case
15872 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15873 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15874 as the old Netscape_RSA functions except they have an additional
15875 'sgckey' parameter which uses the modified algorithm. Also added
15876 an -sgckey command line option to the rsa utility. Thanks to
15877 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15878 algorithm to openssl-dev.
15882 * The evp_local.h macros were using 'c.##kname' which resulted in
15883 invalid expansion on some systems (SCO 5.0.5 for example).
15884 Corrected to 'c.kname'.
15886 *Phillip Porch <root@theporch.com>*
15888 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15889 a STACK of email addresses from a certificate or request, these look
15890 in the subject name and the subject alternative name extensions and
15891 omit any duplicate addresses.
15895 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15896 This makes DSA verification about 2 % faster.
15900 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15901 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15902 plus overhead for 1024 bit moduli).
15903 This makes exponentiations about 0.5 % faster for 1024 bit
15904 exponents (as measured by "openssl speed rsa2048").
15908 * Rename memory handling macros to avoid conflicts with other
15910 Malloc => OPENSSL_malloc
15911 Malloc_locked => OPENSSL_malloc_locked
15912 Realloc => OPENSSL_realloc
15913 Free => OPENSSL_free
15917 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15918 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15922 * CygWin32 support.
15924 *John Jarvie <jjarvie@newsguy.com>*
15926 * The type-safe stack code has been rejigged. It is now only compiled
15927 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15928 by default all type-specific stack functions are "#define"d back to
15929 standard stack functions. This results in more streamlined output
15930 but retains the type-safety checking possibilities of the original
15935 * The STACK code has been cleaned up, and certain type declarations
15936 that didn't make a lot of sense have been brought in line. This has
15937 also involved a cleanup of sorts in safestack.h to more correctly
15938 map type-safe stack functions onto their plain stack counterparts.
15939 This work has also resulted in a variety of "const"ifications of
15940 lots of the code, especially `_cmp` operations which should normally
15941 be prototyped with "const" parameters anyway.
15945 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15946 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15947 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15948 where all of 'md' is used each time the PRNG is used, but 'state'
15949 is used only indexed by a cyclic counter. As entropy may not be
15950 well distributed from the beginning, 'md' is important as a
15951 chaining variable. However, the output function chains only half
15952 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15953 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15954 in all of 'state' being rewritten, with the new values depending
15955 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15959 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15960 the handshake is continued after ssl_verify_cert_chain();
15961 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15962 can lead to 'unexplainable' connection aborts later.
15964 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15966 * Major EVP API cipher revision.
15967 Add hooks for extra EVP features. This allows various cipher
15968 parameters to be set in the EVP interface. Support added for variable
15969 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15970 setting of RC2 and RC5 parameters.
15972 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15975 Remove lots of duplicated code from the EVP library. For example *every*
15976 cipher init() function handles the 'iv' in the same way according to the
15977 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15978 for CFB and OFB modes they zero ctx->num.
15980 New functionality allows removal of S/MIME code RC2 hack.
15982 Most of the routines have the same form and so can be declared in terms
15985 By shifting this to the top level EVP_CipherInit() it can be removed from
15986 all individual ciphers. If the cipher wants to handle IVs or keys
15987 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15990 Change lots of functions like EVP_EncryptUpdate() to now return a
15991 value: although software versions of the algorithms cannot fail
15992 any installed hardware versions can.
15996 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15997 this option is set, tolerate broken clients that send the negotiated
15998 protocol version number instead of the requested protocol version
16003 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16004 i.e. non-zero for export ciphersuites, zero otherwise.
16005 Previous versions had this flag inverted, inconsistent with
16006 rsa_tmp_cb (..._TMP_RSA_CB).
16008 *Bodo Moeller; problem reported by Amit Chopra*
16010 * Add missing DSA library text string. Work around for some IIS
16011 key files with invalid SEQUENCE encoding.
16015 * Add a document (doc/standards.txt) that list all kinds of standards
16016 and so on that are implemented in OpenSSL.
16020 * Enhance c_rehash script. Old version would mishandle certificates
16021 with the same subject name hash and wouldn't handle CRLs at all.
16022 Added -fingerprint option to crl utility, to support new c_rehash
16027 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16031 * Fix for SSL server purpose checking. Server checking was
16032 rejecting certificates which had extended key usage present
16033 but no ssl client purpose.
16035 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16037 * Make PKCS#12 code work with no password. The PKCS#12 spec
16038 is a little unclear about how a blank password is handled.
16039 Since the password in encoded as a BMPString with terminating
16040 double NULL a zero length password would end up as just the
16041 double NULL. However no password at all is different and is
16042 handled differently in the PKCS#12 key generation code. NS
16043 treats a blank password as zero length. MSIE treats it as no
16044 password on export: but it will try both on import. We now do
16045 the same: PKCS12_parse() tries zero length and no password if
16046 the password is set to "" or NULL (NULL is now a valid password:
16047 it wasn't before) as does the pkcs12 application.
16051 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16052 perror when PEM_read_bio_X509_REQ fails, the error message must
16053 be obtained from the error queue.
16057 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16058 it in ERR_remove_state if appropriate, and change ERR_get_state
16059 accordingly to avoid race conditions (this is necessary because
16060 thread_hash is no longer constant once set).
16064 * Bugfix for linux-elf makefile.one.
16068 * RSA_get_default_method() will now cause a default
16069 RSA_METHOD to be chosen if one doesn't exist already.
16070 Previously this was only set during a call to RSA_new()
16071 or RSA_new_method(NULL) meaning it was possible for
16072 RSA_get_default_method() to return NULL.
16076 * Added native name translation to the existing DSO code
16077 that will convert (if the flag to do so is set) filenames
16078 that are sufficiently small and have no path information
16079 into a canonical native form. Eg. "blah" converted to
16080 "libblah.so" or "blah.dll" etc.
16084 * New function ERR_error_string_n(e, buf, len) which is like
16085 ERR_error_string(e, buf), but writes at most 'len' bytes
16086 including the 0 terminator. For ERR_error_string_n, 'buf'
16089 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16091 * CONF library reworked to become more general. A new CONF
16092 configuration file reader "class" is implemented as well as a
16093 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16094 old `CONF_*` functions are still there, but are reimplemented to
16095 work in terms of the new functions. Also, a set of functions
16096 to handle the internal storage of the configuration data is
16097 provided to make it easier to write new configuration file
16098 reader "classes" (I can definitely see something reading a
16099 configuration file in XML format, for example), called `_CONF_*`,
16100 or "the configuration storage API"...
16102 The new configuration file reading functions are:
16104 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16105 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16107 NCONF_default, NCONF_WIN32
16109 NCONF_dump_fp, NCONF_dump_bio
16111 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16112 NCONF_new creates a new CONF object. This works in the same way
16113 as other interfaces in OpenSSL, like the BIO interface.
16114 `NCONF_dump_*` dump the internal storage of the configuration file,
16115 which is useful for debugging. All other functions take the same
16116 arguments as the old `CONF_*` functions with the exception of the
16117 first that must be a `CONF *` instead of a `LHASH *`.
16119 To make it easier to use the new classes with the old `CONF_*` functions,
16120 the function CONF_set_default_method is provided.
16124 * Add '-tls1' option to 'openssl ciphers', which was already
16125 mentioned in the documentation but had not been implemented.
16126 (This option is not yet really useful because even the additional
16127 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16131 * Initial DSO code added into libcrypto for letting OpenSSL (and
16132 OpenSSL-based applications) load shared libraries and bind to
16133 them in a portable way.
16135 *Geoff Thorpe, with contributions from Richard Levitte*
16137 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16139 * Make sure _lrotl and _lrotr are only used with MSVC.
16141 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16142 (the default implementation of RAND_status).
16144 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16145 to '-clrext' (= clear extensions), as intended and documented.
16146 *Bodo Moeller; inconsistency pointed out by Michael Attili
16147 <attili@amaxo.com>*
16149 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16150 was larger than the MD block size.
16152 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16154 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16155 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16156 using the passed key: if the passed key was a private key the result
16157 of X509_print(), for example, would be to print out all the private key
16162 * des_quad_cksum() byte order bug fix.
16163 *Ulf Möller, using the problem description in krb4-0.9.7, where
16164 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16166 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16169 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16171 * For easily testing in shell scripts whether some command
16172 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16173 returns with exit code 0 iff no command of the given name is available.
16174 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16175 the output goes to stdout and nothing is printed to stderr.
16176 Additional arguments are always ignored.
16178 Since for each cipher there is a command of the same name,
16179 the 'no-cipher' compilation switches can be tested this way.
16181 ('openssl no-XXX' is not able to detect pseudo-commands such
16182 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16186 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16190 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16191 is set; it will be thrown away anyway because each handshake creates
16193 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16194 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16195 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16196 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16200 * New s_client option -ign_eof: EOF at stdin is ignored, and
16201 'Q' and 'R' lose their special meanings (quit/renegotiate).
16202 This is part of what -quiet does; unlike -quiet, -ign_eof
16203 does not suppress any output.
16207 * Add compatibility options to the purpose and trust code. The
16208 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16209 accepts a certificate or CA, this was the previous behaviour,
16210 with all the associated security issues.
16212 X509_TRUST_COMPAT is the old trust behaviour: only and
16213 automatically trust self signed roots in certificate store. A
16214 new trust setting X509_TRUST_DEFAULT is used to specify that
16215 a purpose has no associated trust setting and it should instead
16216 use the value in the default purpose.
16220 * Fix the PKCS#8 DSA private key code so it decodes keys again
16221 and fix a memory leak.
16225 * In util/mkerr.pl (which implements 'make errors'), preserve
16226 reason strings from the previous version of the .c file, as
16227 the default to have only downcase letters (and digits) in
16228 automatically generated reasons codes is not always appropriate.
16232 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16233 using strerror. Previously, ERR_reason_error_string() returned
16234 library names as reason strings for SYSerr; but SYSerr is a special
16235 case where small numbers are errno values, not library numbers.
16239 * Add '-dsaparam' option to 'openssl dhparam' application. This
16240 converts DSA parameters into DH parameters. (When creating parameters,
16241 DSA_generate_parameters is used.)
16245 * Include 'length' (recommended exponent length) in C code generated
16246 by 'openssl dhparam -C'.
16250 * The second argument to set_label in perlasm was already being used
16251 so couldn't be used as a "file scope" flag. Moved to third argument
16256 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16257 instead of RAND_bytes for encryption IVs and salts.
16261 * Include RAND_status() into RAND_METHOD instead of implementing
16262 it only for md_rand.c Otherwise replacing the PRNG by calling
16263 RAND_set_rand_method would be impossible.
16267 * Don't let DSA_generate_key() enter an infinite loop if the random
16268 number generation fails.
16272 * New 'rand' application for creating pseudo-random output.
16276 * Added configuration support for Linux/IA64
16278 *Rolf Haberrecker <rolf@suse.de>*
16280 * Assembler module support for Mingw32.
16284 * Shared library support for HPUX (in shlib/).
16286 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16288 * Shared library support for Solaris gcc.
16290 *Lutz Behnke <behnke@trustcenter.de>*
16292 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16294 * PKCS7_encrypt() was adding text MIME headers twice because they
16295 were added manually and by SMIME_crlf_copy().
16299 * In bntest.c don't call BN_rand with zero bits argument.
16301 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16303 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16304 case was implemented. This caused BN_div_recp() to fail occasionally.
16308 * Add an optional second argument to the set_label() in the perl
16309 assembly language builder. If this argument exists and is set
16310 to 1 it signals that the assembler should use a symbol whose
16311 scope is the entire file, not just the current function. This
16312 is needed with MASM which uses the format label:: for this scope.
16314 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16316 * Change the ASN1 types so they are typedefs by default. Before
16317 almost all types were #define'd to ASN1_STRING which was causing
16318 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16323 * Change names of new functions to the new get1/get0 naming
16324 convention: After 'get1', the caller owns a reference count
16325 and has to call `..._free`; 'get0' returns a pointer to some
16326 data structure without incrementing reference counters.
16327 (Some of the existing 'get' functions increment a reference
16328 counter, some don't.)
16329 Similarly, 'set1' and 'add1' functions increase reference
16330 counters or duplicate objects.
16334 * Allow for the possibility of temp RSA key generation failure:
16335 the code used to assume it always worked and crashed on failure.
16339 * Fix potential buffer overrun problem in BIO_printf().
16340 *Ulf Möller, using public domain code by Patrick Powell; problem
16341 pointed out by David Sacerdote <das33@cornell.edu>*
16343 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16344 RAND_egd() and RAND_status(). In the command line application,
16345 the EGD socket can be specified like a seed file using RANDFILE
16350 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16351 Some CAs (e.g. Verisign) distribute certificates in this form.
16355 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16356 list to exclude them. This means that no special compilation option
16357 is needed to use anonymous DH: it just needs to be included in the
16362 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16363 EVP_MD_type. The old functionality is available in a new macro called
16364 EVP_MD_md(). Change code that uses it and update docs.
16368 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16369 where the `void *` argument is replaced by a function pointer argument.
16370 Previously `void *` was abused to point to functions, which works on
16371 many platforms, but is not correct. As these functions are usually
16372 called by macros defined in OpenSSL header files, most source code
16373 should work without changes.
16377 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16378 sections with information on -D... compiler switches used for
16379 compiling the library so that applications can see them. To enable
16380 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16381 must be defined. E.g.,
16382 #define OPENSSL_ALGORITHM_DEFINES
16383 #include <openssl/opensslconf.h>
16384 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16386 *Richard Levitte, Ulf and Bodo Möller*
16388 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16393 * Change the 'other' type in certificate aux info to a STACK_OF
16394 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16395 the required ASN1 format: arbitrary types determined by an OID.
16399 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16400 argument to 'req'. This is not because the function is newer or
16401 better than others it just uses the work 'NEW' in the certificate
16402 request header lines. Some software needs this.
16406 * Reorganise password command line arguments: now passwords can be
16407 obtained from various sources. Delete the PEM_cb function and make
16408 it the default behaviour: i.e. if the callback is NULL and the
16409 usrdata argument is not NULL interpret it as a null terminated pass
16410 phrase. If usrdata and the callback are NULL then the pass phrase
16411 is prompted for as usual.
16415 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16416 the support is automatically enabled. The resulting binaries will
16417 autodetect the card and use it if present.
16419 *Ben Laurie and Compaq Inc.*
16421 * Work around for Netscape hang bug. This sends certificate request
16422 and server done in one record. Since this is perfectly legal in the
16423 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16424 the bugs/SSLv3 entry for more info.
16428 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16432 * Add -rand argument to smime and pkcs12 applications and read/write
16437 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16441 * Add command line password options to the remaining applications.
16445 * Bug fix for BN_div_recp() for numerators with an even number of
16450 * More tests in bntest.c, and changed test_bn output.
16454 * ./config recognizes MacOS X now.
16458 * Bug fix for BN_div() when the first words of num and divisor are
16459 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16463 * Add support for various broken PKCS#8 formats, and command line
16464 options to produce them.
16468 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16469 get temporary BIGNUMs from a BN_CTX.
16473 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16478 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16479 include a #define from the old name to the new. The original intent
16480 was that statically linked binaries could for example just call
16481 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16482 link with digests. This never worked because SSLeay_add_all_digests()
16483 and SSLeay_add_all_ciphers() were in the same source file so calling
16484 one would link with the other. They are now in separate source files.
16488 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16492 * Use a less unusual form of the Miller-Rabin primality test (it used
16493 a binary algorithm for exponentiation integrated into the Miller-Rabin
16494 loop, our standard modexp algorithms are faster).
16498 * Support for the EBCDIC character set completed.
16500 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16502 * Source code cleanups: use const where appropriate, eliminate casts,
16503 use `void *` instead of `char *` in lhash.
16507 * Bugfix: ssl3_send_server_key_exchange was not restartable
16508 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16509 this the server could overwrite ephemeral keys that the client
16514 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16515 using 50 iterations of the Rabin-Miller test.
16517 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16518 iterations of the Rabin-Miller test as required by the appendix
16519 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16520 As BN_is_prime_fasttest includes trial division, DSA parameter
16521 generation becomes much faster.
16523 This implies a change for the callback functions in DSA_is_prime
16524 and DSA_generate_parameters: The callback function is called once
16525 for each positive witness in the Rabin-Miller test, not just
16526 occasionally in the inner loop; and the parameters to the
16527 callback function now provide an iteration count for the outer
16528 loop rather than for the current invocation of the inner loop.
16529 DSA_generate_parameters additionally can call the callback
16530 function with an 'iteration count' of -1, meaning that a
16531 candidate has passed the trial division test (when q is generated
16532 from an application-provided seed, trial division is skipped).
16536 * New function BN_is_prime_fasttest that optionally does trial
16537 division before starting the Rabin-Miller test and has
16538 an additional BN_CTX * argument (whereas BN_is_prime always
16539 has to allocate at least one BN_CTX).
16540 'callback(1, -1, cb_arg)' is called when a number has passed the
16541 trial division stage.
16545 * Fix for bug in CRL encoding. The validity dates weren't being handled
16550 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16554 * New function BN_pseudo_rand().
16558 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16559 bignum version of BN_from_montgomery() with the working code from
16560 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16565 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16566 made it impossible to use the same SSL_SESSION data structure in
16567 SSL2 clients in multiple threads.
16571 * The return value of RAND_load_file() no longer counts bytes obtained
16572 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16573 to seed the PRNG (previously an explicit byte count was required).
16575 *Ulf Möller, Bodo Möller*
16577 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16578 used `char *` instead of `void *` and had casts all over the place.
16582 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16586 * Retain source code compatibility for BN_prime_checks macro:
16587 BN_is_prime(..., BN_prime_checks, ...) now uses
16588 BN_prime_checks_for_size to determine the appropriate number of
16589 Rabin-Miller iterations.
16593 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16594 DH_CHECK_P_NOT_SAFE_PRIME.
16595 (Check if this is true? OpenPGP calls them "strong".)
16599 * Merge the functionality of "dh" and "gendh" programs into a new program
16600 "dhparam". The old programs are retained for now but will handle DH keys
16601 (instead of parameters) in future.
16605 * Make the ciphers, s_server and s_client programs check the return values
16606 when a new cipher list is set.
16610 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16611 ciphers. Before when the 56bit ciphers were enabled the sorting was
16614 The syntax for the cipher sorting has been extended to support sorting by
16615 cipher-strength (using the strength_bits hard coded in the tables).
16616 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16618 Fix a bug in the cipher-command parser: when supplying a cipher command
16619 string with an "undefined" symbol (neither command nor alphanumeric
16620 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16621 an error is flagged.
16623 Due to the strength-sorting extension, the code of the
16624 ssl_create_cipher_list() function was completely rearranged. I hope that
16625 the readability was also increased :-)
16627 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16629 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16630 for the first serial number and places 2 in the serial number file. This
16631 avoids problems when the root CA is created with serial number zero and
16632 the first user certificate has the same issuer name and serial number
16637 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16638 the new code. Add documentation for this stuff.
16642 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16643 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16644 structures and behave in an analogous way to the X509v3 functions:
16645 they shouldn't be called directly but wrapper functions should be used
16648 So we also now have some wrapper functions that call the X509at functions
16649 when passed certificate requests. (TO DO: similar things can be done with
16650 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16651 things. Some of these need some d2i or i2d and print functionality
16652 because they handle more complex structures.)
16656 * Add missing #ifndefs that caused missing symbols when building libssl
16657 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16658 NO_RSA in `ssl/s2*.c`.
16660 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16662 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16663 has a return value which indicates the quality of the random data
16664 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16665 error queue. New function RAND_pseudo_bytes() generates output that is
16666 guaranteed to be unique but not unpredictable. RAND_add is like
16667 RAND_seed, but takes an extra argument for an entropy estimate
16668 (RAND_seed always assumes full entropy).
16672 * Do more iterations of Rabin-Miller probable prime test (specifically,
16673 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16674 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16675 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16676 false-positive rate of at most 2^-80 for random input.
16680 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16684 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16685 in the 0.9.5 release), this returns the chain
16686 from an X509_CTX structure with a dup of the stack and all
16687 the X509 reference counts upped: so the stack will exist
16688 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16691 Also make SSL_SESSION_print() print out the verify return
16696 * Add manpage for the pkcs12 command. Also change the default
16697 behaviour so MAC iteration counts are used unless the new
16698 -nomaciter option is used. This improves file security and
16699 only older versions of MSIE (4.0 for example) need it.
16703 * Honor the no-xxx Configure options when creating .DEF files.
16707 * Add PKCS#10 attributes to field table: challengePassword,
16708 unstructuredName and unstructuredAddress. These are taken from
16709 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16710 international characters are used.
16712 More changes to X509_ATTRIBUTE code: allow the setting of types
16713 based on strings. Remove the 'loc' parameter when adding
16714 attributes because these will be a SET OF encoding which is sorted
16719 * Initial changes to the 'req' utility to allow request generation
16720 automation. This will allow an application to just generate a template
16721 file containing all the field values and have req construct the
16724 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16725 used all over the place including certificate requests and PKCS#7
16726 structures. They are currently handled manually where necessary with
16727 some primitive wrappers for PKCS#7. The new functions behave in a
16728 manner analogous to the X509 extension functions: they allow
16729 attributes to be looked up by NID and added.
16731 Later something similar to the X509V3 code would be desirable to
16732 automatically handle the encoding, decoding and printing of the
16733 more complex types. The string types like challengePassword can
16734 be handled by the string table functions.
16736 Also modified the multi byte string table handling. Now there is
16737 a 'global mask' which masks out certain types. The table itself
16738 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16739 is useful when for example there is only one permissible type
16740 (as in countryName) and using the mask might result in no valid
16745 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16746 SSL_get_peer_finished to allow applications to obtain the latest
16747 Finished messages sent to the peer or expected from the peer,
16748 respectively. (SSL_get_peer_finished is usually the Finished message
16749 actually received from the peer, otherwise the protocol will be aborted.)
16751 As the Finished message are message digests of the complete handshake
16752 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16753 be used for external authentication procedures when the authentication
16754 provided by SSL/TLS is not desired or is not enough.
16758 * Enhanced support for Alpha Linux is added. Now ./config checks if
16759 the host supports BWX extension and if Compaq C is present on the
16760 $PATH. Just exploiting of the BWX extension results in 20-30%
16761 performance kick for some algorithms, e.g. DES and RC4 to mention
16762 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16767 * Add support for MS "fast SGC". This is arguably a violation of the
16768 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16769 weak crypto and after checking the certificate is SGC a second one
16770 with strong crypto. MS SGC stops the first handshake after receiving
16771 the server certificate message and sends a second client hello. Since
16772 a server will typically do all the time consuming operations before
16773 expecting any further messages from the client (server key exchange
16774 is the most expensive) there is little difference between the two.
16776 To get OpenSSL to support MS SGC we have to permit a second client
16777 hello message after we have sent server done. In addition we have to
16778 reset the MAC if we do get this second client hello.
16782 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16783 if a DER encoded private key is RSA or DSA traditional format. Changed
16784 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16785 format DER encoded private key. Newer code should use PKCS#8 format which
16786 has the key type encoded in the ASN1 structure. Added DER private key
16787 support to pkcs8 application.
16791 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16792 ciphersuites has been selected (as required by the SSL 3/TLS 1
16793 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16794 is set, we interpret this as a request to violate the specification
16795 (the worst that can happen is a handshake failure, and 'correct'
16796 behaviour would result in a handshake failure anyway).
16800 * In SSL_CTX_add_session, take into account that there might be multiple
16801 SSL_SESSION structures with the same session ID (e.g. when two threads
16802 concurrently obtain them from an external cache).
16803 The internal cache can handle only one SSL_SESSION with a given ID,
16804 so if there's a conflict, we now throw out the old one to achieve
16809 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16810 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16811 some routines that use cipher OIDs: some ciphers do not have OIDs
16812 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16817 * Simplify the trust setting structure and code. Now we just have
16818 two sequences of OIDs for trusted and rejected settings. These will
16819 typically have values the same as the extended key usage extension
16820 and any application specific purposes.
16822 The trust checking code now has a default behaviour: it will just
16823 check for an object with the same NID as the passed id. Functions can
16824 be provided to override either the default behaviour or the behaviour
16825 for a given id. SSL client, server and email already have functions
16826 in place for compatibility: they check the NID and also return "trusted"
16827 if the certificate is self signed.
16831 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16832 traditional format into an EVP_PKEY structure.
16836 * Add a password callback function PEM_cb() which either prompts for
16837 a password if usr_data is NULL or otherwise assumes it is a null
16838 terminated password. Allow passwords to be passed on command line
16839 environment or config files in a few more utilities.
16843 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16844 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16845 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16846 Update documentation.
16850 * Support for ASN1 "NULL" type. This could be handled before by using
16851 ASN1_TYPE but there wasn't any function that would try to read a NULL
16852 and produce an error if it couldn't. For compatibility we also have
16853 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16854 don't allocate anything because they don't need to.
16858 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16861 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16863 * Rebuild of the memory allocation routines used by OpenSSL code and
16864 possibly others as well. The purpose is to make an interface that
16865 provide hooks so anyone can build a separate set of allocation and
16866 deallocation routines to be used by OpenSSL, for example memory
16867 pool implementations, or something else, which was previously hard
16868 since Malloc(), Realloc() and Free() were defined as macros having
16869 the values malloc, realloc and free, respectively (except for Win32
16870 compilations). The same is provided for memory debugging code.
16871 OpenSSL already comes with functionality to find memory leaks, but
16872 this gives people a chance to debug other memory problems.
16874 With these changes, a new set of functions and macros have appeared:
16876 CRYPTO_set_mem_debug_functions() [F]
16877 CRYPTO_get_mem_debug_functions() [F]
16878 CRYPTO_dbg_set_options() [F]
16879 CRYPTO_dbg_get_options() [F]
16880 CRYPTO_malloc_debug_init() [M]
16882 The memory debug functions are NULL by default, unless the library
16883 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16884 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16885 gives the standard debugging functions that come with OpenSSL) or
16886 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16887 provided by the library user) must be used. When the standard
16888 debugging functions are used, CRYPTO_dbg_set_options can be used to
16889 request additional information:
16890 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16891 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16893 Also, things like CRYPTO_set_mem_functions will always give the
16894 expected result (the new set of functions is used for allocation
16895 and deallocation) at all times, regardless of platform and compiler
16898 To finish it up, some functions that were never use in any other
16899 way than through macros have a new API and new semantic:
16901 CRYPTO_dbg_malloc()
16902 CRYPTO_dbg_realloc()
16905 All macros of value have retained their old syntax.
16907 *Richard Levitte and Bodo Moeller*
16909 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16910 ordering of SMIMECapabilities wasn't in "strength order" and there
16911 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16916 * Some ASN1 types with illegal zero length encoding (INTEGER,
16917 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16919 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16921 * Merge in my S/MIME library for OpenSSL. This provides a simple
16922 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16923 functionality to handle multipart/signed properly) and a utility
16924 called 'smime' to call all this stuff. This is based on code I
16925 originally wrote for Celo who have kindly allowed it to be
16926 included in OpenSSL.
16930 * Add variants des_set_key_checked and des_set_key_unchecked of
16931 des_set_key (aka des_key_sched). Global variable des_check_key
16932 decides which of these is called by des_set_key; this way
16933 des_check_key behaves as it always did, but applications and
16934 the library itself, which was buggy for des_check_key == 1,
16935 have a cleaner way to pick the version they need.
16939 * New function PKCS12_newpass() which changes the password of a
16944 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16945 dynamic mix. In both cases the ids can be used as an index into the
16946 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16947 functions so they accept a list of the field values and the
16948 application doesn't need to directly manipulate the X509_TRUST
16953 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16958 * Modify the way the V3 extension code looks up extensions. This now
16959 works in a similar way to the object code: we have some "standard"
16960 extensions in a static table which is searched with OBJ_bsearch()
16961 and the application can add dynamic ones if needed. The file
16962 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16963 updated whenever a new extension is added to the core code and kept
16964 in ext_nid order. There is a simple program 'tabtest.c' which checks
16965 this. New extensions are not added too often so this file can readily
16966 be maintained manually.
16968 There are two big advantages in doing things this way. The extensions
16969 can be looked up immediately and no longer need to be "added" using
16970 X509V3_add_standard_extensions(): this function now does nothing.
16971 Side note: I get *lots* of email saying the extension code doesn't
16972 work because people forget to call this function.
16973 Also no dynamic allocation is done unless new extensions are added:
16974 so if we don't add custom extensions there is no need to call
16975 X509V3_EXT_cleanup().
16979 * Modify enc utility's salting as follows: make salting the default. Add a
16980 magic header, so unsalted files fail gracefully instead of just decrypting
16981 to garbage. This is because not salting is a big security hole, so people
16982 should be discouraged from doing it.
16986 * Fixes and enhancements to the 'x509' utility. It allowed a message
16987 digest to be passed on the command line but it only used this
16988 parameter when signing a certificate. Modified so all relevant
16989 operations are affected by the digest parameter including the
16990 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16991 DSA key was used because it didn't fix the digest.
16995 * Initial certificate chain verify code. Currently tests the untrusted
16996 certificates for consistency with the verify purpose (which is set
16997 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16999 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17000 this is because it will reject chains with invalid extensions whereas
17001 every previous version of OpenSSL and SSLeay made no checks at all.
17003 Trust code: checks the root CA for the relevant trust settings. Trust
17004 settings have an initial value consistent with the verify purpose: e.g.
17005 if the verify purpose is for SSL client use it expects the CA to be
17006 trusted for SSL client use. However the default value can be changed to
17007 permit custom trust settings: one example of this would be to only trust
17008 certificates from a specific "secure" set of CAs.
17010 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17011 which should be used for version portability: especially since the
17012 verify structure is likely to change more often now.
17014 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17015 to set them. If not set then assume SSL clients will verify SSL servers
17018 Two new options to the verify program: -untrusted allows a set of
17019 untrusted certificates to be passed in and -purpose which sets the
17020 intended purpose of the certificate. If a purpose is set then the
17021 new chain verify code is used to check extension consistency.
17025 * Support for the authority information access extension.
17029 * Modify RSA and DSA PEM read routines to transparently handle
17030 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17031 public keys in a format compatible with certificate
17032 SubjectPublicKeyInfo structures. Unfortunately there were already
17033 functions called *_PublicKey_* which used various odd formats so
17034 these are retained for compatibility: however the DSA variants were
17035 never in a public release so they have been deleted. Changed dsa/rsa
17036 utilities to handle the new format: note no releases ever handled public
17037 keys so we should be OK.
17039 The primary motivation for this change is to avoid the same fiasco
17040 that dogs private keys: there are several incompatible private key
17041 formats some of which are standard and some OpenSSL specific and
17042 require various evil hacks to allow partial transparent handling and
17043 even then it doesn't work with DER formats. Given the option anything
17044 other than PKCS#8 should be dumped: but the other formats have to
17045 stay in the name of compatibility.
17047 With public keys and the benefit of hindsight one standard format
17048 is used which works with EVP_PKEY, RSA or DSA structures: though
17049 it clearly returns an error if you try to read the wrong kind of key.
17051 Added a -pubkey option to the 'x509' utility to output the public key.
17052 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17053 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17054 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17055 that do the same as the `EVP_PKEY_assign_*()` except they up the
17056 reference count of the added key (they don't "swallow" the
17061 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17062 CRLs would fail if the file contained no certificates or no CRLs:
17063 added a new function to read in both types and return the number
17064 read: this means that if none are read it will be an error. The
17065 DER versions of the certificate and CRL reader would always fail
17066 because it isn't possible to mix certificates and CRLs in DER format
17067 without choking one or the other routine. Changed this to just read
17068 a certificate: this is the best we can do. Also modified the code
17069 in `apps/verify.c` to take notice of return codes: it was previously
17070 attempting to read in certificates from NULL pointers and ignoring
17071 any errors: this is one reason why the cert and CRL reader seemed
17072 to work. It doesn't check return codes from the default certificate
17073 routines: these may well fail if the certificates aren't installed.
17077 * Code to support otherName option in GeneralName.
17081 * First update to verify code. Change the verify utility
17082 so it warns if it is passed a self signed certificate:
17083 for consistency with the normal behaviour. X509_verify
17084 has been modified to it will now verify a self signed
17085 certificate if *exactly* the same certificate appears
17086 in the store: it was previously impossible to trust a
17087 single self signed certificate. This means that:
17088 openssl verify ss.pem
17089 now gives a warning about a self signed certificate but
17090 openssl verify -CAfile ss.pem ss.pem
17095 * For servers, store verify_result in SSL_SESSION data structure
17096 (and add it to external session representation).
17097 This is needed when client certificate verifications fails,
17098 but an application-provided verification callback (set by
17099 SSL_CTX_set_cert_verify_callback) allows accepting the session
17100 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17101 but returns 1): When the session is reused, we have to set
17102 ssl->verify_result to the appropriate error code to avoid
17105 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17107 * Fix a bug in the new PKCS#7 code: it didn't consider the
17108 case in PKCS7_dataInit() where the signed PKCS7 structure
17109 didn't contain any existing data because it was being created.
17111 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17113 * Add a salt to the key derivation routines in enc.c. This
17114 forms the first 8 bytes of the encrypted file. Also add a
17115 -S option to allow a salt to be input on the command line.
17119 * New function X509_cmp(). Oddly enough there wasn't a function
17120 to compare two certificates. We do this by working out the SHA1
17121 hash and comparing that. X509_cmp() will be needed by the trust
17126 * SSL_get1_session() is like SSL_get_session(), but increments
17127 the reference count in the SSL_SESSION returned.
17129 *Geoff Thorpe <geoff@eu.c2.net>*
17131 * Fix for 'req': it was adding a null to request attributes.
17132 Also change the X509_LOOKUP and X509_INFO code to handle
17133 certificate auxiliary information.
17137 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17142 * Add the possibility to add extra information to the memory leak
17143 detecting output, to form tracebacks, showing from where each
17144 allocation was originated: CRYPTO_push_info("constant string") adds
17145 the string plus current file name and line number to a per-thread
17146 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17147 is like calling CYRPTO_pop_info() until the stack is empty.
17148 Also updated memory leak detection code to be multi-thread-safe.
17152 * Add options -text and -noout to pkcs7 utility and delete the
17153 encryption options which never did anything. Update docs.
17157 * Add options to some of the utilities to allow the pass phrase
17158 to be included on either the command line (not recommended on
17159 OSes like Unix) or read from the environment. Update the
17160 manpages and fix a few bugs.
17164 * Add a few manpages for some of the openssl commands.
17168 * Fix the -revoke option in ca. It was freeing up memory twice,
17169 leaking and not finding already revoked certificates.
17173 * Extensive changes to support certificate auxiliary information.
17174 This involves the use of X509_CERT_AUX structure and X509_AUX
17175 functions. An X509_AUX function such as PEM_read_X509_AUX()
17176 can still read in a certificate file in the usual way but it
17177 will also read in any additional "auxiliary information". By
17178 doing things this way a fair degree of compatibility can be
17179 retained: existing certificates can have this information added
17180 using the new 'x509' options.
17182 Current auxiliary information includes an "alias" and some trust
17183 settings. The trust settings will ultimately be used in enhanced
17184 certificate chain verification routines: currently a certificate
17185 can only be trusted if it is self signed and then it is trusted
17190 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17191 The problem was that one of the replacement routines had not been working
17192 since SSLeay releases. For now the offending routine has been replaced
17193 with non-optimised assembler. Even so, this now gives around 95%
17194 performance improvement for 1024 bit RSA signs.
17198 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17199 handling. Most clients have the effective key size in bits equal to
17200 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17201 A few however don't do this and instead use the size of the decrypted key
17202 to determine the RC2 key length and the AlgorithmIdentifier to determine
17203 the effective key length. In this case the effective key length can still
17204 be 40 bits but the key length can be 168 bits for example. This is fixed
17205 by manually forcing an RC2 key into the EVP_PKEY structure because the
17206 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17207 the key length and effective key length are equal.
17211 * Add a bunch of functions that should simplify the creation of
17212 X509_NAME structures. Now you should be able to do:
17213 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17214 and have it automatically work out the correct field type and fill in
17215 the structures. The more adventurous can try:
17216 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17217 and it will (hopefully) work out the correct multibyte encoding.
17221 * Change the 'req' utility to use the new field handling and multibyte
17222 copy routines. Before the DN field creation was handled in an ad hoc
17223 way in req, ca, and x509 which was rather broken and didn't support
17224 BMPStrings or UTF8Strings. Since some software doesn't implement
17225 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17226 using the dirstring_type option. See the new comment in the default
17227 openssl.cnf for more info.
17231 * Make crypto/rand/md_rand.c more robust:
17232 - Assure unique random numbers after fork().
17233 - Make sure that concurrent threads access the global counter and
17234 md serializably so that we never lose entropy in them
17235 or use exactly the same state in multiple threads.
17236 Access to the large state is not always serializable because
17237 the additional locking could be a performance killer, and
17238 md should be large enough anyway.
17242 * New file `apps/app_rand.c` with commonly needed functionality
17243 for handling the random seed file.
17245 Use the random seed file in some applications that previously did not:
17247 dsaparam -genkey (which also ignored its '-rand' option),
17250 x509 (when signing).
17251 Except on systems with /dev/urandom, it is crucial to have a random
17252 seed file at least for key creation, DSA signing, and for DH exchanges;
17253 for RSA signatures we could do without one.
17255 gendh and gendsa (unlike genrsa) used to read only the first byte
17256 of each file listed in the '-rand' option. The function as previously
17257 found in genrsa is now in app_rand.c and is used by all programs
17258 that support '-rand'.
17262 * In RAND_write_file, use mode 0600 for creating files;
17263 don't just chmod when it may be too late.
17267 * Report an error from X509_STORE_load_locations
17268 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17272 * New function ASN1_mbstring_copy() this copies a string in either
17273 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17274 into an ASN1_STRING type. A mask of permissible types is passed
17275 and it chooses the "minimal" type to use or an error if not type
17280 * Add function equivalents to the various macros in asn1.h. The old
17281 macros are retained with an `M_` prefix. Code inside the library can
17282 use the `M_` macros. External code (including the openssl utility)
17283 should *NOT* in order to be "shared library friendly".
17287 * Add various functions that can check a certificate's extensions
17288 to see if it usable for various purposes such as SSL client,
17289 server or S/MIME and CAs of these types. This is currently
17290 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17291 verification. Also added a -purpose flag to x509 utility to
17292 print out all the purposes.
17296 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17301 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17302 for, obtain and decode and extension and obtain its critical flag.
17303 This allows all the necessary extension code to be handled in a
17304 single function call.
17308 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17309 platforms. See crypto/rc4/rc4_enc.c for further details.
17313 * New -noout option to asn1parse. This causes no output to be produced
17314 its main use is when combined with -strparse and -out to extract data
17315 from a file (which may not be in ASN.1 format).
17319 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17320 when producing the local key id.
17322 *Richard Levitte <levitte@stacken.kth.se>*
17324 * New option -dhparam in s_server. This allows a DH parameter file to be
17325 stated explicitly. If it is not stated then it tries the first server
17326 certificate file. The previous behaviour hard coded the filename
17331 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17332 a public key to be input or output. For example:
17333 openssl rsa -in key.pem -pubout -out pubkey.pem
17334 Also added necessary DSA public key functions to handle this.
17338 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17339 in the message. This was handled by allowing
17340 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17342 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17344 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17345 to the end of the strings whereas this didn't. This would cause problems
17346 if strings read with d2i_ASN1_bytes() were later modified.
17348 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17350 * Fix for base64 decode bug. When a base64 bio reads only one line of
17351 data and it contains EOF it will end up returning an error. This is
17352 caused by input 46 bytes long. The cause is due to the way base64
17353 BIOs find the start of base64 encoded data. They do this by trying a
17354 trial decode on each line until they find one that works. When they
17355 do a flag is set and it starts again knowing it can pass all the
17356 data directly through the decoder. Unfortunately it doesn't reset
17357 the context it uses. This means that if EOF is reached an attempt
17358 is made to pass two EOFs through the context and this causes the
17359 resulting error. This can also cause other problems as well. As is
17360 usual with these problems it takes *ages* to find and the fix is
17361 trivial: move one line.
17363 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17365 * Ugly workaround to get s_client and s_server working under Windows. The
17366 old code wouldn't work because it needed to select() on sockets and the
17367 tty (for keypresses and to see if data could be written). Win32 only
17368 supports select() on sockets so we select() with a 1s timeout on the
17369 sockets and then see if any characters are waiting to be read, if none
17370 are present then we retry, we also assume we can always write data to
17371 the tty. This isn't nice because the code then blocks until we've
17372 received a complete line of data and it is effectively polling the
17373 keyboard at 1s intervals: however it's quite a bit better than not
17374 working at all :-) A dedicated Windows application might handle this
17375 with an event loop for example.
17379 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17380 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17381 will be called when RSA_sign() and RSA_verify() are used. This is useful
17382 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17383 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17384 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17385 This necessitated the support of an extra signature type NID_md5_sha1
17386 for SSL signatures and modifications to the SSL library to use it instead
17387 of calling RSA_public_decrypt() and RSA_private_encrypt().
17391 * Add new -verify -CAfile and -CApath options to the crl program, these
17392 will lookup a CRL issuers certificate and verify the signature in a
17393 similar way to the verify program. Tidy up the crl program so it
17394 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17395 less strict. It will now permit CRL extensions even if it is not
17396 a V2 CRL: this will allow it to tolerate some broken CRLs.
17400 * Initialize all non-automatic variables each time one of the openssl
17401 sub-programs is started (this is necessary as they may be started
17402 multiple times from the "OpenSSL>" prompt).
17404 *Lennart Bang, Bodo Moeller*
17406 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17407 removing all other RSA functionality (this is what NO_RSA does). This
17408 is so (for example) those in the US can disable those operations covered
17409 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17414 * Non-copying interface to BIO pairs.
17415 (still largely untested)
17419 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17420 ASCII string. This was handled independently in various places before.
17424 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17425 UTF8 strings a character at a time.
17429 * Use client_version from client hello to select the protocol
17430 (s23_srvr.c) and for RSA client key exchange verification
17431 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17435 * Add various utility functions to handle SPKACs, these were previously
17436 handled by poking round in the structure internals. Added new function
17437 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17438 print, verify and generate SPKACs. Based on an original idea from
17439 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17443 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17447 * Allow the config file extension section to be overwritten on the
17448 command line. Based on an original idea from Massimiliano Pala
17449 <madwolf@comune.modena.it>. The new option is called -extensions
17450 and can be applied to ca, req and x509. Also -reqexts to override
17451 the request extensions in req and -crlexts to override the crl extensions
17456 * Add new feature to the SPKAC handling in ca. Now you can include
17457 the same field multiple times by preceding it by "XXXX." for example:
17460 this is the same syntax as used in the req config file.
17464 * Allow certificate extensions to be added to certificate requests. These
17465 are specified in a 'req_extensions' option of the req section of the
17466 config file. They can be printed out with the -text option to req but
17467 are otherwise ignored at present.
17471 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17472 data read consists of only the final block it would not decrypted because
17473 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17474 A misplaced 'break' also meant the decrypted final block might not be
17475 copied until the next read.
17479 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17480 a few extra parameters to the DH structure: these will be useful if
17481 for example we want the value of 'q' or implement X9.42 DH.
17485 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17486 provides hooks that allow the default DSA functions or functions on a
17487 "per key" basis to be replaced. This allows hardware acceleration and
17488 hardware key storage to be handled without major modification to the
17489 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17490 associated functions.
17494 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17495 as "read only": it can't be written to and the buffer it points to will
17496 not be freed. Reading from a read only BIO is much more efficient than
17497 a normal memory BIO. This was added because there are several times when
17498 an area of memory needs to be read from a BIO. The previous method was
17499 to create a memory BIO and write the data to it, this results in two
17500 copies of the data and an O(n^2) reading algorithm. There is a new
17501 function BIO_new_mem_buf() which creates a read only memory BIO from
17502 an area of memory. Also modified the PKCS#7 routines to use read only
17507 * Bugfix: ssl23_get_client_hello did not work properly when called in
17508 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17509 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17510 but a retry condition occurred while trying to read the rest.
17514 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17515 NID_pkcs7_encrypted by default: this was wrong since this should almost
17516 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17517 the encrypted data type: this is a more sensible place to put it and it
17518 allows the PKCS#12 code to be tidied up that duplicated this
17523 * Changed obj_dat.pl script so it takes its input and output files on
17524 the command line. This should avoid shell escape redirection problems
17529 * Initial support for certificate extension requests, these are included
17530 in things like Xenroll certificate requests. Included functions to allow
17531 extensions to be obtained and added.
17535 * -crlf option to s_client and s_server for sending newlines as
17536 CRLF (as required by many protocols).
17540 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17542 * Install libRSAglue.a when OpenSSL is built with RSAref.
17544 *Ralf S. Engelschall*
17546 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17548 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17550 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17555 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17556 DH parameters/keys (q is lost during that conversion, but the resulting
17557 DH parameters contain its length).
17559 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17560 much faster than DH_generate_parameters (which creates parameters
17561 where `p = 2*q + 1`), and also the smaller q makes DH computations
17562 much more efficient (160-bit exponentiation instead of 1024-bit
17563 exponentiation); so this provides a convenient way to support DHE
17564 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17565 utter importance to use
17566 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17568 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17569 when such DH parameters are used, because otherwise small subgroup
17570 attacks may become possible!
17574 * Avoid memory leak in i2d_DHparams.
17578 * Allow the -k option to be used more than once in the enc program:
17579 this allows the same encrypted message to be read by multiple recipients.
17583 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17584 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17585 it will always use the numerical form of the OID, even if it has a short
17590 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17591 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17592 otherwise bn_mod_exp was called. In the case of hardware keys for example
17593 no private key components need be present and it might store extra data
17594 in the RSA structure, which cannot be accessed from bn_mod_exp.
17595 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17596 private key operations.
17600 * Added support for SPARC Linux.
17604 * pem_password_cb function type incompatibly changed from
17605 typedef int pem_password_cb(char *buf, int size, int rwflag);
17607 ....(char *buf, int size, int rwflag, void *userdata);
17608 so that applications can pass data to their callbacks:
17609 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17610 additional void * argument, which is just handed through whenever
17611 the password callback is called.
17613 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17615 New function SSL_CTX_set_default_passwd_cb_userdata.
17617 Compatibility note: As many C implementations push function arguments
17618 onto the stack in reverse order, the new library version is likely to
17619 interoperate with programs that have been compiled with the old
17620 pem_password_cb definition (PEM_whatever takes some data that
17621 happens to be on the stack as its last argument, and the callback
17622 just ignores this garbage); but there is no guarantee whatsoever that
17625 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17626 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17627 problems not only on Windows, but also on some Unix platforms.
17628 To avoid problematic command lines, these definitions are now in an
17629 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17630 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17634 * MIPS III/IV assembler module is reimplemented.
17638 * More DES library cleanups: remove references to srand/rand and
17639 delete an unused file.
17643 * Add support for the free Netwide assembler (NASM) under Win32,
17644 since not many people have MASM (ml) and it can be hard to obtain.
17645 This is currently experimental but it seems to work OK and pass all
17646 the tests. Check out INSTALL.W32 for info.
17650 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17651 without temporary keys kept an extra copy of the server key,
17652 and connections with temporary keys did not free everything in case
17657 * New function RSA_check_key and new openssl rsa option -check
17658 for verifying the consistency of RSA keys.
17660 *Ulf Moeller, Bodo Moeller*
17662 * Various changes to make Win32 compile work:
17663 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17664 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17665 comparison" warnings.
17666 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17670 * Add a debugging option to PKCS#5 v2 key generation function: when
17671 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17672 derived keys are printed to stderr.
17676 * Copy the flags in ASN1_STRING_dup().
17678 *Roman E. Pavlov <pre@mo.msk.ru>*
17680 * The x509 application mishandled signing requests containing DSA
17681 keys when the signing key was also DSA and the parameters didn't match.
17683 It was supposed to omit the parameters when they matched the signing key:
17684 the verifying software was then supposed to automatically use the CA's
17685 parameters if they were absent from the end user certificate.
17687 Omitting parameters is no longer recommended. The test was also
17688 the wrong way round! This was probably due to unusual behaviour in
17689 EVP_cmp_parameters() which returns 1 if the parameters match.
17690 This meant that parameters were omitted when they *didn't* match and
17691 the certificate was useless. Certificates signed with 'ca' didn't have
17694 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17696 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17697 The interface is as follows:
17698 Applications can use
17699 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17700 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17701 "off" is now the default.
17702 The library internally uses
17703 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17704 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17705 to disable memory-checking temporarily.
17707 Some inconsistent states that previously were possible (and were
17708 even the default) are now avoided.
17710 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17711 with each memory chunk allocated; this is occasionally more helpful
17712 than just having a counter.
17714 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17716 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17721 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17722 which largely parallels "options", but is for changing API behaviour,
17723 whereas "options" are about protocol behaviour.
17724 Initial "mode" flags are:
17726 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17727 a single record has been written.
17728 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17729 retries use the same buffer location.
17730 (But all of the contents must be
17735 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17738 * Fix problems with no-hmac etc.
17740 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17742 * New functions RSA_get_default_method(), RSA_set_method() and
17743 RSA_get_method(). These allows replacement of RSA_METHODs without having
17744 to mess around with the internals of an RSA structure.
17748 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17749 Also really enable memory leak checks in openssl.c and in some
17752 *Chad C. Mulligan, Bodo Moeller*
17754 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17755 up the length of negative integers. This has now been simplified to just
17756 store the length when it is first determined and use it later, rather
17757 than trying to keep track of where data is copied and updating it to
17759 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17761 * Add a new function PKCS7_signatureVerify. This allows the verification
17762 of a PKCS#7 signature but with the signing certificate passed to the
17763 function itself. This contrasts with PKCS7_dataVerify which assumes the
17764 certificate is present in the PKCS#7 structure. This isn't always the
17765 case: certificates can be omitted from a PKCS#7 structure and be
17766 distributed by "out of band" means (such as a certificate database).
17770 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17771 function prototypes in pem.h, also change util/mkdef.pl to add the
17772 necessary function names.
17776 * mk1mf.pl (used by Windows builds) did not properly read the
17777 options set by Configure in the top level Makefile, and Configure
17778 was not even able to write more than one option correctly.
17779 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17783 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17784 file to be loaded from a BIO or FILE pointer. The BIO version will
17785 for example allow memory BIOs to contain config info.
17789 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17790 Whoever hopes to achieve shared-library compatibility across versions
17791 must use this, not the compile-time macro.
17792 (Exercise 0.9.4: Which is the minimum library version required by
17794 Note: All this applies only to multi-threaded programs, others don't
17799 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17800 through a BIO pair triggered the default case, i.e.
17801 SSLerr(...,SSL_R_UNKNOWN_STATE).
17805 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17806 can use the SSL library even if none of the specific BIOs is
17811 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17812 for the encoded length.
17814 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17816 * Add initial documentation of the X509V3 functions.
17820 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17821 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17822 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17823 secure PKCS#8 private key format with a high iteration count.
17827 * Fix determination of Perl interpreter: A perl or perl5
17828 *directory* in $PATH was also accepted as the interpreter.
17830 *Ralf S. Engelschall*
17832 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17833 wrong with it but it was very old and did things like calling
17834 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17835 unusual formatting.
17839 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17840 to use the new extension code.
17844 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17845 with macros. This should make it easier to change their form, add extra
17846 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17851 * Add to configuration table a new entry that can specify an alternative
17852 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17853 according to Mark Crispin <MRC@Panda.COM>.
17857 * DES CBC did not update the IV. Weird.
17861 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17862 Changing the behaviour of the former might break existing programs --
17863 where IV updating is needed, des_ncbc_encrypt can be used.
17866 * When bntest is run from "make test" it drives bc to check its
17867 calculations, as well as internally checking them. If an internal check
17868 fails, it needs to cause bc to give a non-zero result or make test carries
17869 on without noticing the failure. Fixed.
17873 * DES library cleanups.
17877 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17878 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17879 ciphers. NOTE: although the key derivation function has been verified
17880 against some published test vectors it has not been extensively tested
17881 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17886 * Instead of "mkdir -p", which is not fully portable, use new
17887 Perl script "util/mkdir-p.pl".
17891 * Rewrite the way password based encryption (PBE) is handled. It used to
17892 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17893 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17894 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17895 the 'parameter' field of the AlgorithmIdentifier is passed to the
17896 underlying key generation function so it must do its own ASN1 parsing.
17897 This has also changed the EVP_PBE_CipherInit() function which now has a
17898 'parameter' argument instead of literal salt and iteration count values
17899 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17903 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17904 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17905 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17906 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17907 value was just used as a "magic string" and not used directly its
17908 value doesn't matter.
17912 * Introduce some semblance of const correctness to BN. Shame C doesn't
17917 * "linux-sparc64" configuration (ultrapenguin).
17919 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17920 "linux-sparc" configuration.
17922 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17924 * config now generates no-xxx options for missing ciphers.
17928 * Support the EBCDIC character set (work in progress).
17929 File ebcdic.c not yet included because it has a different license.
17931 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17933 * Support BS2000/OSD-POSIX.
17935 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17937 * Make callbacks for key generation use `void *` instead of `char *`.
17941 * Make S/MIME samples compile (not yet tested).
17945 * Additional typesafe stacks.
17949 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17953 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17955 * New configuration variant "sco5-gcc".
17957 * Updated some demos.
17959 *Sean O Riordain, Wade Scholine*
17961 * Add missing BIO_free at exit of pkcs12 application.
17965 * Fix memory leak in conf.c.
17969 * Updates for Win32 to assembler version of MD5.
17973 * Set #! path to perl in `apps/der_chop` to where we found it
17974 instead of using a fixed path.
17978 * SHA library changes for irix64-mips4-cc.
17982 * Improvements for VMS support.
17986 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17988 * Bignum library bug fix. IRIX 6 passes "make test" now!
17989 This also avoids the problems with SC4.2 and unpatched SC5.
17991 *Andy Polyakov <appro@fy.chalmers.se>*
17993 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17994 These are required because of the typesafe stack would otherwise break
17995 existing code. If old code used a structure member which used to be STACK
17996 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17997 sk_num or sk_value it would produce an error because the num, data members
17998 are not present in STACK_OF. Now it just produces a warning. sk_set
17999 replaces the old method of assigning a value to sk_value
18000 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18001 that does this will no longer work (and should use sk_set instead) but
18002 this could be regarded as a "questionable" behaviour anyway.
18006 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18007 correctly handle encrypted S/MIME data.
18011 * Change type of various DES function arguments from des_cblock
18012 (which means, in function argument declarations, pointer to char)
18013 to des_cblock * (meaning pointer to array with 8 char elements),
18014 which allows the compiler to do more typechecking; it was like
18015 that back in SSLeay, but with lots of ugly casts.
18017 Introduce new type const_des_cblock.
18021 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18022 problems: find RecipientInfo structure that matches recipient certificate
18023 and initialise the ASN1 structures properly based on passed cipher.
18027 * Belatedly make the BN tests actually check the results.
18031 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18032 to and from BNs: it was completely broken. New compilation option
18033 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18034 key elements as negative integers.
18038 * Reorganize and speed up MD5.
18040 *Andy Polyakov <appro@fy.chalmers.se>*
18044 *Richard Levitte <richard@levitte.org>*
18046 * New option -out to asn1parse to allow the parsed structure to be
18047 output to a file. This is most useful when combined with the -strparse
18048 option to examine the output of things like OCTET STRINGS.
18052 * Make SSL library a little more fool-proof by not requiring any longer
18053 that `SSL_set_{accept,connect}_state` be called before
18054 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18055 in many applications because usually everything *appeared* to work as
18056 intended anyway -- now it really works as intended).
18060 * Move openssl.cnf out of lib/.
18064 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18065 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18066 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18068 *Ralf S. Engelschall*
18070 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18071 handle PKCS#7 enveloped data properly.
18073 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18075 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18076 copying pointers. The cert_st handling is changed by this in
18077 various ways (and thus what used to be known as ctx->default_cert
18078 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18079 any longer when s->cert does not give us what we need).
18080 ssl_cert_instantiate becomes obsolete by this change.
18081 As soon as we've got the new code right (possibly it already is?),
18082 we have solved a couple of bugs of the earlier code where s->cert
18083 was used as if it could not have been shared with other SSL structures.
18085 Note that using the SSL API in certain dirty ways now will result
18086 in different behaviour than observed with earlier library versions:
18087 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18088 does not influence s as it used to.
18090 In order to clean up things more thoroughly, inside SSL_SESSION
18091 we don't use CERT any longer, but a new structure SESS_CERT
18092 that holds per-session data (if available); currently, this is
18093 the peer's certificate chain and, for clients, the server's certificate
18094 and temporary key. CERT holds only those values that can have
18095 meaningful defaults in an SSL_CTX.
18099 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18100 from the internal representation. Various PKCS#7 fixes: remove some
18101 evil casts and set the enc_dig_alg field properly based on the signing
18106 * Allow PKCS#12 password to be set from the command line or the
18107 environment. Let 'ca' get its config file name from the environment
18108 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18113 * Allow certificate policies extension to use an IA5STRING for the
18114 organization field. This is contrary to the PKIX definition but
18115 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18120 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18121 without disallowing inline assembler and the like for non-pedantic builds.
18125 * Support Borland C++ builder.
18127 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18133 * SHA-1 cleanups and performance enhancements.
18135 *Andy Polyakov <appro@fy.chalmers.se>*
18137 * Sparc v8plus assembler for the bignum library.
18139 *Andy Polyakov <appro@fy.chalmers.se>*
18141 * Accept any -xxx and +xxx compiler options in Configure.
18145 * Update HPUX configuration.
18149 * Add missing `sk_<type>_unshift()` function to safestack.h
18151 *Ralf S. Engelschall*
18153 * New function SSL_CTX_use_certificate_chain_file that sets the
18154 "extra_cert"s in addition to the certificate. (This makes sense
18155 only for "PEM" format files, as chains as a whole are not
18160 * Support verify_depth from the SSL API.
18161 x509_vfy.c had what can be considered an off-by-one-error:
18162 Its depth (which was not part of the external interface)
18163 was actually counting the number of certificates in a chain;
18164 now it really counts the depth.
18168 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18169 instead of X509err, which often resulted in confusing error
18170 messages since the error codes are not globally unique
18171 (e.g. an alleged error in ssl3_accept when a certificate
18172 didn't match the private key).
18174 * New function SSL_CTX_set_session_id_context that allows to set a default
18175 value (so that you don't need SSL_set_session_id_context for each
18176 connection using the SSL_CTX).
18180 * OAEP decoding bug fix.
18184 * Support INSTALL_PREFIX for package builders, as proposed by
18189 * New Configure options "threads" and "no-threads". For systems
18190 where the proper compiler options are known (currently Solaris
18191 and Linux), "threads" is the default.
18195 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18199 * Install various scripts to $(OPENSSLDIR)/misc, not to
18200 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18201 such as /usr/local/bin.
18205 * "make linux-shared" to build shared libraries.
18207 *Niels Poppe <niels@netbox.org>*
18209 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18213 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18214 extension adding in x509 utility.
18218 * Remove NOPROTO sections and error code comments.
18222 * Partial rewrite of the DEF file generator to now parse the ANSI
18227 * New Configure options --prefix=DIR and --openssldir=DIR.
18231 * Complete rewrite of the error code script(s). It is all now handled
18232 by one script at the top level which handles error code gathering,
18233 header rewriting and C source file generation. It should be much better
18234 than the old method: it now uses a modified version of Ulf's parser to
18235 read the ANSI prototypes in all header files (thus the old K&R definitions
18236 aren't needed for error creation any more) and do a better job of
18237 translating function codes into names. The old 'ASN1 error code embedded
18238 in a comment' is no longer necessary and it doesn't use .err files which
18239 have now been deleted. Also the error code call doesn't have to appear all
18240 on one line (which resulted in some large lines...).
18244 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18248 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18249 0 (which usually indicates a closed connection), but continue reading.
18253 * Fix some race conditions.
18257 * Add support for CRL distribution points extension. Add Certificate
18258 Policies and CRL distribution points documentation.
18262 * Move the autogenerated header file parts to crypto/opensslconf.h.
18266 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18267 8 of keying material. Merlin has also confirmed interop with this fix
18268 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18270 *Merlin Hughes <merlin@baltimore.ie>*
18272 * Fix lots of warnings.
18274 *Richard Levitte <levitte@stacken.kth.se>*
18276 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18277 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18279 *Richard Levitte <levitte@stacken.kth.se>*
18281 * Fix problems with sizeof(long) == 8.
18283 *Andy Polyakov <appro@fy.chalmers.se>*
18285 * Change functions to ANSI C.
18289 * Fix typos in error codes.
18291 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18293 * Remove defunct assembler files from Configure.
18297 * SPARC v8 assembler BIGNUM implementation.
18299 *Andy Polyakov <appro@fy.chalmers.se>*
18301 * Support for Certificate Policies extension: both print and set.
18302 Various additions to support the r2i method this uses.
18306 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18307 return a const string when you are expecting an allocated buffer.
18311 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18312 types DirectoryString and DisplayText.
18316 * Add code to allow r2i extensions to access the configuration database,
18317 add an LHASH database driver and add several ctx helper functions.
18321 * Fix an evil bug in bn_expand2() which caused various BN functions to
18322 fail when they extended the size of a BIGNUM.
18326 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18327 support typesafe stack.
18331 * Fix typo in SSL_[gs]et_options().
18333 *Nils Frostberg <nils@medcom.se>*
18335 * Delete various functions and files that belonged to the (now obsolete)
18336 old X509V3 handling code.
18340 * New Configure option "rsaref".
18344 * Don't auto-generate pem.h.
18348 * Introduce type-safe ASN.1 SETs.
18352 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18354 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18356 * Introduce type-safe STACKs. This will almost certainly break lots of code
18357 that links with OpenSSL (well at least cause lots of warnings), but fear
18358 not: the conversion is trivial, and it eliminates loads of evil casts. A
18359 few STACKed things have been converted already. Feel free to convert more.
18360 In the fullness of time, I'll do away with the STACK type altogether.
18364 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18365 specified in `<certfile>` by updating the entry in the index.txt file.
18366 This way one no longer has to edit the index.txt file manually for
18367 revoking a certificate. The -revoke option does the gory details now.
18369 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18371 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18372 `-text` option at all and this way the `-noout -text` combination was
18373 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18375 *Ralf S. Engelschall*
18377 * Make sure a corresponding plain text error message exists for the
18378 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18379 verify callback function determined that a certificate was revoked.
18381 *Ralf S. Engelschall*
18383 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18384 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18385 all available ciphers including rc5, which was forgotten until now.
18386 In order to let the testing shell script know which algorithms
18387 are available, a new (up to now undocumented) command
18388 `openssl list-cipher-commands` is used.
18392 * Bugfix: s_client occasionally would sleep in select() when
18393 it should have checked SSL_pending() first.
18397 * New functions DSA_do_sign and DSA_do_verify to provide access to
18398 the raw DSA values prior to ASN.1 encoding.
18402 * Tweaks to Configure
18404 *Niels Poppe <niels@netbox.org>*
18406 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18411 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18415 * New config option to avoid instructions that are illegal on the 80386.
18416 The default code is faster, but requires at least a 486.
18420 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18421 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18422 same as SSL2_VERSION anyway.
18426 * New "-showcerts" option for s_client.
18430 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18431 application. Various cleanups and fixes.
18435 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18436 modify error routines to work internally. Add error codes and PBE init
18437 to library startup routines.
18441 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18442 packing functions to asn1 and evp. Changed function names and error
18443 codes along the way.
18447 * PKCS12 integration: and so it begins... First of several patches to
18448 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18449 objects to objects.h
18453 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18454 and display support for Thawte strong extranet extension.
18458 * Add LinuxPPC support.
18460 *Jeff Dubrule <igor@pobox.org>*
18462 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18463 bn_div_words in alpha.s.
18465 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18467 * Make sure the RSA OAEP test is skipped under -DRSAref because
18468 OAEP isn't supported when OpenSSL is built with RSAref.
18470 *Ulf Moeller <ulf@fitug.de>*
18472 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18473 so they no longer are missing under -DNOPROTO.
18475 *Soren S. Jorvang <soren@t.dk>*
18477 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18479 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18480 doesn't work when the session is reused. Coming soon!
18484 * Fix a security hole, that allows sessions to be reused in the wrong
18485 context thus bypassing client cert protection! All software that uses
18486 client certs and session caches in multiple contexts NEEDS PATCHING to
18487 allow session reuse! A fuller solution is in the works.
18489 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18491 * Some more source tree cleanups (removed obsolete files
18492 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18493 permission on "config" script to be executable) and a fix for the INSTALL
18496 *Ulf Moeller <ulf@fitug.de>*
18498 * Remove some legacy and erroneous uses of malloc, free instead of
18501 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18503 * Make rsa_oaep_test return non-zero on error.
18505 *Ulf Moeller <ulf@fitug.de>*
18507 * Add support for native Solaris shared libraries. Configure
18508 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18509 if someone would make that last step automatic.
18511 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18513 * ctx_size was not built with the right compiler during "make links". Fixed.
18517 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18518 except NULL ciphers". This means the default cipher list will no longer
18519 enable NULL ciphers. They need to be specifically enabled e.g. with
18520 the string "DEFAULT:eNULL".
18524 * Fix to RSA private encryption routines: if p < q then it would
18525 occasionally produce an invalid result. This will only happen with
18526 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18530 * Be less restrictive and allow also `perl util/perlpath.pl
18531 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18532 because this way one can also use an interpreter named `perl5` (which is
18533 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18534 installed as `perl`).
18536 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18538 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18540 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18542 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18543 advapi32.lib to Win32 build and change the pem test comparison
18544 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18545 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18546 and crypto/des/ede_cbcm_enc.c.
18550 * DES quad checksum was broken on big-endian architectures. Fixed.
18554 * Comment out two functions in bio.h that aren't implemented. Fix up the
18555 Win32 test batch file so it (might) work again. The Win32 test batch file
18556 is horrible: I feel ill....
18560 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18561 in e_os.h. Audit of header files to check ANSI and non ANSI
18562 sections: 10 functions were absent from non ANSI section and not exported
18563 from Windows DLLs. Fixed up libeay.num for new functions.
18567 * Make `openssl version` output lines consistent.
18569 *Ralf S. Engelschall*
18571 * Fix Win32 symbol export lists for BIO functions: Added
18572 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18573 to ms/libeay{16,32}.def.
18575 *Ralf S. Engelschall*
18577 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18578 fine under Unix and passes some trivial tests I've now added. But the
18579 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18580 added to make sure no one expects that this stuff really works in the
18581 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18582 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18585 *Ralf S. Engelschall*
18587 * Fix the generation of two part addresses in perl.
18589 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18591 * Add config entry for Linux on MIPS.
18593 *John Tobey <jtobey@channel1.com>*
18595 * Make links whenever Configure is run, unless we are on Windoze.
18599 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18600 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18605 * Add a useful kludge to allow package maintainers to specify compiler and
18606 other platforms details on the command line without having to patch the
18607 Configure script every time: One now can use
18608 `perl Configure <id>:<details>`,
18609 i.e. platform ids are allowed to have details appended
18610 to them (separated by colons). This is treated as there would be a static
18611 pre-configured entry in Configure's %table under key `<id>` with value
18612 `<details>` and `perl Configure <id>` is called. So, when you want to
18613 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18614 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18615 now, which overrides the FreeBSD-elf entry on-the-fly.
18617 *Ralf S. Engelschall*
18619 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18623 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18624 on the `perl Configure ...` command line. This way one can compile
18625 OpenSSL libraries with Position Independent Code (PIC) which is needed
18626 for linking it into DSOs.
18628 *Ralf S. Engelschall*
18630 * Remarkably, export ciphers were totally broken and no-one had noticed!
18635 * Cleaned up the LICENSE document: The official contact for any license
18636 questions now is the OpenSSL core team under openssl-core@openssl.org.
18637 And add a paragraph about the dual-license situation to make sure people
18638 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18639 to the OpenSSL toolkit.
18641 *Ralf S. Engelschall*
18643 * General source tree makefile cleanups: Made `making xxx in yyy...`
18644 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18645 Additionally cleaned up the `make links` target: Remove unnecessary
18646 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18647 to speed processing and no longer clutter the display with confusing
18648 stuff. Instead only the actually done links are displayed.
18650 *Ralf S. Engelschall*
18652 * Permit null encryption ciphersuites, used for authentication only. It used
18653 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18654 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18659 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18660 signed attributes when verifying signatures (this would break them),
18661 the detached data encoding was wrong and public keys obtained using
18662 X509_get_pubkey() weren't freed.
18666 * Add text documentation for the BUFFER functions. Also added a work around
18667 to a Win95 console bug. This was triggered by the password read stuff: the
18668 last character typed gets carried over to the next fread(). If you were
18669 generating a new cert request using 'req' for example then the last
18670 character of the passphrase would be CR which would then enter the first
18675 * Added the new 'Includes OpenSSL Cryptography Software' button as
18676 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18677 button and can be used by applications based on OpenSSL to show the
18678 relationship to the OpenSSL project.
18680 *Ralf S. Engelschall*
18682 * Remove confusing variables in function signatures in files
18683 ssl/ssl_lib.c and ssl/ssl.h.
18685 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18687 * Don't install bss_file.c under PREFIX/include/
18689 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18691 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18692 functions that return function pointers and has support for NT specific
18693 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18694 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18695 unsigned to signed types: this was killing the Win32 compile.
18699 * Add new certificate file to stack functions,
18700 SSL_add_dir_cert_subjects_to_stack() and
18701 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18702 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18703 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18704 This means that Apache-SSL and similar packages don't have to mess around
18705 to add as many CAs as they want to the preferred list.
18709 * Experiment with doxygen documentation. Currently only partially applied to
18711 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18712 openssl.doxy as the configuration file.
18716 * Get rid of remaining C++-style comments which strict C compilers hate.
18718 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18720 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18721 compiled in by default: it has problems with large keys.
18725 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18726 DH private keys and/or callback functions which directly correspond to
18727 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18728 is needed for applications which have to configure certificates on a
18729 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18731 For the RSA certificate situation is makes no difference, but
18732 for the DSA certificate situation this fixes the "no shared cipher"
18733 problem where the OpenSSL cipher selection procedure failed because the
18734 temporary keys were not overtaken from the context and the API provided
18735 no way to reconfigure them.
18736 The new functions now let applications reconfigure the stuff and they
18737 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18738 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18739 non-public-API function ssl_cert_instantiate() is used as a helper
18740 function and also to reduce code redundancy inside ssl_rsa.c.
18742 *Ralf S. Engelschall*
18744 * Move s_server -dcert and -dkey options out of the undocumented feature
18745 area because they are useful for the DSA situation and should be
18746 recognized by the users.
18748 *Ralf S. Engelschall*
18750 * Fix the cipher decision scheme for export ciphers: the export bits are
18751 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18752 SSL_EXP_MASK. So, the original variable has to be used instead of the
18753 already masked variable.
18755 *Richard Levitte <levitte@stacken.kth.se>*
18757 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18759 *Richard Levitte <levitte@stacken.kth.se>*
18761 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18762 from `int` to `unsigned int` because it is a length and initialized by
18763 EVP_DigestFinal() which expects an `unsigned int *`.
18765 *Richard Levitte <levitte@stacken.kth.se>*
18767 * Don't hard-code path to Perl interpreter on shebang line of Configure
18768 script. Instead use the usual Shell->Perl transition trick.
18770 *Ralf S. Engelschall*
18772 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18773 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18774 -noout -modulus` as it's already the case for `openssl rsa -noout
18775 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18776 currently the public key is printed (a decision which was already done by
18777 `openssl dsa -modulus` in the past) which serves a similar purpose.
18778 Additionally the NO_RSA no longer completely removes the whole -modulus
18779 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18782 *Ralf S. Engelschall*
18784 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18785 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18787 *Arne Ansper <arne@ats.cyber.ee>*
18789 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18790 to be added. Now both 'req' and 'ca' can use new objects defined in the
18795 * Add cool BIO that does syslog (or event log on NT).
18797 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18799 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18800 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18801 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18802 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18806 * Add preliminary config info for new extension code.
18810 * Make RSA_NO_PADDING really use no padding.
18812 *Ulf Moeller <ulf@fitug.de>*
18814 * Generate errors when private/public key check is done.
18818 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18819 for some CRL extensions and new objects added.
18823 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18824 key usage extension and fuller support for authority key id.
18828 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18829 padding method for RSA, which is recommended for new applications in PKCS
18830 #1 v2.0 (RFC 2437, October 1998).
18831 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18832 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18833 against Bleichbacher's attack on RSA.
18834 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18837 * Updates to the new SSL compression code
18839 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18841 * Fix so that the version number in the master secret, when passed
18842 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18843 (because the server will not accept higher), that the version number
18844 is 0x03,0x01, not 0x03,0x00
18846 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18848 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18849 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18850 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18854 * Support for RAW extensions where an arbitrary extension can be
18855 created by including its DER encoding. See `apps/openssl.cnf` for
18860 * Make sure latest Perl versions don't interpret some generated C array
18861 code as Perl array code in the crypto/err/err_genc.pl script.
18863 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18865 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18866 not many people have the assembler. Various Win32 compilation fixes and
18867 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18868 build instructions.
18872 * Modify configure script 'Configure' to automatically create crypto/date.h
18873 file under Win32 and also build pem.h from pem.org. New script
18874 util/mkfiles.pl to create the MINFO file on environments that can't do a
18875 'make files': perl util/mkfiles.pl >MINFO should work.
18879 * Major rework of DES function declarations, in the pursuit of correctness
18880 and purity. As a result, many evil casts evaporated, and some weirdness,
18881 too. You may find this causes warnings in your code. Zapping your evil
18882 casts will probably fix them. Mostly.
18886 * Fix for a typo in asn1.h. Bug fix to object creation script
18887 obj_dat.pl. It considered a zero in an object definition to mean
18888 "end of object": none of the objects in objects.h have any zeros
18889 so it wasn't spotted.
18891 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18893 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18894 Masking (CBCM). In the absence of test vectors, the best I have been able
18895 to do is check that the decrypt undoes the encrypt, so far. Send me test
18896 vectors if you have them.
18900 * Correct calculation of key length for export ciphers (too much space was
18901 allocated for null ciphers). This has not been tested!
18905 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18906 message is now correct (it understands "crypto" and "ssl" on its
18907 command line). There is also now an "update" option. This will update
18908 the util/ssleay.num and util/libeay.num files with any new functions.
18910 perl util/mkdef.pl crypto ssl update
18911 it will update them.
18915 * Overhauled the Perl interface:
18916 - ported BN stuff to OpenSSL's different BN library
18917 - made the perl/ source tree CVS-aware
18918 - renamed the package from SSLeay to OpenSSL (the files still contain
18919 their history because I've copied them in the repository)
18920 - removed obsolete files (the test scripts will be replaced
18921 by better Test::Harness variants in the future)
18923 *Ralf S. Engelschall*
18925 * First cut for a very conservative source tree cleanup:
18926 1. merge various obsolete readme texts into doc/ssleay.txt
18927 where we collect the old documents and readme texts.
18928 2. remove the first part of files where I'm already sure that we no
18929 longer need them because of three reasons: either they are just temporary
18930 files which were left by Eric or they are preserved original files where
18931 I've verified that the diff is also available in the CVS via "cvs diff
18932 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18933 the crypto/md/ stuff).
18935 *Ralf S. Engelschall*
18937 * More extension code. Incomplete support for subject and issuer alt
18938 name, issuer and authority key id. Change the i2v function parameters
18939 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18940 what that's for :-) Fix to ASN1 macro which messed up
18941 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18945 * Preliminary support for ENUMERATED type. This is largely copied from the
18950 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18952 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18954 * Make sure `make rehash` target really finds the `openssl` program.
18956 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18958 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18959 like to hear about it if this slows down other processors.
18963 * Add CygWin32 platform information to Configure script.
18965 *Alan Batie <batie@aahz.jf.intel.com>*
18967 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18969 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18971 * New program nseq to manipulate netscape certificate sequences
18975 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18980 * Fixes to BN code. Previously the default was to define BN_RECURSION
18981 but the BN code had some problems that would cause failures when
18982 doing certificate verification and some other functions.
18984 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18986 * Add ASN1 and PEM code to support netscape certificate sequences.
18990 * Add ASN1 and PEM code to support netscape certificate sequences.
18994 * Add several PKIX and private extended key usage OIDs.
18998 * Modify the 'ca' program to handle the new extension code. Modify
18999 openssl.cnf for new extension format, add comments.
19003 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19004 and add a sample to openssl.cnf so req -x509 now adds appropriate
19009 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19010 error code, add initial support to X509_print() and x509 application.
19014 * Takes a deep breath and start adding X509 V3 extension support code. Add
19015 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19016 stuff is currently isolated and isn't even compiled yet.
19020 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19021 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19022 Removed the versions check from X509 routines when loading extensions:
19023 this allows certain broken certificates that don't set the version
19024 properly to be processed.
19028 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19029 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19030 can still be regenerated with "make depend".
19034 * Spelling mistake in C version of CAST-128.
19036 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19038 * Changes to the error generation code. The perl script err-code.pl
19039 now reads in the old error codes and retains the old numbers, only
19040 adding new ones if necessary. It also only changes the .err files if new
19041 codes are added. The makefiles have been modified to only insert errors
19042 when needed (to avoid needlessly modifying header files). This is done
19043 by only inserting errors if the .err file is newer than the auto generated
19044 C file. To rebuild all the error codes from scratch (the old behaviour)
19045 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19046 or delete all the .err files.
19050 * CAST-128 was incorrectly implemented for short keys. The C version has
19051 been fixed, but is untested. The assembler versions are also fixed, but
19052 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19053 to regenerate it if needed.
19054 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19055 Hagino <itojun@kame.net>*
19057 * File was opened incorrectly in randfile.c.
19059 *Ulf Möller <ulf@fitug.de>*
19061 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19062 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19063 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19064 al: it's just almost always a UTCTime. Note this patch adds new error
19065 codes so do a "make errors" if there are problems.
19069 * Correct Linux 1 recognition in config.
19071 *Ulf Möller <ulf@fitug.de>*
19073 * Remove pointless MD5 hash when using DSA keys in ca.
19075 *Anonymous <nobody@replay.com>*
19077 * Generate an error if given an empty string as a cert directory. Also
19078 generate an error if handed NULL (previously returned 0 to indicate an
19079 error, but didn't set one).
19081 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19083 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19087 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19088 parameters. This was causing a warning which killed off the Win32 compile.
19092 * Remove C++ style comments from crypto/bn/bn_local.h.
19094 *Neil Costigan <neil.costigan@celocom.com>*
19096 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19097 based on a text string, looking up short and long names and finally
19098 "dot" format. The "dot" format stuff didn't work. Added new function
19099 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19100 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19101 OID is not part of the table.
19105 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19106 X509_LOOKUP_by_alias().
19110 * Sort openssl functions by name.
19114 * Get the `gendsa` command working and add it to the `list` command. Remove
19115 encryption from sample DSA keys (in case anyone is interested the password
19120 * Make *all* `*_free` functions accept a NULL pointer.
19122 *Frans Heymans <fheymans@isaserver.be>*
19124 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19127 *Anonymous <nobody@replay.com>*
19129 * s_server should send the CAfile as acceptable CAs, not its own cert.
19131 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19133 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19135 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19137 * Temp key "for export" tests were wrong in s3_srvr.c.
19139 *Anonymous <nobody@replay.com>*
19141 * Add prototype for temp key callback functions
19142 SSL_CTX_set_tmp_{rsa,dh}_callback().
19146 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19147 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19151 * X509_name_add_entry() freed the wrong thing after an error.
19153 *Arne Ansper <arne@ats.cyber.ee>*
19155 * rsa_eay.c would attempt to free a NULL context.
19157 *Arne Ansper <arne@ats.cyber.ee>*
19159 * BIO_s_socket() had a broken should_retry() on Windoze.
19161 *Arne Ansper <arne@ats.cyber.ee>*
19163 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19165 *Arne Ansper <arne@ats.cyber.ee>*
19167 * Make sure the already existing X509_STORE->depth variable is initialized
19168 in X509_STORE_new(), but document the fact that this variable is still
19169 unused in the certificate verification process.
19171 *Ralf S. Engelschall*
19173 * Fix the various library and `apps/` files to free up pkeys obtained from
19174 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19178 * Fix reference counting in X509_PUBKEY_get(). This makes
19179 demos/maurice/example2.c work, amongst others, probably.
19181 *Steve Henson and Ben Laurie*
19183 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19184 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19185 are no longer created. This way we have a single and consistent command
19186 line interface `openssl <command>`, similar to `cvs <command>`.
19188 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19190 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19191 BIT STRING wrapper always have zero unused bits.
19195 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19199 * Make the top-level INSTALL documentation easier to understand.
19203 * Makefiles updated to exit if an error occurs in a sub-directory
19204 make (including if user presses ^C) [Paul Sutton]
19206 * Make Montgomery context stuff explicit in RSA data structure.
19210 * Fix build order of pem and err to allow for generated pem.h.
19214 * Fix renumbering bug in X509_NAME_delete_entry().
19218 * Enhanced the err-ins.pl script so it makes the error library number
19219 global and can add a library name. This is needed for external ASN1 and
19220 other error libraries.
19224 * Fixed sk_insert which never worked properly.
19228 * Fix ASN1 macros so they can handle indefinite length constructed
19229 EXPLICIT tags. Some non standard certificates use these: they can now
19234 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19235 into a single doc/ssleay.txt bundle. This way the information is still
19236 preserved but no longer messes up this directory. Now it's new room for
19237 the new set of documentation files.
19239 *Ralf S. Engelschall*
19241 * SETs were incorrectly DER encoded. This was a major pain, because they
19242 shared code with SEQUENCEs, which aren't coded the same. This means that
19243 almost everything to do with SETs or SEQUENCEs has either changed name or
19244 number of arguments.
19246 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19248 * Fix test data to work with the above.
19252 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19253 was already fixed by Eric for 0.9.1 it seems.
19255 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19257 * Autodetect FreeBSD3.
19261 * Fix various bugs in Configure. This affects the following platforms:
19265 unixware-2.0-pentium
19270 * Eliminate generated files from CVS. Reorder tests to regenerate files
19271 before they are needed.
19275 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19279 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19281 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19282 changed SSLeay to OpenSSL in version strings.
19284 *Ralf S. Engelschall*
19286 * Some fixups to the top-level documents.
19290 * Fixed the nasty bug where rsaref.h was not found under compile-time
19291 because the symlink to include/ was missing.
19293 *Ralf S. Engelschall*
19295 * Incorporated the popular no-RSA/DSA-only patches
19296 which allow to compile a RSA-free SSLeay.
19298 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19300 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19301 when "ssleay" is still not found.
19303 *Ralf S. Engelschall*
19305 * Added more platforms to Configure: Cray T3E, HPUX 11,
19307 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19309 * Updated the README file.
19311 *Ralf S. Engelschall*
19313 * Added various .cvsignore files in the CVS repository subdirs
19314 to make a "cvs update" really silent.
19316 *Ralf S. Engelschall*
19318 * Recompiled the error-definition header files and added
19319 missing symbols to the Win32 linker tables.
19321 *Ralf S. Engelschall*
19323 * Cleaned up the top-level documents;
19324 o new files: CHANGES and LICENSE
19325 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19326 o merged COPYRIGHT into LICENSE
19327 o removed obsolete TODO file
19328 o renamed MICROSOFT to INSTALL.W32
19330 *Ralf S. Engelschall*
19332 * Removed dummy files from the 0.9.1b source tree:
19333 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19334 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19335 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19336 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19337 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19339 *Ralf S. Engelschall*
19341 * Added various platform portability fixes.
19345 * The Genesis of the OpenSSL rpject:
19346 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19347 Young and Tim J. Hudson created while they were working for C2Net until
19350 *The OpenSSL Project*
19352 ### Changes between 0.9.0b and 0.9.1b [not released]
19354 * Updated a few CA certificates under certs/
19358 * Changed some BIGNUM api stuff.
19362 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19363 DGUX x86, Linux Alpha, etc.
19367 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19368 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19373 * Add -strparse option to asn1pars program which parses nested
19376 *Dr Stephen Henson <shenson@bigfoot.com>*
19378 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19382 * DSA fix for "ca" program.
19386 * Added "-genkey" option to "dsaparam" program.
19390 * Added RIPE MD160 (rmd160) message digest.
19394 * Added -a (all) option to "ssleay version" command.
19398 * Added PLATFORM define which is the id given to Configure.
19402 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19406 * Extended the ASN.1 parser routines.
19410 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19414 * Added a BN_CTX to the BN library.
19418 * Fixed the weak key values in DES library
19422 * Changed API in EVP library for cipher aliases.
19426 * Added support for RC2/64bit cipher.
19430 * Converted the lhash library to the crypto/mem.c functions.
19434 * Added more recognized ASN.1 object ids.
19438 * Added more RSA padding checks for SSL/TLS.
19442 * Added BIO proxy/filter functionality.
19446 * Added extra_certs to SSL_CTX which can be used
19447 send extra CA certificates to the client in the CA cert chain sending
19448 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19452 * Now Fortezza is denied in the authentication phase because
19453 this is key exchange mechanism is not supported by SSLeay at all.
19457 * Additional PKCS1 checks.
19461 * Support the string "TLSv1" for all TLS v1 ciphers.
19465 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19466 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19470 * Fixed a few memory leaks.
19474 * Fixed various code and comment typos.
19478 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19479 bytes sent in the client random.
19481 *Edward Bishop <ebishop@spyglass.com>*
19485 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19486 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19487 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19488 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19489 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19490 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19491 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19492 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19493 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19494 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19495 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19496 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19497 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19498 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19499 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19500 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19501 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19502 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19503 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19504 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19505 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19506 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19507 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19508 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19509 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19510 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19511 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19512 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19513 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19514 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19515 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19516 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19517 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19518 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19519 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19520 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19521 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19522 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19523 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19524 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19525 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19526 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19527 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19528 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19529 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19530 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19531 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19532 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19533 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19534 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19535 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19536 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19537 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19538 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19539 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19540 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19541 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19542 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19543 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19544 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19545 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19546 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19547 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19548 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19549 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19550 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19551 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19552 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19553 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19554 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19555 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19556 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19557 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19558 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19559 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19560 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19561 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19562 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19563 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19564 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19565 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19566 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19567 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19568 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19569 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19570 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19571 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19572 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19573 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19574 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19575 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19576 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19577 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19578 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19579 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19580 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19581 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19582 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19583 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19584 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19585 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19586 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19587 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19588 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19589 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19590 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19591 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19592 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19593 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19594 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19595 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19596 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19597 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19598 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19599 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19600 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19601 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19602 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19603 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19604 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19605 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19606 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19607 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19608 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19609 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19610 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19611 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19612 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19613 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19614 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19615 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19616 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19617 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19618 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19619 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19620 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19621 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19622 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19623 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19624 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19625 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19626 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19627 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19628 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19629 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19630 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19631 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19632 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19633 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19634 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19635 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19636 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19637 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19638 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19639 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19640 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19641 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19642 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19643 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19644 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19645 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19646 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19647 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19648 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655