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 3.1.0 [xx XXX xxxx]
27 * Added support for KMAC in KBKDF.
31 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
32 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
33 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
34 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
35 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
36 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
37 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
38 for legacy EC and SM2 keys is also changed similarly to honor the
39 equivalent conversion format flag as specified in the underlying
40 `EC_KEY` object being exported to a provider, when this function is
41 called through `EVP_PKEY_export()`.
45 * RNDR and RNDRRS support in provider functions to provide
46 random number generation for Arm CPUs (aarch64).
50 * s_client and s_server apps now explicitly say when the TLS version
51 does not include the renegotiation mechanism. This avoids confusion
52 between that scenario versus when the TLS version includes secure
53 renegotiation but the peer lacks support for it.
57 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
59 *Tomasz Kantecki, Andrey Matyukov*
61 * The various OBJ_* functions have been made thread safe.
65 * Parallel dual-prime 1536/2048-bit modular exponentiation for
66 AVX512_IFMA capable processors.
68 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
70 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
71 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
72 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
73 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
74 `OPENSSL_NO_DEPRECATED_3_1`.
76 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
77 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
78 definitions for these functions regardless of whether
79 `OPENSSL_NO_DEPRECATED_3_1` is defined.
81 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
82 functions regardless of whether they are using them. It is recommended that
83 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
87 * When generating safe-prime DH parameters set the recommended private key
88 length equivalent to minimum key lengths as in RFC 7919.
92 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
93 maximum size that is smaller or equal to the digest length to comply with
94 FIPS 186-4 section 5. This is implemented by a new option
95 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
96 `rsa_pss_saltlen` parameter, which is now the default. Signature
97 verification is not affected by this change and continues to work as before.
104 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
105 listed here are only a brief description.
106 The migration guide contains more detailed information related to new features,
107 breaking changes, and mappings for the large list of deprecated functions.
109 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
111 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
113 * Fixed two buffer overflows in punycode decoding functions.
115 A buffer overrun can be triggered in X.509 certificate verification,
116 specifically in name constraint checking. Note that this occurs after
117 certificate chain signature verification and requires either a CA to
118 have signed the malicious certificate or for the application to continue
119 certificate verification despite failure to construct a path to a trusted
122 In a TLS client, this can be triggered by connecting to a malicious
123 server. In a TLS server, this can be triggered if the server requests
124 client authentication and a malicious client connects.
126 An attacker can craft a malicious email address to overflow
127 an arbitrary number of bytes containing the `.` character (decimal 46)
128 on the stack. This buffer overflow could result in a crash (causing a
132 An attacker can craft a malicious email address to overflow four
133 attacker-controlled bytes on the stack. This buffer overflow could
134 result in a crash (causing a denial of service) or potentially remote code
135 execution depending on stack layout for any given platform/compiler.
140 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
141 parameters in OpenSSL code.
142 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
143 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
144 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
145 Using these invalid names may cause algorithms to use slower methods
146 that ignore the CRT parameters.
150 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
155 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
156 data to be signed before signing the certificate.
160 * Added RIPEMD160 to the default provider.
164 * Ensured that the key share group sent or accepted for the key exchange
165 is allowed for the protocol version.
169 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
171 * OpenSSL supports creating a custom cipher via the legacy
172 EVP_CIPHER_meth_new() function and associated function calls. This function
173 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
174 to use the new provider mechanism in order to implement custom ciphers.
176 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
177 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
178 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
179 and decryption initialisation functions). Instead of using the custom cipher
180 directly it incorrectly tries to fetch an equivalent cipher from the
181 available providers. An equivalent cipher is found based on the NID passed to
182 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
183 given cipher. However it is possible for an application to incorrectly pass
184 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
185 is used in this way the OpenSSL encryption/decryption initialisation function
186 will match the NULL cipher as being equivalent and will fetch this from the
187 available providers. This will succeed if the default provider has been
188 loaded (or if a third party provider has been loaded that offers this
189 cipher). Using the NULL cipher means that the plaintext is emitted as the
192 Applications are only affected by this issue if they call
193 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
194 encryption/decryption initialisation function. Applications that only use
195 SSL/TLS are not impacted by this issue.
200 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
205 * Fixed the linux-mips64 Configure target which was missing the
206 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
211 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
216 * Correctly handle a retransmitted ClientHello in DTLS
220 * Fixed detection of ktls support in cross-compile environment on Linux
224 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
229 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
230 report correct results in some cases
234 * Fix UWP builds by defining VirtualLock
238 * For known safe primes use the minimum key length according to RFC 7919.
239 Longer private key sizes unnecessarily raise the cycles needed to compute the
240 shared secret without any increase of the real security. This fixes a
241 regression from 1.1.1 where these shorter keys were generated for the known
246 * Added the loongarch64 target
250 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
251 only passed to the FIPS provider and not to the default or legacy provider.
255 * Fixed reported performance degradation on aarch64. Restored the
256 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
257 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
258 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
259 The new algorithm is still used for 32 bit targets.
263 * Added a missing header for memcmp that caused compilation failure on some
268 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
270 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
271 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
272 This issue makes the RSA implementation with 2048 bit private keys
273 incorrect on such machines and memory corruption will happen during
274 the computation. As a consequence of the memory corruption an attacker
275 may be able to trigger a remote code execution on the machine performing
278 SSL/TLS servers or other servers using 2048 bit RSA private keys running
279 on machines supporting AVX512IFMA instructions of the X86_64 architecture
280 are affected by this issue.
285 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
286 implementation would not encrypt the entirety of the data under some
287 circumstances. This could reveal sixteen bytes of data that was
288 preexisting in the memory that wasn't written. In the special case of
289 "in place" encryption, sixteen bytes of the plaintext would be revealed.
291 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
292 they are both unaffected.
295 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
297 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
299 * In addition to the c_rehash shell command injection identified in
300 CVE-2022-1292, further bugs where the c_rehash script does not
301 properly sanitise shell metacharacters to prevent command injection have been
304 When the CVE-2022-1292 was fixed it was not discovered that there
305 are other places in the script where the file names of certificates
306 being hashed were possibly passed to a command executed through the shell.
308 This script is distributed by some operating systems in a manner where
309 it is automatically executed. On such operating systems, an attacker
310 could execute arbitrary commands with the privileges of the script.
312 Use of the c_rehash script is considered obsolete and should be replaced
313 by the OpenSSL rehash command line tool.
316 *Daniel Fiala, Tomáš Mráz*
318 * Case insensitive string comparison no longer uses locales. It has instead
319 been directly implemented.
323 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
325 * Case insensitive string comparison is reimplemented via new locale-agnostic
326 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
327 comparison. The previous implementation had problems when the Turkish locale
332 * Fixed a bug in the c_rehash script which was not properly sanitising shell
333 metacharacters to prevent command injection. This script is distributed by
334 some operating systems in a manner where it is automatically executed. On
335 such operating systems, an attacker could execute arbitrary commands with the
336 privileges of the script.
338 Use of the c_rehash script is considered obsolete and should be replaced
339 by the OpenSSL rehash command line tool.
344 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
345 certificate on an OCSP response. The bug caused the function in the case
346 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
347 response (meaning a successful verification) even in the case where the
348 response signing certificate fails to verify.
350 It is anticipated that most users of `OCSP_basic_verify` will not use the
351 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
352 a negative value (indicating a fatal error) in the case of a certificate
353 verification failure. The normal expected return value in this case would be
356 This issue also impacts the command line OpenSSL "ocsp" application. When
357 verifying an ocsp response with the "-no_cert_checks" option the command line
358 application will report that the verification is successful even though it
359 has in fact failed. In this case the incorrect successful response will also
360 be accompanied by error messages showing the failure and contradicting the
361 apparently successful result.
366 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
367 AAD data as the MAC key. This made the MAC key trivially predictable.
369 An attacker could exploit this issue by performing a man-in-the-middle attack
370 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
371 that the modified data would still pass the MAC integrity check.
373 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
374 endpoint will always be rejected by the recipient and the connection will
375 fail at that point. Many application protocols require data to be sent from
376 the client to the server first. Therefore, in such a case, only an OpenSSL
377 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
379 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
380 sent in both directions. In this case both clients and servers could be
381 affected, regardless of the application protocol.
383 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
384 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
385 the handshake when using this ciphersuite.
387 The confidentiality of data is not impacted by this issue, i.e. an attacker
388 cannot decrypt data that has been encrypted using this ciphersuite - they can
391 In order for this attack to work both endpoints must legitimately negotiate
392 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
393 OpenSSL 3.0, and is not available within the default provider or the default
394 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
395 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
396 following must have occurred:
398 1) OpenSSL must have been compiled with the (non-default) compile time option
399 enable-weak-ssl-ciphers
401 2) OpenSSL must have had the legacy provider explicitly loaded (either
402 through application code or via configuration)
404 3) The ciphersuite must have been explicitly added to the ciphersuite list
406 4) The libssl security level must have been set to 0 (default is 1)
408 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
410 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
411 others that both endpoints have in common
416 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
417 occupied by the removed hash table entries.
419 This function is used when decoding certificates or keys. If a long lived
420 process periodically decodes certificates or keys its memory usage will
421 expand without bounds and the process might be terminated by the operating
422 system causing a denial of service. Also traversing the empty hash table
423 entries will take increasingly more time.
425 Typically such long lived processes might be TLS clients or TLS servers
426 configured to accept client certificate authentication.
429 *Hugo Landau, Aliaksei Levin*
431 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
432 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
433 statistics are no longer supported. For compatibility, these statistics are
434 still listed in the output but are now always reported as zero.
438 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
440 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
441 for non-prime moduli.
443 Internally this function is used when parsing certificates that contain
444 elliptic curve public keys in compressed form or explicit elliptic curve
445 parameters with a base point encoded in compressed form.
447 It is possible to trigger the infinite loop by crafting a certificate that
448 has invalid explicit curve parameters.
450 Since certificate parsing happens prior to verification of the certificate
451 signature, any process that parses an externally supplied certificate may thus
452 be subject to a denial of service attack. The infinite loop can also be
453 reached when parsing crafted private keys as they can contain explicit
454 elliptic curve parameters.
456 Thus vulnerable situations include:
458 - TLS clients consuming server certificates
459 - TLS servers consuming client certificates
460 - Hosting providers taking certificates or private keys from customers
461 - Certificate authorities parsing certification requests from subscribers
462 - Anything else which parses ASN.1 elliptic curve parameters
464 Also any other applications that use the BN_mod_sqrt() where the attacker
465 can control the parameter values are vulnerable to this DoS issue.
470 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
471 to the list of ciphersuites providing Perfect Forward Secrecy as
472 required by SECLEVEL >= 3.
474 *Dmitry Belyavskiy, Nicola Tuveri*
476 * Made the AES constant time code for no-asm configurations
477 optional due to the resulting 95% performance degradation.
478 The AES constant time code can be enabled, for no assembly
479 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
483 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
488 * The negative return value handling of the certificate verification callback
489 was reverted. The replacement is to set the verification retry state with
490 the SSL_set_retry_verify() function.
494 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
496 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
497 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
498 verify a certificate supplied by a server. That function may return a
499 negative return value to indicate an internal error (for example out of
500 memory). Such a negative return value is mishandled by OpenSSL and will cause
501 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
502 success and a subsequent call to SSL_get_error() to return the value
503 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
504 returned by OpenSSL if the application has previously called
505 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
506 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
507 totally unexpected and applications may not behave correctly as a result. The
508 exact behaviour will depend on the application but it could result in
509 crashes, infinite loops or other similar incorrect responses.
511 This issue is made more serious in combination with a separate bug in OpenSSL
512 3.0 that will cause X509_verify_cert() to indicate an internal error when
513 processing a certificate chain. This will occur where a certificate does not
514 include the Subject Alternative Name extension but where a Certificate
515 Authority has enforced name constraints. This issue can occur even with valid
521 * Corrected a few file name and file reference bugs in the build,
522 installation and setup scripts, which lead to installation verification
523 failures. Slightly enhanced the installation verification script.
527 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
532 * Fixed PVK encoder to properly query for the passphrase.
536 * Multiple fixes in the OSSL_HTTP API functions.
540 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
541 OSSL_PARAM_INTEGER data type and return error on negative numbers
542 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
543 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
547 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
551 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
555 * Multiple threading fixes.
559 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
563 * Allow fetching an operation from the provider that owns an unexportable key
564 as a fallback if that is still allowed by the property query.
568 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
570 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
575 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
576 S390X capability vector to zero. This simplifies testing of different code
577 paths on S390X architecture.
581 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
582 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
583 SP 800-38D". The communication will fail at this point.
587 * The EC_GROUP_clear_free() function is deprecated as there is nothing
588 confidential in EC_GROUP data.
592 * The byte order mark (BOM) character is ignored if encountered at the
593 beginning of a PEM-formatted file.
597 * Added CMS support for the Russian GOST algorithms.
601 * Due to move of the implementation of cryptographic operations
602 to the providers, validation of various operation parameters can
603 be postponed until the actual operation is executed where previously
604 it happened immediately when an operation parameter was set.
606 For example when setting an unsupported curve with
607 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
608 fail but later keygen operations with the EVP_PKEY_CTX will fail.
610 *OpenSSL team members and many third party contributors*
612 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
613 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
614 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
615 instead to retrieve these algorithms from a provider.
619 * On build targets where the multilib postfix is set in the build
620 configuration the libdir directory was changing based on whether
621 the lib directory with the multilib postfix exists on the system
622 or not. This unpredictable behavior was removed and eventual
623 multilib postfix is now always added to the default libdir. Use
624 `--libdir=lib` to override the libdir if adding the postfix is
629 * The triple DES key wrap functionality now conforms to RFC 3217 but is
630 no longer interoperable with OpenSSL 1.1.1.
634 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
635 function codes, this function can only cause problems for calling
640 * Add a configurable flag to output date formats as ISO 8601. Does not
641 change the default date format.
645 * Version of MSVC earlier than 1300 could get link warnings, which could
646 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
647 Support for this flag has been removed.
651 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
652 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
653 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
654 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
655 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
659 * The signatures of the functions to get and set options on SSL and
660 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
661 Some source code changes may be required.
665 * The public definitions of conf_method_st and conf_st have been
666 deprecated. They will be made opaque in a future release.
668 *Rich Salz and Tomáš Mráz*
670 * Client-initiated renegotiation is disabled by default. To allow it, use
671 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
672 flag, or the "ClientRenegotiation" config parameter as appropriate.
676 * Add "abspath" and "includedir" pragma's to config files, to prevent,
677 or modify relative pathname inclusion.
681 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
682 validated. Please consult the README-FIPS and
683 README-PROVIDERS files, as well as the migration guide.
685 *OpenSSL team members and many third party contributors*
687 * For the key types DH and DHX the allowed settable parameters are now different.
691 * The openssl commands that read keys, certificates, and CRLs now
692 automatically detect the PEM or DER format of the input files.
694 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
696 * Added enhanced PKCS#12 APIs which accept a library context.
700 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
704 * Added support for Kernel TLS (KTLS).
706 *Boris Pismenny, John Baldwin and Andrew Gallatin*
708 * Support for RFC 5746 secure renegotiation is now required by default for
709 SSL or TLS connections to succeed.
713 * The signature of the `copy` functional parameter of the
714 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
715 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
716 the signature of the `pub_decode` functional parameter of the
717 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
718 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
722 * The error return values from some control calls (ctrl) have changed.
726 * A public key check is now performed during EVP_PKEY_derive_set_peer().
730 * Many functions in the EVP_ namespace that are getters of values from
731 implementations or contexts were renamed to include get or get0 in their
732 names. Old names are provided as macro aliases for compatibility and
737 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
738 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
739 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
744 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
747 * The output from the command line applications may have minor
752 * The output from numerous "printing" may have minor changes.
756 * Windows thread synchronization uses read/write primitives (SRWLock) when
757 supported by the OS, otherwise CriticalSection continues to be used.
761 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
762 work on read only BIO source/sinks that do not support these functions.
763 This allows piping or redirection of a file BIO using stdin to be buffered
764 into memory. This is used internally in OSSL_DECODER_from_bio().
768 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
769 this function would return one of the values OSSL_STORE_INFO_NAME,
770 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
771 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
772 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
773 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
774 using this function should be amended to handle the changed return value.
778 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
779 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
780 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
781 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
782 contains more than one certificate identifier: This means that all
783 certificates referenced there MUST be part of the validation chain.
787 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
788 RC5, DESX and DES have been moved to the legacy provider.
792 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
793 RIPEMD-160 have been moved to the legacy provider.
797 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
802 * The deprecated functions EVP_PKEY_get0_RSA(),
803 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
804 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
805 well as the similarly named "get1" functions behave differently in
810 * A number of functions handling low-level keys or engines were deprecated
811 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
812 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
813 EVP_PKEY_get0_siphash().
817 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
818 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
819 will need to load the legacy crypto provider. This includes these PBE
820 algorithms which use this KDF:
821 - NID_pbeWithMD2AndDES_CBC
822 - NID_pbeWithMD5AndDES_CBC
823 - NID_pbeWithSHA1AndRC2_CBC
824 - NID_pbeWithMD2AndRC2_CBC
825 - NID_pbeWithMD5AndRC2_CBC
826 - NID_pbeWithSHA1AndDES_CBC
830 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
831 BIO_debug_callback() functions.
835 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
836 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
840 * The RAND_METHOD APIs have been deprecated.
844 * The SRP APIs have been deprecated.
848 * Add a compile time option to prevent the caching of provider fetched
849 algorithms. This is enabled by including the no-cached-fetch option
850 at configuration time.
854 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
855 count of PKCS12_DEFAULT_ITER.
857 *Tomáš Mráz and Sahana Prasad*
859 * The openssl speed command does not use low-level API calls anymore.
863 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
866 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
868 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
872 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
873 providers may supply their own group implementations (using either the "key
874 exchange" or the "key encapsulation" methods) which will automatically be
875 detected and used by libssl.
877 *Matt Caswell, Nicola Tuveri*
879 * The undocumented function X509_certificate_type() has been deprecated;
883 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
887 * Removed RSA padding mode for SSLv23 (which was only used for
888 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
889 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
894 * Deprecated the obsolete X9.31 RSA key generation related functions.
896 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
897 is not allowed to return a value > 1, this is no more taken as failure.
899 *Viktor Dukhovni and David von Oheimb*
901 * Deprecated the obsolete X9.31 RSA key generation related functions
902 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
903 BN_X931_generate_prime_ex().
907 * The default key generation method for the regular 2-prime RSA keys was
908 changed to the FIPS 186-4 B.3.6 method.
912 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
916 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
920 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
921 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
923 *Rich Salz, Richard Levitte, and David von Oheimb*
925 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
929 * Deprecated `OCSP_parse_url()`.
933 * Validation of SM2 keys has been separated from the validation of regular EC
938 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
939 switches: a validation failure triggers an early exit, returning a failure
940 exit status to the parent process.
944 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
945 to ignore unknown ciphers.
949 * The `-cipher-commands` and `-digest-commands` options
950 of the command line utility `list` have been deprecated.
951 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
955 * Added convenience functions for generating asymmetric key pairs:
956 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
957 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
961 * All of the low-level EC_KEY functions have been deprecated.
963 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
965 * Deprecated all the libcrypto and libssl error string loading
970 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
971 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
976 * The `-crypt` option to the `passwd` command line tool has been removed.
980 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
985 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
989 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
990 EVP_PKEY_get1_tls_encodedpoint().
994 * The security callback, which can be customised by application code, supports
995 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
996 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1000 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1001 interface. Their functionality remains unchanged.
1005 * Added new option for 'openssl list', '-providers', which will display the
1006 list of loaded providers, their names, version and status. It optionally
1007 displays their gettable parameters.
1011 * Removed EVP_PKEY_set_alias_type().
1015 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1016 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1020 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1021 parameters are still checked for type safety at compile time via helper
1026 * Remove the RAND_DRBG API
1028 *Paul Dale and Matthias St. Pierre*
1030 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1031 as well as actual hostnames.
1035 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1036 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1037 conversely, silently ignore DTLS protocol version bounds when configuring
1038 TLS-based contexts. The commands can be repeated to set bounds of both
1039 types. The same applies with the corresponding "min_protocol" and
1040 "max_protocol" command-line switches, in case some application uses both TLS
1043 SSL_CTX instances that are created for a fixed protocol version (e.g.
1044 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1045 attempts to apply bounds to these protocol versions would result in an
1046 error. Now only the "version-flexible" SSL_CTX instances are subject to
1047 limits in configuration files in command-line options.
1051 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1056 * Reworked the recorded ERR codes to make better space for system errors.
1057 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1058 given code is a system error (true) or an OpenSSL error (false).
1062 * Reworked the test perl framework to better allow parallel testing.
1064 *Nicola Tuveri and David von Oheimb*
1066 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1067 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1071 * 'Configure' has been changed to figure out the configuration target if
1072 none is given on the command line. Consequently, the 'config' script is
1073 now only a mere wrapper. All documentation is changed to only mention
1076 *Rich Salz and Richard Levitte*
1078 * Added a library context `OSSL_LIB_CTX` that applications as well as
1079 other libraries can use to form a separate context within which
1080 libcrypto operations are performed.
1084 * Added various `_ex` functions to the OpenSSL API that support using
1085 a non-default `OSSL_LIB_CTX`.
1089 * Handshake now fails if Extended Master Secret extension is dropped
1094 * Dropped interactive mode from the `openssl` program.
1098 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1100 *David von Oheimb and Shane Lontis*
1102 * Deprecated `EC_METHOD_get_field_type()`.
1106 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1107 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1108 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1112 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1116 * Add CAdES-BES signature verification support, mostly derived
1117 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1119 *Filipe Raimundo da Silva*
1121 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1125 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1126 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1130 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1134 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1135 EC_KEY_precompute_mult().
1139 * Deprecated EC_POINTs_mul().
1143 * Removed FIPS_mode() and FIPS_mode_set().
1147 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1151 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1152 EC_POINT_get_Jprojective_coordinates_GFp().
1156 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1157 arrays to be more easily constructed via a series of utility functions.
1158 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1159 the various push functions and finally convert to a passable OSSL_PARAM
1160 array using OSSL_PARAM_BLD_to_param().
1164 * The security strength of SHA1 and MD5 based signatures in TLS has been
1169 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1170 contain a provider side internal key.
1174 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1178 * Project text documents not yet having a proper file name extension
1179 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1180 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1184 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1185 have been converted to Markdown with the goal to produce documents
1186 which not only look pretty when viewed online in the browser, but
1187 remain well readable inside a plain text editor.
1189 To achieve this goal, a 'minimalistic' Markdown style has been applied
1190 which avoids formatting elements that interfere too much with the
1191 reading flow in the text file. For example, it
1193 * avoids [ATX headings][] and uses [setext headings][] instead
1194 (which works for `<h1>` and `<h2>` headings only).
1195 * avoids [inline links][] and uses [reference links][] instead.
1196 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1198 [ATX headings]: https://github.github.com/gfm/#atx-headings
1199 [setext headings]: https://github.github.com/gfm/#setext-headings
1200 [inline links]: https://github.github.com/gfm/#inline-link
1201 [reference links]: https://github.github.com/gfm/#reference-link
1202 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1203 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1205 *Matthias St. Pierre*
1207 * The test suite is changed to preserve results of each test recipe.
1208 A new directory test-runs/ with subdirectories named like the
1209 test recipes are created in the build tree for this purpose.
1213 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1214 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1215 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1217 *David von Oheimb, Martin Peylo*
1219 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1220 It supports arbitrary request and response content types, GET redirection,
1221 TLS, connections via HTTP(S) proxies, connections and exchange via
1222 user-defined BIOs (allowing implicit connections), persistent connections,
1223 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1224 The legacy OCSP-focused (and only partly documented) API
1225 is retained for backward compatibility, while most of it is deprecated.
1229 * Added `util/check-format.pl`, a tool for checking adherence to the
1230 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1231 The checks performed are incomplete and yield some false positives.
1232 Still the tool should be useful for detecting most typical glitches.
1236 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1237 If domain name resolution yields multiple IP addresses all of them are tried
1238 after `connect()` failures.
1242 * All of the low-level RSA functions have been deprecated.
1246 * X509 certificates signed using SHA1 are no longer allowed at security
1251 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1252 modified to use PKEY APIs. These commands are now in maintenance mode
1253 and no new features will be added to them.
1257 * The command line utility rsautl has been deprecated.
1261 * The command line utilities genrsa and rsa have been modified to use PKEY
1262 APIs. They now write PKCS#8 keys by default. These commands are now in
1263 maintenance mode and no new features will be added to them.
1267 * All of the low-level DH functions have been deprecated.
1269 *Paul Dale and Matt Caswell*
1271 * All of the low-level DSA functions have been deprecated.
1275 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1276 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1280 * Deprecated low-level ECDH and ECDSA functions.
1284 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1288 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1289 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1290 a new formulation to include all the things it can be used for,
1291 as well as words of caution.
1295 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1299 * All of the low-level HMAC functions have been deprecated.
1301 *Paul Dale and David von Oheimb*
1303 * Over two thousand fixes were made to the documentation, including:
1304 - Common options (such as -rand/-writerand, TLS version control, etc)
1305 were refactored and point to newly-enhanced descriptions in openssl.pod.
1306 - Added style conformance for all options (with help from Richard Levitte),
1307 documented all reported missing options, added a CI build to check
1308 that all options are documented and that no unimplemented options
1310 - Documented some internals, such as all use of environment variables.
1311 - Addressed all internal broken L<> references.
1315 * All of the low-level CMAC functions have been deprecated.
1319 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1320 functions have been deprecated.
1322 *Paul Dale and David von Oheimb*
1324 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1325 set of functions. The documentation mentioned negative values for some
1326 errors, but this was never the case, so the mention of negative values
1329 Code that followed the documentation and thereby check with something
1330 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1334 * All of the low-level cipher functions have been deprecated.
1336 *Matt Caswell and Paul Dale*
1338 * Removed include/openssl/opensslconf.h.in and replaced it with
1339 include/openssl/configuration.h.in, which differs in not including
1340 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1341 was added to include both.
1343 This allows internal hacks where one might need to modify the set
1344 of configured macros, for example this if deprecated symbols are
1345 still supposed to be available internally:
1347 #include <openssl/configuration.h>
1349 #undef OPENSSL_NO_DEPRECATED
1350 #define OPENSSL_SUPPRESS_DEPRECATED
1352 #include <openssl/macros.h>
1354 This should not be used by applications that use the exported
1355 symbols, as that will lead to linking errors.
1359 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1360 used in exponentiation with 512-bit moduli. No EC algorithms are
1361 affected. Analysis suggests that attacks against 2-prime RSA1024,
1362 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1363 difficult to perform and are not believed likely. Attacks against DH512
1364 are considered just feasible. However, for an attack the target would
1365 have to re-use the DH512 private key, which is not recommended anyway.
1366 Also applications directly using the low-level API BN_mod_exp may be
1367 affected if they use BN_FLG_CONSTTIME.
1372 * Most memory-debug features have been deprecated, and the functionality
1373 replaced with no-ops.
1377 * Added documentation for the STACK API.
1381 * Introduced a new method type and API, OSSL_ENCODER, to represent
1382 generic encoders. These do the same sort of job that PEM writers
1383 and d2i functions do, but with support for methods supplied by
1384 providers, and the possibility for providers to support other
1389 * Introduced a new method type and API, OSSL_DECODER, to represent
1390 generic decoders. These do the same sort of job that PEM readers
1391 and i2d functions do, but with support for methods supplied by
1392 providers, and the possibility for providers to support other
1397 * Added a .pragma directive to the syntax of configuration files, to
1398 allow varying behavior in a supported and predictable manner.
1399 Currently added pragma:
1403 This allows dollar signs to be a keyword character unless it's
1404 followed by a opening brace or parenthesis. This is useful for
1405 platforms where dollar signs are commonly used in names, such as
1406 volume names and system directory names on VMS.
1410 * Added functionality to create an EVP_PKEY from user data.
1414 * Change the interpretation of the '--api' configuration option to
1415 mean that this is a desired API compatibility level with no
1416 further meaning. The previous interpretation, that this would
1417 also mean to remove all deprecated symbols up to and including
1418 the given version, no requires that 'no-deprecated' is also used
1419 in the configuration.
1421 When building applications, the desired API compatibility level
1422 can be set with the OPENSSL_API_COMPAT macro like before. For
1423 API compatibility version below 3.0, the old style numerical
1424 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1425 For version 3.0 and on, the value is expected to be the decimal
1426 value calculated from the major and minor version like this:
1428 MAJOR * 10000 + MINOR * 100
1432 -DOPENSSL_API_COMPAT=30000 For 3.0
1433 -DOPENSSL_API_COMPAT=30200 For 3.2
1435 To hide declarations that are deprecated up to and including the
1436 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1437 given when building the application as well.
1441 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1442 access to certificate and CRL stores via URIs and OSSL_STORE
1445 This adds the following functions:
1447 - X509_LOOKUP_store()
1448 - X509_STORE_load_file()
1449 - X509_STORE_load_path()
1450 - X509_STORE_load_store()
1451 - SSL_add_store_cert_subjects_to_stack()
1452 - SSL_CTX_set_default_verify_store()
1453 - SSL_CTX_load_verify_file()
1454 - SSL_CTX_load_verify_dir()
1455 - SSL_CTX_load_verify_store()
1459 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1460 The presence of this system service is determined at run-time.
1464 * Added functionality to create an EVP_PKEY context based on data
1465 for methods from providers. This takes an algorithm name and a
1466 property query string and simply stores them, with the intent
1467 that any operation that uses this context will use those strings
1468 to fetch the needed methods implicitly, thereby making the port
1469 of application written for pre-3.0 OpenSSL easier.
1473 * The undocumented function NCONF_WIN32() has been deprecated; for
1474 conversion details see the HISTORY section of doc/man5/config.pod
1478 * Introduced the new functions EVP_DigestSignInit_ex() and
1479 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1480 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1481 pages for further details.
1485 * Over two thousand fixes were made to the documentation, including:
1486 adding missing command flags, better style conformance, documentation
1489 *Rich Salz, Richard Levitte*
1491 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1492 X25519, X448, Ed25519 and Ed448.
1496 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1501 * Deprecated the public definition of `ERR_STATE` as well as the function
1502 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1507 * Added ERR functionality to give callers access to the stored function
1508 names that have replaced the older function code based functions.
1510 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1511 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1512 ERR_peek_error_all() and ERR_peek_last_error_all().
1514 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1515 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1516 ERR_func_error_string().
1520 * Extended testing to be verbose for failing tests only. The make variables
1521 VERBOSE_FAILURE or VF can be used to enable this:
1523 $ make VF=1 test # Unix
1524 $ mms /macro=(VF=1) test ! OpenVMS
1525 $ nmake VF=1 test # Windows
1529 * Added the `-copy_extensions` option to the `x509` command for use with
1530 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1531 all extensions in the request are copied to the certificate or vice versa.
1533 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1535 * Added the `-copy_extensions` option to the `req` command for use with
1536 `-x509`. When given with the `copy` or `copyall` argument,
1537 all extensions in the certification request are copied to the certificate.
1541 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1542 they generate are by default RFC 5280 compliant in the following sense:
1543 There is a subjectKeyIdentifier extension with a hash value of the public key
1544 and for not self-signed certs there is an authorityKeyIdentifier extension
1545 with a keyIdentifier field or issuer information identifying the signing key.
1546 This is done unless some configuration overrides the new default behavior,
1547 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1551 * Added several checks to `X509_verify_cert()` according to requirements in
1552 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1553 (which may be done by using the CLI option `-x509_strict`):
1554 * The basicConstraints of CA certificates must be marked critical.
1555 * CA certificates must explicitly include the keyUsage extension.
1556 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1557 * The issuer name of any certificate must not be empty.
1558 * The subject name of CA certs, certs with keyUsage crlSign,
1559 and certs without subjectAlternativeName must not be empty.
1560 * If a subjectAlternativeName extension is given it must not be empty.
1561 * The signatureAlgorithm field and the cert signature must be consistent.
1562 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1563 must not be marked critical.
1564 * The authorityKeyIdentifier must be given for X.509v3 certs
1565 unless they are self-signed.
1566 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1570 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1571 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1575 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1576 used even when parsing explicit parameters, when loading a encoded key
1577 or calling `EC_GROUP_new_from_ecpkparameters()`/
1578 `EC_GROUP_new_from_ecparameters()`.
1579 This prevents bypass of security hardening and performance gains,
1580 especially for curves with specialized EC_METHODs.
1581 By default, if a key encoded with explicit parameters is loaded and later
1582 encoded, the output is still encoded with explicit parameters, even if
1583 internally a "named" EC_GROUP is used for computation.
1587 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1588 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1589 NULL. After this change, only the cofactor parameter can be NULL. It also
1590 does some minimal sanity checks on the passed order.
1595 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1596 An attack is simple, if the first CMS_recipientInfo is valid but the
1597 second CMS_recipientInfo is chosen ciphertext. If the second
1598 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1599 encryption key will be replaced by garbage, and the message cannot be
1600 decoded, but if the RSA decryption fails, the correct encryption key is
1601 used and the recipient will not notice the attack.
1602 As a work around for this potential attack the length of the decrypted
1603 key must be equal to the cipher default key length, in case the
1604 certificate is not given and all recipientInfo are tried out.
1605 The old behaviour can be re-enabled in the CMS code by setting the
1606 CMS_DEBUG_DECRYPT flag.
1610 * Early start up entropy quality from the DEVRANDOM seed source has been
1611 improved for older Linux systems. The RAND subsystem will wait for
1612 /dev/random to be producing output before seeding from /dev/urandom.
1613 The seeded state is stored for future library initialisations using
1614 a system global shared memory segment. The shared memory identifier
1615 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1616 the desired value. The default identifier is 114.
1620 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1621 when primes for RSA keys are computed.
1622 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1623 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1624 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1625 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1626 This avoids possible fingerprinting of newly generated RSA modules.
1630 * Correct the extended master secret constant on EBCDIC systems. Without this
1631 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1632 negotiate EMS will fail. Unfortunately this also means that TLS connections
1633 between EBCDIC systems with this fix, and EBCDIC systems without this
1634 fix will fail if they negotiate EMS.
1638 * Changed the library initialisation so that the config file is now loaded
1639 by default. This was already the case for libssl. It now occurs for both
1640 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1641 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1645 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1646 where the former acts as a replacement for `ERR_put_error()`, and the
1647 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1648 `ERR_raise_data()` adds more flexibility by taking a format string and
1649 an arbitrary number of arguments following it, to be processed with
1654 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1655 to check if a named provider is loaded and available. When called, it
1656 will also activate all fallback providers if such are still present.
1660 * Enforce a minimum DH modulus size of 512 bits.
1664 * Changed DH parameters to generate the order q subgroup instead of 2q.
1665 Previously generated DH parameters are still accepted by DH_check
1666 but DH_generate_key works around that by clearing bit 0 of the
1667 private key for those. This avoids leaking bit 0 of the private key.
1671 * Significantly reduce secure memory usage by the randomness pools.
1675 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1680 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1681 algorithms. An implementation of a key exchange algorithm can be obtained
1682 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1683 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1684 the older EVP_PKEY_derive_init() function. See the man pages for the new
1685 functions for further details.
1689 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1693 * Removed the function names from error messages and deprecated the
1698 * Removed NextStep support and the macro OPENSSL_UNISTD
1702 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1703 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1704 Also removed "export var as function" capability; we do not export
1705 variables, only functions.
1709 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1710 an error and 1 indicating success. In previous versions of OpenSSL this
1711 was a void type. If a key was set longer than the maximum possible this
1716 * Support SM2 signing and verification schemes with X509 certificate.
1720 * Use SHA256 as the default digest for TS query in the `ts` app.
1724 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1728 * Default cipher lists/suites are now available via a function, the
1729 #defines are deprecated.
1733 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1734 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1735 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1739 * Join the directories crypto/x509 and crypto/x509v3
1743 * Added command 'openssl kdf' that uses the EVP_KDF API.
1747 * Added command 'openssl mac' that uses the EVP_MAC API.
1751 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1752 as default directories. Also added the command 'openssl info'
1753 for scripting purposes.
1757 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1762 * Add prediction resistance to the DRBG reseeding process.
1766 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1767 mandated by IEEE Std 1619-2018.
1771 * Added newline escaping functionality to a filename when using openssl dgst.
1772 This output format is to replicate the output format found in the `*sum`
1773 checksum programs. This aims to preserve backward compatibility.
1775 *Matt Eaton, Richard Levitte, and Paul Dale*
1777 * Removed the heartbeat message in DTLS feature, as it has very
1778 little usage and doesn't seem to fulfill a valuable purpose.
1779 The configuration option is now deprecated.
1783 * Changed the output of 'openssl {digestname} < file' to display the
1784 digest name in its output.
1788 * Added a new generic trace API which provides support for enabling
1789 instrumentation through trace output.
1791 *Richard Levitte & Matthias St. Pierre*
1793 * Added build tests for C++. These are generated files that only do one
1794 thing, to include one public OpenSSL head file each. This tests that
1795 the public header files can be usefully included in a C++ application.
1797 This test isn't enabled by default. It can be enabled with the option
1798 'enable-buildtest-c++'.
1802 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1806 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1810 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1814 * Added KMAC to EVP_MAC.
1818 * Added property based algorithm implementation selection framework to
1823 * Added SCA hardening for modular field inversion in EC_GROUP through
1824 a new dedicated field_inv() pointer in EC_METHOD.
1825 This also addresses a leakage affecting conversions from projective
1826 to affine coordinates.
1828 *Billy Bob Brumley, Nicola Tuveri*
1830 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1831 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1832 those algorithms that were already supported through the EVP_PKEY API
1833 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1834 and scrypt are now wrappers that call EVP_KDF.
1838 * Build devcrypto engine as a dynamic engine.
1840 *Eneas U de Queiroz*
1842 * Add keyed BLAKE2 to EVP_MAC.
1846 * Fix a bug in the computation of the endpoint-pair shared secret used
1847 by DTLS over SCTP. This breaks interoperability with older versions
1848 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1849 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1850 interoperability with such broken implementations. However, enabling
1851 this switch breaks interoperability with correct implementations.
1853 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1854 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1858 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1862 * Changed the license to the Apache License v2.0.
1866 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1868 - Major releases (indicated by incrementing the MAJOR release number)
1869 may introduce incompatible API/ABI changes.
1870 - Minor releases (indicated by incrementing the MINOR release number)
1871 may introduce new features but retain API/ABI compatibility.
1872 - Patch releases (indicated by incrementing the PATCH number)
1873 are intended for bug fixes and other improvements of existing
1874 features only (like improving performance or adding documentation)
1875 and retain API/ABI compatibility.
1879 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1883 * Remove the 'dist' target and add a tarball building script. The
1884 'dist' target has fallen out of use, and it shouldn't be
1885 necessary to configure just to create a source distribution.
1889 * Recreate the OS390-Unix config target. It no longer relies on a
1890 special script like it did for OpenSSL pre-1.1.0.
1894 * Instead of having the source directories listed in Configure, add
1895 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1900 * Add GMAC to EVP_MAC.
1904 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1908 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1909 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1910 to facilitate the continued use of MACs through raw private keys in
1911 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1915 * Deprecate ECDH_KDF_X9_62().
1919 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1920 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1921 are retained for backwards compatibility.
1925 * AES-XTS mode now enforces that its two keys are different to mitigate
1926 the attacked described in "Efficient Instantiations of Tweakable
1927 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1928 Details of this attack can be obtained from:
1929 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1933 * Rename the object files, i.e. give them other names than in previous
1934 versions. Their names now include the name of the final product, as
1935 well as its type mnemonic (bin, lib, shlib).
1939 * Added new option for 'openssl list', '-objects', which will display the
1940 list of built in objects, i.e. OIDs with names.
1944 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1945 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1950 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1951 improves application performance by removing data copies and providing
1952 applications with zero-copy system calls such as sendfile and splice.
1956 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1960 * `PKCS12_parse` now maintains the order of the parsed certificates
1961 when outputting them via `*ca` (rather than reversing it).
1965 * Deprecated pthread fork support methods.
1969 * Added support for FFDHE key exchange in TLS 1.3.
1973 * Added a new concept for OpenSSL plugability: providers. This
1974 functionality is designed to replace the ENGINE API and ENGINE
1975 implementations, and to be much more dynamic, allowing provider
1976 authors to introduce new algorithms among other things, as long as
1977 there's an API that supports the algorithm type.
1979 With this concept comes a new core API for interaction between
1980 libcrypto and provider implementations. Public libcrypto functions
1981 that want to use providers do so through this core API.
1983 The main documentation for this core API is found in
1984 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1985 refer to other manuals describing the API specific for supported
1986 algorithm types (also called operations).
1993 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1995 * Avoid loading of a dynamic engine twice.
1999 * Prioritise DANE TLSA issuer certs over peer certs
2003 * Fixed random API for MacOS prior to 10.12
2005 These MacOS versions don't support the CommonCrypto APIs
2009 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2011 * Fixed an SM2 Decryption Buffer Overflow.
2013 In order to decrypt SM2 encrypted data an application is expected to
2014 call the API function EVP_PKEY_decrypt(). Typically an application will
2015 call this function twice. The first time, on entry, the "out" parameter
2016 can be NULL and, on exit, the "outlen" parameter is populated with the
2017 buffer size required to hold the decrypted plaintext. The application
2018 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2019 again, but this time passing a non-NULL value for the "out" parameter.
2021 A bug in the implementation of the SM2 decryption code means that the
2022 calculation of the buffer size required to hold the plaintext returned
2023 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2024 size required by the second call. This can lead to a buffer overflow
2025 when EVP_PKEY_decrypt() is called by the application a second time with
2026 a buffer that is too small.
2028 A malicious attacker who is able present SM2 content for decryption to
2029 an application could cause attacker chosen data to overflow the buffer
2030 by up to a maximum of 62 bytes altering the contents of other data held
2031 after the buffer, possibly changing application behaviour or causing
2032 the application to crash. The location of the buffer is application
2033 dependent but is typically heap allocated.
2038 * Fixed various read buffer overruns processing ASN.1 strings
2040 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2041 structure which contains a buffer holding the string data and a field
2042 holding the buffer length. This contrasts with normal C strings which
2043 are represented as a buffer for the string data which is terminated
2044 with a NUL (0) byte.
2046 Although not a strict requirement, ASN.1 strings that are parsed using
2047 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2048 well as any string whose value has been set with the ASN1_STRING_set()
2049 function will additionally NUL terminate the byte array in the
2050 ASN1_STRING structure.
2052 However, it is possible for applications to directly construct valid
2053 ASN1_STRING structures which do not NUL terminate the byte array by
2054 directly setting the "data" and "length" fields in the ASN1_STRING
2055 array. This can also happen by using the ASN1_STRING_set0() function.
2057 Numerous OpenSSL functions that print ASN.1 data have been found to
2058 assume that the ASN1_STRING byte array will be NUL terminated, even
2059 though this is not guaranteed for strings that have been directly
2060 constructed. Where an application requests an ASN.1 structure to be
2061 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2062 been directly constructed by the application without NUL terminating
2063 the "data" field, then a read buffer overrun can occur.
2065 The same thing can also occur during name constraints processing
2066 of certificates (for example if a certificate has been directly
2067 constructed by the application instead of loading it via the OpenSSL
2068 parsing functions, and the certificate contains non NUL terminated
2069 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2070 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2072 If a malicious actor can cause an application to directly construct an
2073 ASN1_STRING and then process it through one of the affected OpenSSL
2074 functions then this issue could be hit. This might result in a crash
2075 (causing a Denial of Service attack). It could also result in the
2076 disclosure of private memory contents (such as private keys, or
2077 sensitive plaintext).
2082 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2084 * Fixed a problem with verifying a certificate chain when using the
2085 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2086 the certificates present in a certificate chain. It is not set by default.
2088 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2089 the chain that have explicitly encoded elliptic curve parameters was added
2090 as an additional strict check.
2092 An error in the implementation of this check meant that the result of a
2093 previous check to confirm that certificates in the chain are valid CA
2094 certificates was overwritten. This effectively bypasses the check
2095 that non-CA certificates must not be able to issue other certificates.
2097 If a "purpose" has been configured then there is a subsequent opportunity
2098 for checks that the certificate is a valid CA. All of the named "purpose"
2099 values implemented in libcrypto perform this check. Therefore, where
2100 a purpose is set the certificate chain will still be rejected even when the
2101 strict flag has been used. A purpose is set by default in libssl client and
2102 server certificate verification routines, but it can be overridden or
2103 removed by an application.
2105 In order to be affected, an application must explicitly set the
2106 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2107 for the certificate verification or, in the case of TLS client or server
2108 applications, override the default purpose.
2113 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2114 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2115 renegotiation ClientHello omits the signature_algorithms extension (where it
2116 was present in the initial ClientHello), but includes a
2117 signature_algorithms_cert extension then a NULL pointer dereference will
2118 result, leading to a crash and a denial of service attack.
2120 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2121 (which is the default configuration). OpenSSL TLS clients are not impacted by
2125 *Peter Kästle and Samuel Sapalski*
2127 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2129 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2130 create a unique hash value based on the issuer and serial number data
2131 contained within an X509 certificate. However, it was failing to correctly
2132 handle any errors that may occur while parsing the issuer field (which might
2133 occur if the issuer field is maliciously constructed). This may subsequently
2134 result in a NULL pointer deref and a crash leading to a potential denial of
2140 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2141 padding mode to correctly check for rollback attacks. This is considered a
2142 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2147 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2148 functions. Previously they could overflow the output length argument in some
2149 cases where the input length is close to the maximum permissible length for
2150 an integer on the platform. In such cases the return value from the function
2151 call would be 1 (indicating success), but the output length value would be
2152 negative. This could cause applications to behave incorrectly or crash.
2157 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2158 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2159 could be exploited in a side channel attack to recover the password. Since
2160 the attack is local host only this is outside of the current OpenSSL
2161 threat model and therefore no CVE is assigned.
2163 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2168 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2170 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2171 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2172 If an attacker can control both items being compared then this could lead
2173 to a possible denial of service attack. OpenSSL itself uses the
2174 GENERAL_NAME_cmp function for two purposes:
2175 1) Comparing CRL distribution point names between an available CRL and a
2176 CRL distribution point embedded in an X509 certificate
2177 2) When verifying that a timestamp response token signer matches the
2178 timestamp authority name (exposed via the API functions
2179 TS_RESP_verify_response and TS_RESP_verify_token)
2184 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2186 * Certificates with explicit curve parameters are now disallowed in
2187 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2191 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2192 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2193 conversely, silently ignore DTLS protocol version bounds when configuring
2194 TLS-based contexts. The commands can be repeated to set bounds of both
2195 types. The same applies with the corresponding "min_protocol" and
2196 "max_protocol" command-line switches, in case some application uses both TLS
2199 SSL_CTX instances that are created for a fixed protocol version (e.g.
2200 TLSv1_server_method()) also silently ignore version bounds. Previously
2201 attempts to apply bounds to these protocol versions would result in an
2202 error. Now only the "version-flexible" SSL_CTX instances are subject to
2203 limits in configuration files in command-line options.
2207 * Handshake now fails if Extended Master Secret extension is dropped
2212 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2214 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2216 * Fixed segmentation fault in SSL_check_chain()
2217 Server or client applications that call the SSL_check_chain() function
2218 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2219 dereference as a result of incorrect handling of the
2220 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2221 or unrecognised signature algorithm is received from the peer. This could
2222 be exploited by a malicious peer in a Denial of Service attack.
2227 * Added AES consttime code for no-asm configurations
2228 an optional constant time support for AES was added
2229 when building openssl for no-asm.
2230 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2231 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2232 At this time this feature is by default disabled.
2233 It will be enabled by default in 3.0.
2237 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2239 * Revert the change of EOF detection while reading in libssl to avoid
2240 regressions in applications depending on the current way of reporting
2241 the EOF. As the existing method is not fully accurate the change to
2242 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2243 branch and will be present in the 3.0 release.
2247 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2248 when primes for RSA keys are computed.
2249 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2250 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2251 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2252 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2253 This avoids possible fingerprinting of newly generated RSA modules.
2257 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2259 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2260 while reading in libssl then we would report an error back to the
2261 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2262 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2263 therefore give a hint as to what went wrong.
2267 * Check that ed25519 and ed448 are allowed by the security level. Previously
2268 signature algorithms not using an MD were not being checked that they were
2269 allowed by the security level.
2273 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2274 was not quite right. The behaviour was not consistent between resumption
2275 and normal handshakes, and also not quite consistent with historical
2276 behaviour. The behaviour in various scenarios has been clarified and
2277 it has been updated to make it match historical behaviour as closely as
2282 * *[VMS only]* The header files that the VMS compilers include automatically,
2283 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2284 that the C++ compiler doesn't understand. This is a shortcoming in the
2285 compiler, but can be worked around with `__cplusplus` guards.
2287 C++ applications that use OpenSSL libraries must be compiled using the
2288 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2289 functions. Otherwise, only functions with symbols of less than 31
2290 characters can be used, as the linker will not be able to successfully
2291 resolve symbols with longer names.
2295 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2296 The presence of this system service is determined at run-time.
2300 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2305 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2307 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2308 number generator (RNG). This was intended to include protection in the
2309 event of a fork() system call in order to ensure that the parent and child
2310 processes did not share the same RNG state. However, this protection was not
2311 being used in the default case.
2313 A partial mitigation for this issue is that the output from a high
2314 precision timer is mixed into the RNG state so the likelihood of a parent
2315 and child process sharing state is significantly reduced.
2317 If an application already calls OPENSSL_init_crypto() explicitly using
2318 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2321 *Matthias St. Pierre*
2323 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2324 used even when parsing explicit parameters, when loading a encoded key
2325 or calling `EC_GROUP_new_from_ecpkparameters()`/
2326 `EC_GROUP_new_from_ecparameters()`.
2327 This prevents bypass of security hardening and performance gains,
2328 especially for curves with specialized EC_METHODs.
2329 By default, if a key encoded with explicit parameters is loaded and later
2330 encoded, the output is still encoded with explicit parameters, even if
2331 internally a "named" EC_GROUP is used for computation.
2335 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2336 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2337 NULL. After this change, only the cofactor parameter can be NULL. It also
2338 does some minimal sanity checks on the passed order.
2343 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2344 An attack is simple, if the first CMS_recipientInfo is valid but the
2345 second CMS_recipientInfo is chosen ciphertext. If the second
2346 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2347 encryption key will be replaced by garbage, and the message cannot be
2348 decoded, but if the RSA decryption fails, the correct encryption key is
2349 used and the recipient will not notice the attack.
2350 As a work around for this potential attack the length of the decrypted
2351 key must be equal to the cipher default key length, in case the
2352 certificate is not given and all recipientInfo are tried out.
2353 The old behaviour can be re-enabled in the CMS code by setting the
2354 CMS_DEBUG_DECRYPT flag.
2359 * Early start up entropy quality from the DEVRANDOM seed source has been
2360 improved for older Linux systems. The RAND subsystem will wait for
2361 /dev/random to be producing output before seeding from /dev/urandom.
2362 The seeded state is stored for future library initialisations using
2363 a system global shared memory segment. The shared memory identifier
2364 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2365 the desired value. The default identifier is 114.
2369 * Correct the extended master secret constant on EBCDIC systems. Without this
2370 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2371 negotiate EMS will fail. Unfortunately this also means that TLS connections
2372 between EBCDIC systems with this fix, and EBCDIC systems without this
2373 fix will fail if they negotiate EMS.
2377 * Use Windows installation paths in the mingw builds
2379 Mingw isn't a POSIX environment per se, which means that Windows
2380 paths should be used for installation.
2385 * Changed DH_check to accept parameters with order q and 2q subgroups.
2386 With order 2q subgroups the bit 0 of the private key is not secret
2387 but DH_generate_key works around that by clearing bit 0 of the
2388 private key for those. This avoids leaking bit 0 of the private key.
2392 * Significantly reduce secure memory usage by the randomness pools.
2396 * Revert the DEVRANDOM_WAIT feature for Linux systems
2398 The DEVRANDOM_WAIT feature added a select() call to wait for the
2399 /dev/random device to become readable before reading from the
2400 /dev/urandom device.
2402 It turned out that this change had negative side effects on
2403 performance which were not acceptable. After some discussion it
2404 was decided to revert this feature and leave it up to the OS
2405 resp. the platform maintainer to ensure a proper initialization
2406 during early boot time.
2408 *Matthias St. Pierre*
2410 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2412 * Add build tests for C++. These are generated files that only do one
2413 thing, to include one public OpenSSL head file each. This tests that
2414 the public header files can be usefully included in a C++ application.
2416 This test isn't enabled by default. It can be enabled with the option
2417 'enable-buildtest-c++'.
2421 * Enable SHA3 pre-hashing for ECDSA and DSA.
2425 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2426 This changes the size when using the `genpkey` command when no size is given.
2427 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2428 generation commands to use 2048 bits by default.
2432 * Reorganize the manual pages to consistently have RETURN VALUES,
2433 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2434 util/fix-doc-nits accordingly.
2436 *Paul Yang, Joshua Lock*
2438 * Add the missing accessor EVP_PKEY_get0_engine()
2442 * Have commands like `s_client` and `s_server` output the signature scheme
2443 along with other cipher suite parameters when debugging.
2447 * Make OPENSSL_config() error agnostic again.
2451 * Do the error handling in RSA decryption constant time.
2455 * Prevent over long nonces in ChaCha20-Poly1305.
2457 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2458 for every encryption operation. RFC 7539 specifies that the nonce value
2459 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2460 and front pads the nonce with 0 bytes if it is less than 12
2461 bytes. However it also incorrectly allows a nonce to be set of up to 16
2462 bytes. In this case only the last 12 bytes are significant and any
2463 additional leading bytes are ignored.
2465 It is a requirement of using this cipher that nonce values are
2466 unique. Messages encrypted using a reused nonce value are susceptible to
2467 serious confidentiality and integrity attacks. If an application changes
2468 the default nonce length to be longer than 12 bytes and then makes a
2469 change to the leading bytes of the nonce expecting the new value to be a
2470 new unique nonce then such an application could inadvertently encrypt
2471 messages with a reused nonce.
2473 Additionally the ignored bytes in a long nonce are not covered by the
2474 integrity guarantee of this cipher. Any application that relies on the
2475 integrity of these ignored leading bytes of a long nonce may be further
2476 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2477 is safe because no such use sets such a long nonce value. However user
2478 applications that use this cipher directly and set a non-default nonce
2479 length to be longer than 12 bytes may be vulnerable.
2481 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2487 * Add DEVRANDOM_WAIT feature for Linux systems
2489 On older Linux systems where the getrandom() system call is not available,
2490 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2491 Contrary to getrandom(), the /dev/urandom device will not block during
2492 early boot when the kernel CSPRNG has not been seeded yet.
2494 To mitigate this known weakness, use select() to wait for /dev/random to
2495 become readable before reading from /dev/urandom.
2497 * Ensure that SM2 only uses SM3 as digest algorithm
2501 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2503 * Change the info callback signals for the start and end of a post-handshake
2504 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2505 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2506 confused by this and assume that a TLSv1.2 renegotiation has started. This
2507 can break KeyUpdate handling. Instead we no longer signal the start and end
2508 of a post handshake message exchange (although the messages themselves are
2509 still signalled). This could break some applications that were expecting
2510 the old signals. However without this KeyUpdate is not usable for many
2515 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2517 * Timing vulnerability in DSA signature generation
2519 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2520 timing side channel attack. An attacker could use variations in the signing
2521 algorithm to recover the private key.
2523 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2528 * Timing vulnerability in ECDSA signature generation
2530 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2531 timing side channel attack. An attacker could use variations in the signing
2532 algorithm to recover the private key.
2534 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2539 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2540 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2541 of two gigabytes and the error handling improved.
2543 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2544 categorized as a normal bug, not a security issue, because the DRBG reseeds
2545 automatically and is fully functional even without additional randomness
2546 provided by the application.
2548 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2550 * Add a new ClientHello callback. Provides a callback interface that gives
2551 the application the ability to adjust the nascent SSL object at the
2552 earliest stage of ClientHello processing, immediately after extensions have
2553 been collected but before they have been processed. In particular, this
2554 callback can adjust the supported TLS versions in response to the contents
2559 * Add SM2 base algorithm support.
2563 * s390x assembly pack: add (improved) hardware-support for the following
2564 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2565 aes-cfb/cfb8, aes-ecb.
2569 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2570 parameter is no longer accepted, as it leads to a corrupt table. NULL
2571 pem_str is reserved for alias entries only.
2575 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2576 step for prime curves. The new implementation is based on formulae from
2577 differential addition-and-doubling in homogeneous projective coordinates
2578 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2579 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2580 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2581 to work in projective coordinates.
2583 *Billy Bob Brumley, Nicola Tuveri*
2585 * Change generating and checking of primes so that the error rate of not
2586 being prime depends on the intended use based on the size of the input.
2587 For larger primes this will result in more rounds of Miller-Rabin.
2588 The maximal error rate for primes with more than 1080 bits is lowered
2591 *Kurt Roeckx, Annie Yousar*
2593 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2597 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2598 moving between systems, and to avoid confusion when a Windows build is
2599 done with mingw vs with MSVC. For POSIX installs, there's still a
2600 symlink or copy named 'tsget' to avoid that confusion as well.
2604 * Revert blinding in ECDSA sign and instead make problematic addition
2605 length-invariant. Switch even to fixed-length Montgomery multiplication.
2609 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2610 step for binary curves. The new implementation is based on formulae from
2611 differential addition-and-doubling in mixed Lopez-Dahab projective
2612 coordinates, modified to independently blind the operands.
2614 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2616 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2617 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2618 EC_METHODs to implement their own specialized "ladder step", to take
2619 advantage of more favorable coordinate systems or more efficient
2620 differential addition-and-doubling algorithms.
2622 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2624 * Modified the random device based seed sources to keep the relevant
2625 file descriptors open rather than reopening them on each access.
2626 This allows such sources to operate in a chroot() jail without
2627 the associated device nodes being available. This behaviour can be
2628 controlled using RAND_keep_random_devices_open().
2632 * Numerous side-channel attack mitigations have been applied. This may have
2633 performance impacts for some algorithms for the benefit of improved
2634 security. Specific changes are noted in this change log by their respective
2639 * AIX shared library support overhaul. Switch to AIX "natural" way of
2640 handling shared libraries, which means collecting shared objects of
2641 different versions and bitnesses in one common archive. This allows to
2642 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2643 doesn't affect the way 3rd party applications are linked, only how
2644 multi-version installation is managed.
2648 * Make ec_group_do_inverse_ord() more robust and available to other
2649 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2650 mitigations are applied to the fallback BN_mod_inverse().
2651 When using this function rather than BN_mod_inverse() directly, new
2652 EC cryptosystem implementations are then safer-by-default.
2656 * Add coordinate blinding for EC_POINT and implement projective
2657 coordinate blinding for generic prime curves as a countermeasure to
2658 chosen point SCA attacks.
2660 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2662 * Add blinding to ECDSA and DSA signatures to protect against side channel
2663 attacks discovered by Keegan Ryan (NCC Group).
2667 * Enforce checking in the `pkeyutl` command to ensure that the input
2668 length does not exceed the maximum supported digest length when performing
2669 a sign, verify or verifyrecover operation.
2673 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2674 I/O in combination with something like select() or poll() will hang. This
2675 can be turned off again using SSL_CTX_clear_mode().
2676 Many applications do not properly handle non-application data records, and
2677 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2678 around the problems in those applications, but can also break some.
2679 It's recommended to read the manpages about SSL_read(), SSL_write(),
2680 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2681 SSL_CTX_set_read_ahead() again.
2685 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2686 now allow empty (zero character) pass phrases.
2690 * Apply blinding to binary field modular inversion and remove patent
2691 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2695 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2696 binary and prime elliptic curves.
2700 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2701 constant time fixed point multiplication.
2705 * Revise elliptic curve scalar multiplication with timing attack
2706 defenses: ec_wNAF_mul redirects to a constant time implementation
2707 when computing fixed point and variable point multiplication (which
2708 in OpenSSL are mostly used with secret scalars in keygen, sign,
2709 ECDH derive operations).
2710 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2713 * Updated CONTRIBUTING
2717 * Updated DRBG / RAND to request nonce and additional low entropy
2718 randomness from the system.
2720 *Matthias St. Pierre*
2722 * Updated 'openssl rehash' to use OpenSSL consistent default.
2726 * Moved the load of the ssl_conf module to libcrypto, which helps
2727 loading engines that libssl uses before libssl is initialised.
2731 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2735 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2737 *Ingo Schwarze, Rich Salz*
2739 * Added output of accepting IP address and port for 'openssl s_server'
2743 * Added a new API for TLSv1.3 ciphersuites:
2744 SSL_CTX_set_ciphersuites()
2745 SSL_set_ciphersuites()
2749 * Memory allocation failures consistently add an error to the error
2754 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2755 in libcrypto when run as setuid/setgid.
2759 * Load any config file by default when libssl is used.
2763 * Added new public header file <openssl/rand_drbg.h> and documentation
2764 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2766 *Matthias St. Pierre*
2768 * QNX support removed (cannot find contributors to get their approval
2769 for the license change).
2773 * TLSv1.3 replay protection for early data has been implemented. See the
2774 SSL_read_early_data() man page for further details.
2778 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2779 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2780 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2781 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2782 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2783 configuration has been separated out. See the ciphers man page or the
2784 SSL_CTX_set_ciphersuites() man page for more information.
2788 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2789 in responder mode now supports the new "-multi" option, which
2790 spawns the specified number of child processes to handle OCSP
2791 requests. The "-timeout" option now also limits the OCSP
2792 responder's patience to wait to receive the full client request
2793 on a newly accepted connection. Child processes are respawned
2794 as needed, and the CA index file is automatically reloaded
2795 when changed. This makes it possible to run the "ocsp" responder
2796 as a long-running service, making the OpenSSL CA somewhat more
2797 feature-complete. In this mode, most diagnostic messages logged
2798 after entering the event loop are logged via syslog(3) rather than
2803 * Added support for X448 and Ed448. Heavily based on original work by
2808 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2809 objects loaded. This adds the functions OSSL_STORE_expect() and
2810 OSSL_STORE_find() as well as needed tools to construct searches and
2811 get the search data out of them.
2815 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2816 version of OpenSSL should review their configuration settings to ensure
2817 that they are still appropriate for TLSv1.3. For further information see:
2818 <https://wiki.openssl.org/index.php/TLS1.3>
2822 * Grand redesign of the OpenSSL random generator
2824 The default RAND method now utilizes an AES-CTR DRBG according to
2825 NIST standard SP 800-90Ar1. The new random generator is essentially
2826 a port of the default random generator from the OpenSSL FIPS 2.0
2827 object module. It is a hybrid deterministic random bit generator
2828 using an AES-CTR bit stream and which seeds and reseeds itself
2829 automatically using trusted system entropy sources.
2831 Some of its new features are:
2832 - Support for multiple DRBG instances with seed chaining.
2833 - The default RAND method makes use of a DRBG.
2834 - There is a public and private DRBG instance.
2835 - The DRBG instances are fork-safe.
2836 - Keep all global DRBG instances on the secure heap if it is enabled.
2837 - The public and private DRBG instance are per thread for lock free
2840 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2842 * Changed Configure so it only says what it does and doesn't dump
2843 so much data. Instead, ./configdata.pm should be used as a script
2844 to display all sorts of configuration data.
2848 * Added processing of "make variables" to Configure.
2852 * Added SHA512/224 and SHA512/256 algorithm support.
2856 * The last traces of Netware support, first removed in 1.1.0, have
2861 * Get rid of Makefile.shared, and in the process, make the processing
2862 of certain files (rc.obj, or the .def/.map/.opt files produced from
2863 the ordinal files) more visible and hopefully easier to trace and
2864 debug (or make silent).
2868 * Make it possible to have environment variable assignments as
2869 arguments to config / Configure.
2873 * Add multi-prime RSA (RFC 8017) support.
2877 * Add SM3 implemented according to GB/T 32905-2016
2878 *Jack Lloyd <jack.lloyd@ribose.com>,*
2879 *Ronald Tse <ronald.tse@ribose.com>,*
2880 *Erick Borsboom <erick.borsboom@ribose.com>*
2882 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2883 as documented in RFC6066.
2884 Based on a patch from Tomasz Moń
2886 *Filipe Raimundo da Silva*
2888 * Add SM4 implemented according to GB/T 32907-2016.
2889 *Jack Lloyd <jack.lloyd@ribose.com>,*
2890 *Ronald Tse <ronald.tse@ribose.com>,*
2891 *Erick Borsboom <erick.borsboom@ribose.com>*
2893 * Reimplement -newreq-nodes and ERR_error_string_n; the
2894 original author does not agree with the license change.
2898 * Add ARIA AEAD TLS support.
2902 * Some macro definitions to support VS6 have been removed. Visual
2903 Studio 6 has not worked since 1.1.0
2907 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2908 without clearing the errors.
2912 * Add "atfork" functions. If building on a system that without
2913 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2914 requirements. The RAND facility now uses/requires this.
2922 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2923 not possible to disable entirely. However, it's still possible to
2924 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2927 To disable, configure with 'no-ui-console'. 'no-ui' is still
2928 possible to use as an alias. Check at compile time with the
2929 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2930 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2934 * Add a STORE module, which implements a uniform and URI based reader of
2935 stores that can contain keys, certificates, CRLs and numerous other
2936 objects. The main API is loosely based on a few stdio functions,
2937 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2938 OSSL_STORE_error and OSSL_STORE_close.
2939 The implementation uses backends called "loaders" to implement arbitrary
2940 URI schemes. There is one built in "loader" for the 'file' scheme.
2944 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2945 then adjusted to work on FreeBSD 8.4 as well.
2946 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2947 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2951 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2952 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2953 error code calls like this:
2955 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2957 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2958 that can be encoded in C. For the foreseeable future, this will only
2961 *Richard Levitte and Tim Hudson*
2963 * Removed BSD cryptodev engine.
2967 * Add a build target 'build_all_generated', to build all generated files
2968 and only that. This can be used to prepare everything that requires
2969 things like perl for a system that lacks perl and then move everything
2970 to that system and do the rest of the build there.
2974 * In the UI interface, make it possible to duplicate the user data. This
2975 can be used by engines that need to retain the data for a longer time
2976 than just the call where this user data is passed.
2980 * Ignore the '-named_curve auto' value for compatibility of applications
2983 *Tomáš Mráz <tmraz@fedoraproject.org>*
2985 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2986 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2987 alerts across multiple records (some of which could be empty). In practice
2988 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2989 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2990 support this at all. Supporting it adds significant complexity to the
2991 record layer, and its removal is unlikely to cause interoperability
2996 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2997 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2998 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3003 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3004 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3006 *Richard Levitte, Andy Polyakov*
3008 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3013 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3014 platform rather than 'mingw'.
3018 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3019 success if they are asked to add an object which already exists
3020 in the store. This change cascades to other functions which load
3021 certificates and CRLs.
3025 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3026 facilitate stack unwinding even from assembly subroutines.
3030 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3031 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3035 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3036 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3037 which is the minimum version we support.
3041 * Certificate time validation (X509_cmp_time) enforces stricter
3042 compliance with RFC 5280. Fractional seconds and timezone offsets
3043 are no longer allowed.
3047 * Add support for ARIA
3051 * s_client will now send the Server Name Indication (SNI) extension by
3052 default unless the new "-noservername" option is used. The server name is
3053 based on the host provided to the "-connect" option unless overridden by
3054 using "-servername".
3058 * Add support for SipHash
3062 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3063 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3064 prevent issues where no progress is being made and the peer continually
3065 sends unrecognised record types, using up resources processing them.
3069 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3070 using the algorithm defined in
3071 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3075 * Heartbeat support has been removed; the ABI is changed for now.
3077 *Richard Levitte, Rich Salz*
3079 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3083 * The RSA "null" method, which was partially supported to avoid patent
3084 issues, has been replaced to always returns NULL.
3091 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3093 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3094 used even when parsing explicit parameters, when loading a encoded key
3095 or calling `EC_GROUP_new_from_ecpkparameters()`/
3096 `EC_GROUP_new_from_ecparameters()`.
3097 This prevents bypass of security hardening and performance gains,
3098 especially for curves with specialized EC_METHODs.
3099 By default, if a key encoded with explicit parameters is loaded and later
3100 encoded, the output is still encoded with explicit parameters, even if
3101 internally a "named" EC_GROUP is used for computation.
3105 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3106 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3107 NULL. After this change, only the cofactor parameter can be NULL. It also
3108 does some minimal sanity checks on the passed order.
3113 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3114 An attack is simple, if the first CMS_recipientInfo is valid but the
3115 second CMS_recipientInfo is chosen ciphertext. If the second
3116 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3117 encryption key will be replaced by garbage, and the message cannot be
3118 decoded, but if the RSA decryption fails, the correct encryption key is
3119 used and the recipient will not notice the attack.
3120 As a work around for this potential attack the length of the decrypted
3121 key must be equal to the cipher default key length, in case the
3122 certificate is not given and all recipientInfo are tried out.
3123 The old behaviour can be re-enabled in the CMS code by setting the
3124 CMS_DEBUG_DECRYPT flag.
3129 * Use Windows installation paths in the mingw builds
3131 Mingw isn't a POSIX environment per se, which means that Windows
3132 paths should be used for installation.
3137 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3139 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3140 This changes the size when using the `genpkey` command when no size is given.
3141 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3142 generation commands to use 2048 bits by default.
3146 * Prevent over long nonces in ChaCha20-Poly1305.
3148 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3149 for every encryption operation. RFC 7539 specifies that the nonce value
3150 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3151 and front pads the nonce with 0 bytes if it is less than 12
3152 bytes. However it also incorrectly allows a nonce to be set of up to 16
3153 bytes. In this case only the last 12 bytes are significant and any
3154 additional leading bytes are ignored.
3156 It is a requirement of using this cipher that nonce values are
3157 unique. Messages encrypted using a reused nonce value are susceptible to
3158 serious confidentiality and integrity attacks. If an application changes
3159 the default nonce length to be longer than 12 bytes and then makes a
3160 change to the leading bytes of the nonce expecting the new value to be a
3161 new unique nonce then such an application could inadvertently encrypt
3162 messages with a reused nonce.
3164 Additionally the ignored bytes in a long nonce are not covered by the
3165 integrity guarantee of this cipher. Any application that relies on the
3166 integrity of these ignored leading bytes of a long nonce may be further
3167 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3168 is safe because no such use sets such a long nonce value. However user
3169 applications that use this cipher directly and set a non-default nonce
3170 length to be longer than 12 bytes may be vulnerable.
3172 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3178 * Added SCA hardening for modular field inversion in EC_GROUP through
3179 a new dedicated field_inv() pointer in EC_METHOD.
3180 This also addresses a leakage affecting conversions from projective
3181 to affine coordinates.
3183 *Billy Bob Brumley, Nicola Tuveri*
3185 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3186 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3190 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3194 * Remove the 'dist' target and add a tarball building script. The
3195 'dist' target has fallen out of use, and it shouldn't be
3196 necessary to configure just to create a source distribution.
3200 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3202 * Timing vulnerability in DSA signature generation
3204 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3205 timing side channel attack. An attacker could use variations in the signing
3206 algorithm to recover the private key.
3208 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3213 * Timing vulnerability in ECDSA signature generation
3215 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3216 timing side channel attack. An attacker could use variations in the signing
3217 algorithm to recover the private key.
3219 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3224 * Add coordinate blinding for EC_POINT and implement projective
3225 coordinate blinding for generic prime curves as a countermeasure to
3226 chosen point SCA attacks.
3228 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3230 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3232 * Client DoS due to large DH parameter
3234 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3235 malicious server can send a very large prime value to the client. This will
3236 cause the client to spend an unreasonably long period of time generating a
3237 key for this prime resulting in a hang until the client has finished. This
3238 could be exploited in a Denial Of Service attack.
3240 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3245 * Cache timing vulnerability in RSA Key Generation
3247 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3248 a cache timing side channel attack. An attacker with sufficient access to
3249 mount cache timing attacks during the RSA key generation process could
3250 recover the private key.
3252 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3253 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3258 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3259 parameter is no longer accepted, as it leads to a corrupt table. NULL
3260 pem_str is reserved for alias entries only.
3264 * Revert blinding in ECDSA sign and instead make problematic addition
3265 length-invariant. Switch even to fixed-length Montgomery multiplication.
3269 * Change generating and checking of primes so that the error rate of not
3270 being prime depends on the intended use based on the size of the input.
3271 For larger primes this will result in more rounds of Miller-Rabin.
3272 The maximal error rate for primes with more than 1080 bits is lowered
3275 *Kurt Roeckx, Annie Yousar*
3277 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3281 * Add blinding to ECDSA and DSA signatures to protect against side channel
3282 attacks discovered by Keegan Ryan (NCC Group).
3286 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3287 now allow empty (zero character) pass phrases.
3291 * Certificate time validation (X509_cmp_time) enforces stricter
3292 compliance with RFC 5280. Fractional seconds and timezone offsets
3293 are no longer allowed.
3297 * Fixed a text canonicalisation bug in CMS
3299 Where a CMS detached signature is used with text content the text goes
3300 through a canonicalisation process first prior to signing or verifying a
3301 signature. This process strips trailing space at the end of lines, converts
3302 line terminators to CRLF and removes additional trailing line terminators
3303 at the end of a file. A bug in the canonicalisation process meant that
3304 some characters, such as form-feed, were incorrectly treated as whitespace
3305 and removed. This is contrary to the specification (RFC5485). This fix
3306 could mean that detached text data signed with an earlier version of
3307 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3308 signed with a fixed OpenSSL may fail to verify with an earlier version of
3309 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3310 and use the "-binary" flag (for the "cms" command line application) or set
3311 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3315 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3317 * Constructed ASN.1 types with a recursive definition could exceed the stack
3319 Constructed ASN.1 types with a recursive definition (such as can be found
3320 in PKCS7) could eventually exceed the stack given malicious input with
3321 excessive recursion. This could result in a Denial Of Service attack. There
3322 are no such structures used within SSL/TLS that come from untrusted sources
3323 so this is considered safe.
3325 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3331 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3333 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3334 effectively reduced to only comparing the least significant bit of each
3335 byte. This allows an attacker to forge messages that would be considered as
3336 authenticated in an amount of tries lower than that guaranteed by the
3337 security claims of the scheme. The module can only be compiled by the
3338 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3340 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3346 * Add a build target 'build_all_generated', to build all generated files
3347 and only that. This can be used to prepare everything that requires
3348 things like perl for a system that lacks perl and then move everything
3349 to that system and do the rest of the build there.
3353 * Backport SSL_OP_NO_RENGOTIATION
3355 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3356 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3357 changes this is no longer possible in 1.1.0. Therefore, the new
3358 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3359 1.1.0 to provide equivalent functionality.
3361 Note that if an application built against 1.1.0h headers (or above) is run
3362 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3363 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3367 * Removed the OS390-Unix config target. It relied on a script that doesn't
3372 * rsaz_1024_mul_avx2 overflow bug on x86_64
3374 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3375 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3376 Analysis suggests that attacks against RSA and DSA as a result of this
3377 defect would be very difficult to perform and are not believed likely.
3378 Attacks against DH1024 are considered just feasible, because most of the
3379 work necessary to deduce information about a private key may be performed
3380 offline. The amount of resources required for such an attack would be
3381 significant. However, for an attack on TLS to be meaningful, the server
3382 would have to share the DH1024 private key among multiple clients, which is
3383 no longer an option since CVE-2016-0701.
3385 This only affects processors that support the AVX2 but not ADX extensions
3386 like Intel Haswell (4th generation).
3388 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3389 was originally found via the OSS-Fuzz project.
3394 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3396 * bn_sqrx8x_internal carry bug on x86_64
3398 There is a carry propagating bug in the x86_64 Montgomery squaring
3399 procedure. No EC algorithms are affected. Analysis suggests that attacks
3400 against RSA and DSA as a result of this defect would be very difficult to
3401 perform and are not believed likely. Attacks against DH are considered just
3402 feasible (although very difficult) because most of the work necessary to
3403 deduce information about a private key may be performed offline. The amount
3404 of resources required for such an attack would be very significant and
3405 likely only accessible to a limited number of attackers. An attacker would
3406 additionally need online access to an unpatched system using the target
3407 private key in a scenario with persistent DH parameters and a private
3408 key that is shared between multiple clients.
3410 This only affects processors that support the BMI1, BMI2 and ADX extensions
3411 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3413 This issue was reported to OpenSSL by the OSS-Fuzz project.
3418 * Malformed X.509 IPAddressFamily could cause OOB read
3420 If an X.509 certificate has a malformed IPAddressFamily extension,
3421 OpenSSL could do a one-byte buffer overread. The most likely result
3422 would be an erroneous display of the certificate in text format.
3424 This issue was reported to OpenSSL by the OSS-Fuzz project.
3429 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3431 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3432 platform rather than 'mingw'.
3436 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3437 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3438 which is the minimum version we support.
3442 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3444 * Encrypt-Then-Mac renegotiation crash
3446 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3447 negotiated where it was not in the original handshake (or vice-versa) then
3448 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3449 and servers are affected.
3451 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3456 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3458 * Truncated packet could crash via OOB read
3460 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3461 cipher is being used, then a truncated packet can cause that host to
3462 perform an out-of-bounds read, usually resulting in a crash.
3464 This issue was reported to OpenSSL by Robert Święcki of Google.
3469 * Bad (EC)DHE parameters cause a client crash
3471 If a malicious server supplies bad parameters for a DHE or ECDHE key
3472 exchange then this can result in the client attempting to dereference a
3473 NULL pointer leading to a client crash. This could be exploited in a Denial
3476 This issue was reported to OpenSSL by Guido Vranken.
3481 * BN_mod_exp may produce incorrect results on x86_64
3483 There is a carry propagating bug in the x86_64 Montgomery squaring
3484 procedure. No EC algorithms are affected. Analysis suggests that attacks
3485 against RSA and DSA as a result of this defect would be very difficult to
3486 perform and are not believed likely. Attacks against DH are considered just
3487 feasible (although very difficult) because most of the work necessary to
3488 deduce information about a private key may be performed offline. The amount
3489 of resources required for such an attack would be very significant and
3490 likely only accessible to a limited number of attackers. An attacker would
3491 additionally need online access to an unpatched system using the target
3492 private key in a scenario with persistent DH parameters and a private
3493 key that is shared between multiple clients. For example this can occur by
3494 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3495 similar to CVE-2015-3193 but must be treated as a separate problem.
3497 This issue was reported to OpenSSL by the OSS-Fuzz project.
3502 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3504 * ChaCha20/Poly1305 heap-buffer-overflow
3506 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3507 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3508 crash. This issue is not considered to be exploitable beyond a DoS.
3510 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3515 * CMS Null dereference
3517 Applications parsing invalid CMS structures can crash with a NULL pointer
3518 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3519 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3520 structure callback if an attempt is made to free certain invalid encodings.
3521 Only CHOICE structures using a callback which do not handle NULL value are
3524 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3529 * Montgomery multiplication may produce incorrect results
3531 There is a carry propagating bug in the Broadwell-specific Montgomery
3532 multiplication procedure that handles input lengths divisible by, but
3533 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3534 and DH private keys are impossible. This is because the subroutine in
3535 question is not used in operations with the private key itself and an input
3536 of the attacker's direct choice. Otherwise the bug can manifest itself as
3537 transient authentication and key negotiation failures or reproducible
3538 erroneous outcome of public-key operations with specially crafted input.
3539 Among EC algorithms only Brainpool P-512 curves are affected and one
3540 presumably can attack ECDH key negotiation. Impact was not analyzed in
3541 detail, because pre-requisites for attack are considered unlikely. Namely
3542 multiple clients have to choose the curve in question and the server has to
3543 share the private key among them, neither of which is default behaviour.
3544 Even then only clients that chose the curve will be affected.
3546 This issue was publicly reported as transient failures and was not
3547 initially recognized as a security issue. Thanks to Richard Morgan for
3548 providing reproducible case.
3553 * Removed automatic addition of RPATH in shared libraries and executables,
3554 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3558 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3560 * Fix Use After Free for large message sizes
3562 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3563 message larger than approx 16k is received then the underlying buffer to
3564 store the incoming message is reallocated and moved. Unfortunately a
3565 dangling pointer to the old location is left which results in an attempt to
3566 write to the previously freed location. This is likely to result in a
3567 crash, however it could potentially lead to execution of arbitrary code.
3569 This issue only affects OpenSSL 1.1.0a.
3571 This issue was reported to OpenSSL by Robert Święcki.
3576 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3578 * OCSP Status Request extension unbounded memory growth
3580 A malicious client can send an excessively large OCSP Status Request
3581 extension. If that client continually requests renegotiation, sending a
3582 large OCSP Status Request extension each time, then there will be unbounded
3583 memory growth on the server. This will eventually lead to a Denial Of
3584 Service attack through memory exhaustion. Servers with a default
3585 configuration are vulnerable even if they do not support OCSP. Builds using
3586 the "no-ocsp" build time option are not affected.
3588 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3593 * SSL_peek() hang on empty record
3595 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3596 sends an empty record. This could be exploited by a malicious peer in a
3597 Denial Of Service attack.
3599 This issue was reported to OpenSSL by Alex Gaynor.
3604 * Excessive allocation of memory in tls_get_message_header() and
3605 dtls1_preprocess_fragment()
3607 A (D)TLS message includes 3 bytes for its length in the header for the
3608 message. This would allow for messages up to 16Mb in length. Messages of
3609 this length are excessive and OpenSSL includes a check to ensure that a
3610 peer is sending reasonably sized messages in order to avoid too much memory
3611 being consumed to service a connection. A flaw in the logic of version
3612 1.1.0 means that memory for the message is allocated too early, prior to
3613 the excessive message length check. Due to way memory is allocated in
3614 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3615 to service a connection. This could lead to a Denial of Service through
3616 memory exhaustion. However, the excessive message length check still takes
3617 place, and this would cause the connection to immediately fail. Assuming
3618 that the application calls SSL_free() on the failed connection in a timely
3619 manner then the 21Mb of allocated memory will then be immediately freed
3620 again. Therefore, the excessive memory allocation will be transitory in
3621 nature. This then means that there is only a security impact if:
3623 1) The application does not call SSL_free() in a timely manner in the event
3624 that the connection fails
3626 2) The application is working in a constrained environment where there is
3627 very little free memory
3629 3) The attacker initiates multiple connection attempts such that there are
3630 multiple connections in a state where memory has been allocated for the
3631 connection; SSL_free() has not yet been called; and there is insufficient
3632 memory to service the multiple requests.
3634 Except in the instance of (1) above any Denial Of Service is likely to be
3635 transitory because as soon as the connection fails the memory is
3636 subsequently freed again in the SSL_free() call. However there is an
3637 increased risk during this period of application crashes due to the lack of
3638 memory - which would then mean a more serious Denial of Service.
3640 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3641 (CVE-2016-6307 and CVE-2016-6308)
3645 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3646 had to be removed. Primary reason is that vendor assembler can't
3647 assemble our modules with -KPIC flag. As result it, assembly
3648 support, was not even available as option. But its lack means
3649 lack of side-channel resistant code, which is incompatible with
3650 security by todays standards. Fortunately gcc is readily available
3651 prepackaged option, which we firmly point at...
3655 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3657 * Windows command-line tool supports UTF-8 opt-in option for arguments
3658 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3659 (to any value) allows Windows user to access PKCS#12 file generated
3660 with Windows CryptoAPI and protected with non-ASCII password, as well
3661 as files generated under UTF-8 locale on Linux also protected with
3666 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3667 have been disabled by default and removed from DEFAULT, just like RC4.
3668 See the RC4 item below to re-enable both.
3672 * The method for finding the storage location for the Windows RAND seed file
3673 has changed. First we check %RANDFILE%. If that is not set then we check
3674 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3675 all else fails we fall back to C:\.
3679 * The EVP_EncryptUpdate() function has had its return type changed from void
3680 to int. A return of 0 indicates and error while a return of 1 indicates
3685 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3686 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3687 off the constant time implementation for RSA, DSA and DH have been made
3688 no-ops and deprecated.
3692 * Windows RAND implementation was simplified to only get entropy by
3693 calling CryptGenRandom(). Various other RAND-related tickets
3696 *Joseph Wylie Yandle, Rich Salz*
3698 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3699 and `OPENSSL_LH_`, respectively. The old names are available
3700 with API compatibility. They new names are now completely documented.
3704 * Unify TYPE_up_ref(obj) methods signature.
3705 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3706 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3707 int (instead of void) like all others TYPE_up_ref() methods.
3708 So now these methods also check the return value of CRYPTO_atomic_add(),
3709 and the validity of object reference counter.
3711 *fdasilvayy@gmail.com*
3713 * With Windows Visual Studio builds, the .pdb files are installed
3714 alongside the installed libraries and executables. For a static
3715 library installation, ossl_static.pdb is the associate compiler
3716 generated .pdb file to be used when linking programs.
3720 * Remove openssl.spec. Packaging files belong with the packagers.
3724 * Automatic Darwin/OSX configuration has had a refresh, it will now
3725 recognise x86_64 architectures automatically. You can still decide
3726 to build for a different bitness with the environment variable
3727 KERNEL_BITS (can be 32 or 64), for example:
3729 KERNEL_BITS=32 ./config
3733 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3734 256 bit AES and HMAC with SHA256.
3738 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3742 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3746 * To enable users to have their own config files and build file templates,
3747 Configure looks in the directory indicated by the environment variable
3748 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3749 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3750 name and is used as is.
3754 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3755 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3756 X509_CERT_FILE_CTX was removed.
3760 * "shared" builds are now the default. To create only static libraries use
3761 the "no-shared" Configure option.
3765 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3766 All of these option have not worked for some while and are fundamental
3771 * Make various cleanup routines no-ops and mark them as deprecated. Most
3772 global cleanup functions are no longer required because they are handled
3773 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3774 Explicitly de-initing can cause problems (e.g. where a library that uses
3775 OpenSSL de-inits, but an application is still using it). The affected
3776 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3777 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3778 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3779 COMP_zlib_cleanup().
3783 * --strict-warnings no longer enables runtime debugging options
3784 such as REF_DEBUG. Instead, debug options are automatically
3785 enabled with '--debug' builds.
3787 *Andy Polyakov, Emilia Käsper*
3789 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3790 have been moved out of the public header files. New functions for managing
3791 these have been added.
3795 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3796 objects have been moved out of the public header files. New
3797 functions for managing these have been added.
3801 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3802 have been moved out of the public header files. New functions for managing
3803 these have been added.
3807 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3808 moved out of the public header files. New functions for managing these
3813 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3817 * Removed the mk1mf build scripts.
3821 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3822 it is always safe to #include a header now.
3826 * Removed the aged BC-32 config and all its supporting scripts
3830 * Removed support for Ultrix, Netware, and OS/2.
3834 * Add support for HKDF.
3836 *Alessandro Ghedini*
3838 * Add support for blake2b and blake2s
3842 * Added support for "pipelining". Ciphers that have the
3843 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3844 encryptions/decryptions simultaneously. There are currently no built-in
3845 ciphers with this property but the expectation is that engines will be able
3846 to offer it to significantly improve throughput. Support has been extended
3847 into libssl so that multiple records for a single connection can be
3848 processed in one go (for >=TLS 1.1).
3852 * Added the AFALG engine. This is an async capable engine which is able to
3853 offload work to the Linux kernel. In this initial version it only supports
3854 AES128-CBC. The kernel must be version 4.1.0 or greater.
3858 * OpenSSL now uses a new threading API. It is no longer necessary to
3859 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3860 are two supported threading models: pthreads and windows threads. It is
3861 also possible to configure OpenSSL at compile time for "no-threads". The
3862 old threading API should no longer be used. The functions have been
3863 replaced with "no-op" compatibility macros.
3865 *Alessandro Ghedini, Matt Caswell*
3867 * Modify behavior of ALPN to invoke callback after SNI/servername
3868 callback, such that updates to the SSL_CTX affect ALPN.
3872 * Add SSL_CIPHER queries for authentication and key-exchange.
3876 * Changes to the DEFAULT cipherlist:
3877 - Prefer (EC)DHE handshakes over plain RSA.
3878 - Prefer AEAD ciphers over legacy ciphers.
3879 - Prefer ECDSA over RSA when both certificates are available.
3880 - Prefer TLSv1.2 ciphers/PRF.
3881 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3886 * Change the ECC default curve list to be this, in order: x25519,
3887 secp256r1, secp521r1, secp384r1.
3891 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3892 disabled by default. They can be re-enabled using the
3893 enable-weak-ssl-ciphers option to Configure.
3897 * If the server has ALPN configured, but supports no protocols that the
3898 client advertises, send a fatal "no_application_protocol" alert.
3899 This behaviour is SHALL in RFC 7301, though it isn't universally
3900 implemented by other servers.
3904 * Add X25519 support.
3905 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3906 for public and private key encoding using the format documented in
3907 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3908 key generation and key derivation.
3910 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3915 * Deprecate SRP_VBASE_get_by_user.
3916 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3917 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3918 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3919 seed, even if the seed is configured.
3921 Users should use SRP_VBASE_get1_by_user instead. Note that in
3922 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3923 also that even though configuring the SRP seed attempts to hide
3924 invalid usernames by continuing the handshake with fake
3925 credentials, this behaviour is not constant time and no strong
3926 guarantees are made that the handshake is indistinguishable from
3927 that of a valid user.
3931 * Configuration change; it's now possible to build dynamic engines
3932 without having to build shared libraries and vice versa. This
3933 only applies to the engines in `engines/`, those in `crypto/engine/`
3934 will always be built into libcrypto (i.e. "static").
3936 Building dynamic engines is enabled by default; to disable, use
3937 the configuration option "disable-dynamic-engine".
3939 The only requirements for building dynamic engines are the
3940 presence of the DSO module and building with position independent
3941 code, so they will also automatically be disabled if configuring
3942 with "disable-dso" or "disable-pic".
3944 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3945 are also taken away from openssl/opensslconf.h, as they are
3950 * Configuration change; if there is a known flag to compile
3951 position independent code, it will always be applied on the
3952 libcrypto and libssl object files, and never on the application
3953 object files. This means other libraries that use routines from
3954 libcrypto / libssl can be made into shared libraries regardless
3955 of how OpenSSL was configured.
3957 If this isn't desirable, the configuration options "disable-pic"
3958 or "no-pic" can be used to disable the use of PIC. This will
3959 also disable building shared libraries and dynamic engines.
3963 * Removed JPAKE code. It was experimental and has no wide use.
3967 * The INSTALL_PREFIX Makefile variable has been renamed to
3968 DESTDIR. That makes for less confusion on what this variable
3969 is for. Also, the configuration option --install_prefix is
3974 * Heartbeat for TLS has been removed and is disabled by default
3975 for DTLS; configure with enable-heartbeats. Code that uses the
3976 old #define's might need to be updated.
3978 *Emilia Käsper, Rich Salz*
3980 * Rename REF_CHECK to REF_DEBUG.
3984 * New "unified" build system
3986 The "unified" build system is aimed to be a common system for all
3987 platforms we support. With it comes new support for VMS.
3989 This system builds supports building in a different directory tree
3990 than the source tree. It produces one Makefile (for unix family
3991 or lookalikes), or one descrip.mms (for VMS).
3993 The source of information to make the Makefile / descrip.mms is
3994 small files called 'build.info', holding the necessary
3995 information for each directory with source to compile, and a
3996 template in Configurations, like unix-Makefile.tmpl or
3999 With this change, the library names were also renamed on Windows
4000 and on VMS. They now have names that are closer to the standard
4001 on Unix, and include the major version number, and in certain
4002 cases, the architecture they are built for. See "Notes on shared
4003 libraries" in INSTALL.
4005 We rely heavily on the perl module Text::Template.
4009 * Added support for auto-initialisation and de-initialisation of the library.
4010 OpenSSL no longer requires explicit init or deinit routines to be called,
4011 except in certain circumstances. See the OPENSSL_init_crypto() and
4012 OPENSSL_init_ssl() man pages for further information.
4016 * The arguments to the DTLSv1_listen function have changed. Specifically the
4017 "peer" argument is now expected to be a BIO_ADDR object.
4019 * Rewrite of BIO networking library. The BIO library lacked consistent
4020 support of IPv6, and adding it required some more extensive
4021 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4022 which hold all types of addresses and chains of address information.
4023 It also introduces a new API, with functions like BIO_socket,
4024 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4025 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4026 have been adapted accordingly.
4030 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4035 * CRIME protection: disable compression by default, even if OpenSSL is
4036 compiled with zlib enabled. Applications can still enable compression
4037 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4038 using the SSL_CONF library to configure compression.
4042 * The signature of the session callback configured with
4043 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4044 was explicitly marked as `const unsigned char*` instead of
4049 * Always DPURIFY. Remove the use of uninitialized memory in the
4050 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4054 * Removed many obsolete configuration items, including
4055 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4056 MD2_CHAR, MD2_INT, MD2_LONG
4058 IDEA_SHORT, IDEA_LONG
4059 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4061 *Rich Salz, with advice from Andy Polyakov*
4063 * Many BN internals have been moved to an internal header file.
4065 *Rich Salz with help from Andy Polyakov*
4067 * Configuration and writing out the results from it has changed.
4068 Files such as Makefile include/openssl/opensslconf.h and are now
4069 produced through general templates, such as Makefile.in and
4070 crypto/opensslconf.h.in and some help from the perl module
4073 Also, the center of configuration information is no longer
4074 Makefile. Instead, Configure produces a perl module in
4075 configdata.pm which holds most of the config data (in the hash
4076 table %config), the target data that comes from the target
4077 configuration in one of the `Configurations/*.conf` files (in
4082 * To clarify their intended purposes, the Configure options
4083 --prefix and --openssldir change their semantics, and become more
4084 straightforward and less interdependent.
4086 --prefix shall be used exclusively to give the location INSTALLTOP
4087 where programs, scripts, libraries, include files and manuals are
4088 going to be installed. The default is now /usr/local.
4090 --openssldir shall be used exclusively to give the default
4091 location OPENSSLDIR where certificates, private keys, CRLs are
4092 managed. This is also where the default openssl.cnf gets
4094 If the directory given with this option is a relative path, the
4095 values of both the --prefix value and the --openssldir value will
4096 be combined to become OPENSSLDIR.
4097 The default for --openssldir is INSTALLTOP/ssl.
4099 Anyone who uses --openssldir to specify where OpenSSL is to be
4100 installed MUST change to use --prefix instead.
4104 * The GOST engine was out of date and therefore it has been removed. An up
4105 to date GOST engine is now being maintained in an external repository.
4106 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4107 support for GOST ciphersuites (these are only activated if a GOST engine
4112 * EGD is no longer supported by default; use enable-egd when
4115 *Ben Kaduk and Rich Salz*
4117 * The distribution now has Makefile.in files, which are used to
4118 create Makefile's when Configure is run. *Configure must be run
4119 before trying to build now.*
4123 * The return value for SSL_CIPHER_description() for error conditions
4128 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4130 Obtaining and performing DNSSEC validation of TLSA records is
4131 the application's responsibility. The application provides
4132 the TLSA records of its choice to OpenSSL, and these are then
4133 used to authenticate the peer.
4135 The TLSA records need not even come from DNS. They can, for
4136 example, be used to implement local end-entity certificate or
4137 trust-anchor "pinning", where the "pin" data takes the form
4138 of TLSA records, which can augment or replace verification
4139 based on the usual WebPKI public certification authorities.
4143 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4144 continues to support deprecated interfaces in default builds.
4145 However, applications are strongly advised to compile their
4146 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4147 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4148 or the 1.1.0 releases.
4150 In environments in which all applications have been ported to
4151 not use any deprecated interfaces OpenSSL's Configure script
4152 should be used with the --api=1.1.0 option to entirely remove
4153 support for the deprecated features from the library and
4154 unconditionally disable them in the installed headers.
4155 Essentially the same effect can be achieved with the "no-deprecated"
4156 argument to Configure, except that this will always restrict
4157 the build to just the latest API, rather than a fixed API
4160 As applications are ported to future revisions of the API,
4161 they should update their compile-time OPENSSL_API_COMPAT define
4162 accordingly, but in most cases should be able to continue to
4163 compile with later releases.
4165 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4166 0x10000000L and 0x00908000L, respectively. However those
4167 versions did not support the OPENSSL_API_COMPAT feature, and
4168 so applications are not typically tested for explicit support
4169 of just the undeprecated features of either release.
4173 * Add support for setting the minimum and maximum supported protocol.
4174 It can bet set via the SSL_set_min_proto_version() and
4175 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4176 MaxProtocol. It's recommended to use the new APIs to disable
4177 protocols instead of disabling individual protocols using
4178 SSL_set_options() or SSL_CONF's Protocol. This change also
4179 removes support for disabling TLS 1.2 in the OpenSSL TLS
4180 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4184 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4188 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4189 and integrates ECDSA and ECDH functionality into EC. Implementations can
4190 now redirect key generation and no longer need to convert to or from
4193 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4194 include the ec.h header file instead.
4198 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4199 ciphers who are no longer supported and drops support the ephemeral RSA key
4200 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4204 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4205 opaque. For HMAC_CTX, the following constructors and destructors
4208 HMAC_CTX *HMAC_CTX_new(void);
4209 void HMAC_CTX_free(HMAC_CTX *ctx);
4211 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4212 destroy such methods has been added. See EVP_MD_meth_new(3) and
4213 EVP_CIPHER_meth_new(3) for documentation.
4216 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4217 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4218 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4219 an already created structure.
4220 2) For consistency with the majority of our object creators and
4221 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4222 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4223 for deprecated builds.
4227 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4228 cryptographic operations to be performed asynchronously as long as an
4229 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4230 further details. Libssl has also had this capability integrated with the
4231 introduction of the new mode SSL_MODE_ASYNC and associated error
4232 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4233 pages. This work was developed in partnership with Intel Corp.
4237 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4238 always enabled now. If you want to disable the support you should
4239 exclude it using the list of supported ciphers. This also means that the
4240 "-no_ecdhe" option has been removed from s_server.
4244 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4245 SSL_{CTX_}set1_curves() which can set a list.
4249 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4250 curve you want to support using SSL_{CTX_}set1_curves().
4254 * State machine rewrite. The state machine code has been significantly
4255 refactored in order to remove much duplication of code and solve issues
4256 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4257 further details). This change does have some associated API changes.
4258 Notably the SSL_state() function has been removed and replaced by
4259 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4260 SSL_set_state() has been removed altogether. The previous handshake states
4261 defined in ssl.h and ssl3.h have also been removed.
4265 * All instances of the string "ssleay" in the public API were replaced
4266 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4267 Some error codes related to internal RSA_eay API's were renamed.
4271 * The demo files in crypto/threads were moved to demo/threads.
4275 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4278 *Matt Caswell, Rich Salz*
4280 * New ASN.1 embed macro.
4282 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4283 structure is not allocated: it is part of the parent. That is instead of
4291 This reduces memory fragmentation and make it impossible to accidentally
4292 set a mandatory field to NULL.
4294 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4295 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4296 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4301 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4305 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4306 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4307 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4308 DES and RC4 ciphersuites.
4312 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4313 This changes the decoding behaviour for some invalid messages,
4314 though the change is mostly in the more lenient direction, and
4315 legacy behaviour is preserved as much as possible.
4319 * Fix no-stdio build.
4320 *David Woodhouse <David.Woodhouse@intel.com> and also*
4321 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4323 * New testing framework
4324 The testing framework has been largely rewritten and is now using
4325 perl and the perl modules Test::Harness and an extended variant of
4326 Test::More called OpenSSL::Test to do its work. All test scripts in
4327 test/ have been rewritten into test recipes, and all direct calls to
4328 executables in test/Makefile have become individual recipes using the
4329 simplified testing OpenSSL::Test::Simple.
4331 For documentation on our testing modules, do:
4333 perldoc test/testlib/OpenSSL/Test/Simple.pm
4334 perldoc test/testlib/OpenSSL/Test.pm
4338 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4339 are used; the latter aborts on memory leaks (usually checked on exit).
4340 Some undocumented "set malloc, etc., hooks" functions were removed
4341 and others were changed. All are now documented.
4345 * In DSA_generate_parameters_ex, if the provided seed is too short,
4348 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4350 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4351 from RFC4279, RFC4785, RFC5487, RFC5489.
4353 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4354 original RSA_PSK patch.
4358 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4359 era flag was never set throughout the codebase (only read). Also removed
4360 SSL3_FLAGS_POP_BUFFER which was only used if
4361 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4365 * Changed the default name options in the "ca", "crl", "req" and "x509"
4366 to be "oneline" instead of "compat".
4370 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4371 not aware of clients that still exhibit this bug, and the workaround
4372 hasn't been working properly for a while.
4376 * The return type of BIO_number_read() and BIO_number_written() as well as
4377 the corresponding num_read and num_write members in the BIO structure has
4378 changed from unsigned long to uint64_t. On platforms where an unsigned
4379 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4384 * Given the pervasive nature of TLS extensions it is inadvisable to run
4385 OpenSSL without support for them. It also means that maintaining
4386 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4387 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4391 * Removed support for the two export grade static DH ciphersuites
4392 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4393 were newly added (along with a number of other static DH ciphersuites) to
4394 1.0.2. However the two export ones have *never* worked since they were
4395 introduced. It seems strange in any case to be adding new export
4396 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4400 * Version negotiation has been rewritten. In particular SSLv23_method(),
4401 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4402 and turned into macros which simply call the new preferred function names
4403 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4404 should use the new names instead. Also as part of this change the ssl23.h
4405 header file has been removed.
4409 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4410 code and the associated standard is no longer considered fit-for-purpose.
4414 * RT2547 was closed. When generating a private key, try to make the
4415 output file readable only by the owner. This behavior change might
4416 be noticeable when interacting with other software.
4418 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4423 * Added HTTP GET support to the ocsp command.
4427 * Changed default digest for the dgst and enc commands from MD5 to
4432 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4436 * Added support for TLS extended master secret from
4437 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4438 initial patch which was a great help during development.
4442 * All libssl internal structures have been removed from the public header
4443 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4444 now redundant). Users should not attempt to access internal structures
4445 directly. Instead they should use the provided API functions.
4449 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4450 Access to deprecated functions can be re-enabled by running config with
4451 "enable-deprecated". In addition applications wishing to use deprecated
4452 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4453 will, by default, disable some transitive includes that previously existed
4454 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4458 * Added support for OCB mode. OpenSSL has been granted a patent license
4459 compatible with the OpenSSL license for use of OCB. Details are available
4460 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4461 for OCB can be removed by calling config with no-ocb.
4465 * SSLv2 support has been removed. It still supports receiving an SSLv2
4466 compatible client hello.
4470 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4471 done while fixing the error code for the key-too-small case.
4473 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4475 * CA.sh has been removed; use CA.pl instead.
4479 * Removed old DES API.
4483 * Remove various unsupported platforms:
4489 Sinix/ReliantUNIX RM400
4494 16-bit platforms such as WIN16
4498 * Clean up OPENSSL_NO_xxx #define's
4499 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4500 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4501 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4502 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4503 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4504 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4505 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4506 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4507 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4508 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4512 * Cleaned up dead code
4513 Remove all but one '#ifdef undef' which is to be looked at.
4517 * Clean up calling of xxx_free routines.
4518 Just like free(), fix most of the xxx_free routines to accept
4519 NULL. Remove the non-null checks from callers. Save much code.
4523 * Add secure heap for storage of private keys (when possible).
4524 Add BIO_s_secmem(), CBIGNUM, etc.
4525 Contributed by Akamai Technologies under our Corporate CLA.
4529 * Experimental support for a new, fast, unbiased prime candidate generator,
4530 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4532 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4534 * New output format NSS in the sess_id command line tool. This allows
4535 exporting the session id and the master key in NSS keylog format.
4537 *Martin Kaiser <martin@kaiser.cx>*
4539 * Harmonize version and its documentation. -f flag is used to display
4542 *mancha <mancha1@zoho.com>*
4544 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4545 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4547 *mancha <mancha1@zoho.com>*
4549 * Fix some double frees. These are not thought to be exploitable.
4551 *mancha <mancha1@zoho.com>*
4553 * A missing bounds check in the handling of the TLS heartbeat extension
4554 can be used to reveal up to 64k of memory to a connected client or
4557 Thanks for Neel Mehta of Google Security for discovering this bug and to
4558 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4559 preparing the fix ([CVE-2014-0160])
4561 *Adam Langley, Bodo Moeller*
4563 * Fix for the attack described in the paper "Recovering OpenSSL
4564 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4565 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4566 <http://eprint.iacr.org/2014/140>
4568 Thanks to Yuval Yarom and Naomi Benger for discovering this
4569 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4571 *Yuval Yarom and Naomi Benger*
4573 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4574 this fixes a limitation in previous versions of OpenSSL.
4578 * Experimental encrypt-then-mac support.
4580 Experimental support for encrypt then mac from
4581 draft-gutmann-tls-encrypt-then-mac-02.txt
4583 To enable it set the appropriate extension number (0x42 for the test
4584 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4586 For non-compliant peers (i.e. just about everything) this should have no
4589 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4593 * Add EVP support for key wrapping algorithms, to avoid problems with
4594 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4595 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4596 algorithms and include tests cases.
4600 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4605 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4606 MGF1 digest and OAEP label.
4610 * Make openssl verify return errors.
4612 *Chris Palmer <palmer@google.com> and Ben Laurie*
4614 * New function ASN1_TIME_diff to calculate the difference between two
4615 ASN1_TIME structures or one structure and the current time.
4619 * Update fips_test_suite to support multiple command line options. New
4620 test to induce all self test errors in sequence and check expected
4625 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4626 sign or verify all in one operation.
4630 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4631 test programs and fips_test_suite. Includes functionality to parse
4632 the minimal script output of fipsalgest.pl directly.
4636 * Add authorisation parameter to FIPS_module_mode_set().
4640 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4644 * Use separate DRBG fields for internal and external flags. New function
4645 FIPS_drbg_health_check() to perform on demand health checking. Add
4646 generation tests to fips_test_suite with reduced health check interval to
4647 demonstrate periodic health checking. Add "nodh" option to
4648 fips_test_suite to skip very slow DH test.
4652 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4657 * More extensive health check for DRBG checking many more failure modes.
4658 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4659 combination: call this in fips_test_suite.
4663 * Add support for canonical generation of DSA parameter 'g'. See
4666 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4667 POST to handle HMAC cases.
4671 * Add functions FIPS_module_version() and FIPS_module_version_text()
4672 to return numerical and string versions of the FIPS module number.
4676 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4677 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4678 outside the validated module in the FIPS capable OpenSSL.
4682 * Minor change to DRBG entropy callback semantics. In some cases
4683 there is no multiple of the block length between min_len and
4684 max_len. Allow the callback to return more than max_len bytes
4685 of entropy but discard any extra: it is the callback's responsibility
4686 to ensure that the extra data discarded does not impact the
4687 requested amount of entropy.
4691 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4692 information in FIPS186-3, SP800-57 and SP800-131A.
4696 * CCM support via EVP. Interface is very similar to GCM case except we
4697 must supply all data in one chunk (i.e. no update, final) and the
4698 message length must be supplied if AAD is used. Add algorithm test
4703 * Initial version of POST overhaul. Add POST callback to allow the status
4704 of POST to be monitored and/or failures induced. Modify fips_test_suite
4705 to use callback. Always run all selftests even if one fails.
4709 * XTS support including algorithm test driver in the fips_gcmtest program.
4710 Note: this does increase the maximum key length from 32 to 64 bytes but
4711 there should be no binary compatibility issues as existing applications
4712 will never use XTS mode.
4716 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4717 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4718 performs algorithm blocking for unapproved PRNG types. Also do not
4719 set PRNG type in FIPS_mode_set(): leave this to the application.
4720 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4721 the standard OpenSSL PRNG: set additional data to a date time vector.
4725 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4726 This shouldn't present any incompatibility problems because applications
4727 shouldn't be using these directly and any that are will need to rethink
4728 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4732 * Extensive self tests and health checking required by SP800-90 DRBG.
4733 Remove strength parameter from FIPS_drbg_instantiate and always
4734 instantiate at maximum supported strength.
4738 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4742 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4746 * New function DH_compute_key_padded() to compute a DH key and pad with
4747 leading zeroes if needed: this complies with SP800-56A et al.
4751 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4752 anything, incomplete, subject to change and largely untested at present.
4756 * Modify fipscanisteronly build option to only build the necessary object
4757 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4761 * Add experimental option FIPSSYMS to give all symbols in
4762 fipscanister.o and FIPS or fips prefix. This will avoid
4763 conflicts with future versions of OpenSSL. Add perl script
4764 util/fipsas.pl to preprocess assembly language source files
4765 and rename any affected symbols.
4769 * Add selftest checks and algorithm block of non-fips algorithms in
4770 FIPS mode. Remove DES2 from selftests.
4774 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4775 return internal method without any ENGINE dependencies. Add new
4776 tiny fips sign and verify functions.
4780 * New build option no-ec2m to disable characteristic 2 code.
4784 * New build option "fipscanisteronly". This only builds fipscanister.o
4785 and (currently) associated fips utilities. Uses the file Makefile.fips
4786 instead of Makefile.org as the prototype.
4790 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4791 Update fips_gcmtest to use IV generator.
4795 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4796 setting output buffer to NULL. The `*Final` function must be
4797 called although it will not retrieve any additional data. The tag
4798 can be set or retrieved with a ctrl. The IV length is by default 12
4799 bytes (96 bits) but can be set to an alternative value. If the IV
4800 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4805 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4806 underlying do_cipher function handles all cipher semantics itself
4807 including padding and finalisation. This is useful if (for example)
4808 an ENGINE cipher handles block padding itself. The behaviour of
4809 do_cipher is subtly changed if this flag is set: the return value
4810 is the number of characters written to the output buffer (zero is
4811 no longer an error code) or a negative error code. Also if the
4812 input buffer is NULL and length 0 finalisation should be performed.
4816 * If a candidate issuer certificate is already part of the constructed
4817 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4821 * Improve forward-security support: add functions
4823 void SSL_CTX_set_not_resumable_session_callback(
4824 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4825 void SSL_set_not_resumable_session_callback(
4826 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4828 for use by SSL/TLS servers; the callback function will be called whenever a
4829 new session is created, and gets to decide whether the session may be
4830 cached to make it resumable (return 0) or not (return 1). (As by the
4831 SSL/TLS protocol specifications, the session_id sent by the server will be
4832 empty to indicate that the session is not resumable; also, the server will
4833 not generate RFC 4507 (RFC 5077) session tickets.)
4835 A simple reasonable callback implementation is to return is_forward_secure.
4836 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4837 by the SSL/TLS server library, indicating whether it can provide forward
4840 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4842 * New -verify_name option in command line utilities to set verification
4847 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4848 Add CMAC pkey methods.
4852 * Experimental renegotiation in s_server -www mode. If the client
4853 browses /reneg connection is renegotiated. If /renegcert it is
4854 renegotiated requesting a certificate.
4858 * Add an "external" session cache for debugging purposes to s_server. This
4859 should help trace issues which normally are only apparent in deployed
4860 multi-process servers.
4864 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4865 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4866 BIO_set_cipher() and some obscure PEM functions were changed so they
4867 can now return an error. The RAND changes required a change to the
4868 RAND_METHOD structure.
4872 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4873 a gcc attribute to warn if the result of a function is ignored. This
4874 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4875 whose return value is often ignored.
4879 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4880 These allow SCTs (signed certificate timestamps) to be requested and
4881 validated when establishing a connection.
4883 *Rob Percival <robpercival@google.com>*
4888 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4890 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4891 used even when parsing explicit parameters, when loading a encoded key
4892 or calling `EC_GROUP_new_from_ecpkparameters()`/
4893 `EC_GROUP_new_from_ecparameters()`.
4894 This prevents bypass of security hardening and performance gains,
4895 especially for curves with specialized EC_METHODs.
4896 By default, if a key encoded with explicit parameters is loaded and later
4897 encoded, the output is still encoded with explicit parameters, even if
4898 internally a "named" EC_GROUP is used for computation.
4902 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4903 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4904 NULL. After this change, only the cofactor parameter can be NULL. It also
4905 does some minimal sanity checks on the passed order.
4910 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4911 An attack is simple, if the first CMS_recipientInfo is valid but the
4912 second CMS_recipientInfo is chosen ciphertext. If the second
4913 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4914 encryption key will be replaced by garbage, and the message cannot be
4915 decoded, but if the RSA decryption fails, the correct encryption key is
4916 used and the recipient will not notice the attack.
4917 As a work around for this potential attack the length of the decrypted
4918 key must be equal to the cipher default key length, in case the
4919 certificate is not given and all recipientInfo are tried out.
4920 The old behaviour can be re-enabled in the CMS code by setting the
4921 CMS_DEBUG_DECRYPT flag.
4926 * Document issue with installation paths in diverse Windows builds
4928 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4929 binaries and run-time config file.
4934 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4936 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4937 This changes the size when using the `genpkey` command when no size is given.
4938 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4939 generation commands to use 2048 bits by default.
4943 * Add FIPS support for Android Arm 64-bit
4945 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4946 Module in Version 2.0.10. For some reason, the corresponding target
4947 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4948 built with FIPS support on Android Arm 64-bit. This omission has been
4951 *Matthias St. Pierre*
4953 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4955 * 0-byte record padding oracle
4957 If an application encounters a fatal protocol error and then calls
4958 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4959 then OpenSSL can respond differently to the calling application if a 0 byte
4960 record is received with invalid padding compared to if a 0 byte record is
4961 received with an invalid MAC. If the application then behaves differently
4962 based on that in a way that is detectable to the remote peer, then this
4963 amounts to a padding oracle that could be used to decrypt data.
4965 In order for this to be exploitable "non-stitched" ciphersuites must be in
4966 use. Stitched ciphersuites are optimised implementations of certain
4967 commonly used ciphersuites. Also the application must call SSL_shutdown()
4968 twice even if a protocol error has occurred (applications should not do
4969 this but some do anyway).
4971 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4972 Aviram, with additional investigation by Steven Collison and Andrew
4973 Hourselt. It was reported to OpenSSL on 10th December 2018.
4978 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4982 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4984 * Microarchitecture timing vulnerability in ECC scalar multiplication
4986 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4987 shown to be vulnerable to a microarchitecture timing side channel attack.
4988 An attacker with sufficient access to mount local timing attacks during
4989 ECDSA signature generation could recover the private key.
4991 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4992 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4998 * Timing vulnerability in DSA signature generation
5000 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5001 timing side channel attack. An attacker could use variations in the signing
5002 algorithm to recover the private key.
5004 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5009 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5010 Module, accidentally introduced while backporting security fixes from the
5011 development branch and hindering the use of ECC in FIPS mode.
5015 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5017 * Client DoS due to large DH parameter
5019 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5020 malicious server can send a very large prime value to the client. This will
5021 cause the client to spend an unreasonably long period of time generating a
5022 key for this prime resulting in a hang until the client has finished. This
5023 could be exploited in a Denial Of Service attack.
5025 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5030 * Cache timing vulnerability in RSA Key Generation
5032 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5033 a cache timing side channel attack. An attacker with sufficient access to
5034 mount cache timing attacks during the RSA key generation process could
5035 recover the private key.
5037 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5038 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5043 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5044 parameter is no longer accepted, as it leads to a corrupt table. NULL
5045 pem_str is reserved for alias entries only.
5049 * Revert blinding in ECDSA sign and instead make problematic addition
5050 length-invariant. Switch even to fixed-length Montgomery multiplication.
5054 * Change generating and checking of primes so that the error rate of not
5055 being prime depends on the intended use based on the size of the input.
5056 For larger primes this will result in more rounds of Miller-Rabin.
5057 The maximal error rate for primes with more than 1080 bits is lowered
5060 *Kurt Roeckx, Annie Yousar*
5062 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5066 * Add blinding to ECDSA and DSA signatures to protect against side channel
5067 attacks discovered by Keegan Ryan (NCC Group).
5071 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5072 now allow empty (zero character) pass phrases.
5076 * Certificate time validation (X509_cmp_time) enforces stricter
5077 compliance with RFC 5280. Fractional seconds and timezone offsets
5078 are no longer allowed.
5082 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5084 * Constructed ASN.1 types with a recursive definition could exceed the stack
5086 Constructed ASN.1 types with a recursive definition (such as can be found
5087 in PKCS7) could eventually exceed the stack given malicious input with
5088 excessive recursion. This could result in a Denial Of Service attack. There
5089 are no such structures used within SSL/TLS that come from untrusted sources
5090 so this is considered safe.
5092 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5098 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5100 * Read/write after SSL object in error state
5102 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5103 mechanism. The intent was that if a fatal error occurred during a handshake
5104 then OpenSSL would move into the error state and would immediately fail if
5105 you attempted to continue the handshake. This works as designed for the
5106 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5107 SSL_connect()), however due to a bug it does not work correctly if
5108 SSL_read() or SSL_write() is called directly. In that scenario, if the
5109 handshake fails then a fatal error will be returned in the initial function
5110 call. If SSL_read()/SSL_write() is subsequently called by the application
5111 for the same SSL object then it will succeed and the data is passed without
5112 being decrypted/encrypted directly from the SSL/TLS record layer.
5114 In order to exploit this issue an application bug would have to be present
5115 that resulted in a call to SSL_read()/SSL_write() being issued after having
5116 already received a fatal error.
5118 This issue was reported to OpenSSL by David Benjamin (Google).
5123 * rsaz_1024_mul_avx2 overflow bug on x86_64
5125 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5126 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5127 Analysis suggests that attacks against RSA and DSA as a result of this
5128 defect would be very difficult to perform and are not believed likely.
5129 Attacks against DH1024 are considered just feasible, because most of the
5130 work necessary to deduce information about a private key may be performed
5131 offline. The amount of resources required for such an attack would be
5132 significant. However, for an attack on TLS to be meaningful, the server
5133 would have to share the DH1024 private key among multiple clients, which is
5134 no longer an option since CVE-2016-0701.
5136 This only affects processors that support the AVX2 but not ADX extensions
5137 like Intel Haswell (4th generation).
5139 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5140 was originally found via the OSS-Fuzz project.
5145 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5147 * bn_sqrx8x_internal carry bug on x86_64
5149 There is a carry propagating bug in the x86_64 Montgomery squaring
5150 procedure. No EC algorithms are affected. Analysis suggests that attacks
5151 against RSA and DSA as a result of this defect would be very difficult to
5152 perform and are not believed likely. Attacks against DH are considered just
5153 feasible (although very difficult) because most of the work necessary to
5154 deduce information about a private key may be performed offline. The amount
5155 of resources required for such an attack would be very significant and
5156 likely only accessible to a limited number of attackers. An attacker would
5157 additionally need online access to an unpatched system using the target
5158 private key in a scenario with persistent DH parameters and a private
5159 key that is shared between multiple clients.
5161 This only affects processors that support the BMI1, BMI2 and ADX extensions
5162 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5164 This issue was reported to OpenSSL by the OSS-Fuzz project.
5169 * Malformed X.509 IPAddressFamily could cause OOB read
5171 If an X.509 certificate has a malformed IPAddressFamily extension,
5172 OpenSSL could do a one-byte buffer overread. The most likely result
5173 would be an erroneous display of the certificate in text format.
5175 This issue was reported to OpenSSL by the OSS-Fuzz project.
5179 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5181 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5182 platform rather than 'mingw'.
5186 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5188 * Truncated packet could crash via OOB read
5190 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5191 cipher is being used, then a truncated packet can cause that host to
5192 perform an out-of-bounds read, usually resulting in a crash.
5194 This issue was reported to OpenSSL by Robert Święcki of Google.
5199 * BN_mod_exp may produce incorrect results on x86_64
5201 There is a carry propagating bug in the x86_64 Montgomery squaring
5202 procedure. No EC algorithms are affected. Analysis suggests that attacks
5203 against RSA and DSA as a result of this defect would be very difficult to
5204 perform and are not believed likely. Attacks against DH are considered just
5205 feasible (although very difficult) because most of the work necessary to
5206 deduce information about a private key may be performed offline. The amount
5207 of resources required for such an attack would be very significant and
5208 likely only accessible to a limited number of attackers. An attacker would
5209 additionally need online access to an unpatched system using the target
5210 private key in a scenario with persistent DH parameters and a private
5211 key that is shared between multiple clients. For example this can occur by
5212 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5213 similar to CVE-2015-3193 but must be treated as a separate problem.
5215 This issue was reported to OpenSSL by the OSS-Fuzz project.
5220 * Montgomery multiplication may produce incorrect results
5222 There is a carry propagating bug in the Broadwell-specific Montgomery
5223 multiplication procedure that handles input lengths divisible by, but
5224 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5225 and DH private keys are impossible. This is because the subroutine in
5226 question is not used in operations with the private key itself and an input
5227 of the attacker's direct choice. Otherwise the bug can manifest itself as
5228 transient authentication and key negotiation failures or reproducible
5229 erroneous outcome of public-key operations with specially crafted input.
5230 Among EC algorithms only Brainpool P-512 curves are affected and one
5231 presumably can attack ECDH key negotiation. Impact was not analyzed in
5232 detail, because pre-requisites for attack are considered unlikely. Namely
5233 multiple clients have to choose the curve in question and the server has to
5234 share the private key among them, neither of which is default behaviour.
5235 Even then only clients that chose the curve will be affected.
5237 This issue was publicly reported as transient failures and was not
5238 initially recognized as a security issue. Thanks to Richard Morgan for
5239 providing reproducible case.
5244 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5245 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5246 prevent issues where no progress is being made and the peer continually
5247 sends unrecognised record types, using up resources processing them.
5251 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5253 * Missing CRL sanity check
5255 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5256 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5257 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5259 This issue only affects the OpenSSL 1.0.2i
5264 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5266 * OCSP Status Request extension unbounded memory growth
5268 A malicious client can send an excessively large OCSP Status Request
5269 extension. If that client continually requests renegotiation, sending a
5270 large OCSP Status Request extension each time, then there will be unbounded
5271 memory growth on the server. This will eventually lead to a Denial Of
5272 Service attack through memory exhaustion. Servers with a default
5273 configuration are vulnerable even if they do not support OCSP. Builds using
5274 the "no-ocsp" build time option are not affected.
5276 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5281 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5284 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5290 * OOB write in MDC2_Update()
5292 An overflow can occur in MDC2_Update() either if called directly or
5293 through the EVP_DigestUpdate() function using MDC2. If an attacker
5294 is able to supply very large amounts of input data after a previous
5295 call to EVP_EncryptUpdate() with a partial block then a length check
5296 can overflow resulting in a heap corruption.
5298 The amount of data needed is comparable to SIZE_MAX which is impractical
5301 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5306 * Malformed SHA512 ticket DoS
5308 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5309 DoS attack where a malformed ticket will result in an OOB read which will
5312 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5313 a custom server callback and ticket lookup mechanism.
5315 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5320 * OOB write in BN_bn2dec()
5322 The function BN_bn2dec() does not check the return value of BN_div_word().
5323 This can cause an OOB write if an application uses this function with an
5324 overly large BIGNUM. This could be a problem if an overly large certificate
5325 or CRL is printed out from an untrusted source. TLS is not affected because
5326 record limits will reject an oversized certificate before it is parsed.
5328 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5333 * OOB read in TS_OBJ_print_bio()
5335 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5336 the total length the OID text representation would use and not the amount
5337 of data written. This will result in OOB reads when large OIDs are
5340 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5345 * Pointer arithmetic undefined behaviour
5347 Avoid some undefined pointer arithmetic
5349 A common idiom in the codebase is to check limits in the following manner:
5352 Where "p" points to some malloc'd data of SIZE bytes and
5355 "len" here could be from some externally supplied data (e.g. from a TLS
5358 The rules of C pointer arithmetic are such that "p + len" is only well
5359 defined where len <= SIZE. Therefore the above idiom is actually
5360 undefined behaviour.
5362 For example this could cause problems if some malloc implementation
5363 provides an address for "p" such that "p + len" actually overflows for
5364 values of len that are too big and therefore p + len < limit.
5366 This issue was reported to OpenSSL by Guido Vranken
5371 * Constant time flag not preserved in DSA signing
5373 Operations in the DSA signing algorithm should run in constant time in
5374 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5375 implementation means that a non-constant time codepath is followed for
5376 certain operations. This has been demonstrated through a cache-timing
5377 attack to be sufficient for an attacker to recover the private DSA key.
5379 This issue was reported by César Pereida (Aalto University), Billy Brumley
5380 (Tampere University of Technology), and Yuval Yarom (The University of
5381 Adelaide and NICTA).
5386 * DTLS buffered message DoS
5388 In a DTLS connection where handshake messages are delivered out-of-order
5389 those messages that OpenSSL is not yet ready to process will be buffered
5390 for later use. Under certain circumstances, a flaw in the logic means that
5391 those messages do not get removed from the buffer even though the handshake
5392 has been completed. An attacker could force up to approx. 15 messages to
5393 remain in the buffer when they are no longer required. These messages will
5394 be cleared when the DTLS connection is closed. The default maximum size for
5395 a message is 100k. Therefore, the attacker could force an additional 1500k
5396 to be consumed per connection. By opening many simultaneous connections an
5397 attacker could cause a DoS attack through memory exhaustion.
5399 This issue was reported to OpenSSL by Quan Luo.
5404 * DTLS replay protection DoS
5406 A flaw in the DTLS replay attack protection mechanism means that records
5407 that arrive for future epochs update the replay protection "window" before
5408 the MAC for the record has been validated. This could be exploited by an
5409 attacker by sending a record for the next epoch (which does not have to
5410 decrypt or have a valid MAC), with a very large sequence number. This means
5411 that all subsequent legitimate packets are dropped causing a denial of
5412 service for a specific DTLS connection.
5414 This issue was reported to OpenSSL by the OCAP audit team.
5419 * Certificate message OOB reads
5421 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5422 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5423 theoretical DoS risk but this has not been observed in practice on common
5426 The messages affected are client certificate, client certificate request
5427 and server certificate. As a result the attack can only be performed
5428 against a client or a server which enables client authentication.
5430 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5435 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5437 * Prevent padding oracle in AES-NI CBC MAC check
5439 A MITM attacker can use a padding oracle attack to decrypt traffic
5440 when the connection uses an AES CBC cipher and the server support
5443 This issue was introduced as part of the fix for Lucky 13 padding
5444 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5445 constant time by making sure that always the same bytes are read and
5446 compared against either the MAC or padding bytes. But it no longer
5447 checked that there was enough data to have both the MAC and padding
5450 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5454 * Fix EVP_EncodeUpdate overflow
5456 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5457 Base64 encoding of binary data. If an attacker is able to supply very large
5458 amounts of input data then a length check can overflow resulting in a heap
5461 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5462 the `PEM_write_bio*` family of functions. These are mainly used within the
5463 OpenSSL command line applications, so any application which processes data
5464 from an untrusted source and outputs it as a PEM file should be considered
5465 vulnerable to this issue. User applications that call these APIs directly
5466 with large amounts of untrusted data may also be vulnerable.
5468 This issue was reported by Guido Vranken.
5473 * Fix EVP_EncryptUpdate overflow
5475 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5476 is able to supply very large amounts of input data after a previous call to
5477 EVP_EncryptUpdate() with a partial block then a length check can overflow
5478 resulting in a heap corruption. Following an analysis of all OpenSSL
5479 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5480 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5481 the first called function after an EVP_EncryptInit(), and therefore that
5482 specific call must be safe. The second form is where the length passed to
5483 EVP_EncryptUpdate() can be seen from the code to be some small value and
5484 therefore there is no possibility of an overflow. Since all instances are
5485 one of these two forms, it is believed that there can be no overflows in
5486 internal code due to this problem. It should be noted that
5487 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5488 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5489 of these calls have also been analysed too and it is believed there are no
5490 instances in internal usage where an overflow could occur.
5492 This issue was reported by Guido Vranken.
5497 * Prevent ASN.1 BIO excessive memory allocation
5499 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5500 a short invalid encoding can cause allocation of large amounts of memory
5501 potentially consuming excessive resources or exhausting memory.
5503 Any application parsing untrusted data through d2i BIO functions is
5504 affected. The memory based functions such as d2i_X509() are *not* affected.
5505 Since the memory based functions are used by the TLS library, TLS
5506 applications are not affected.
5508 This issue was reported by Brian Carpenter.
5515 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5516 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5517 in arbitrary stack data being returned in the buffer.
5519 This issue was reported by Guido Vranken.
5524 * Modify behavior of ALPN to invoke callback after SNI/servername
5525 callback, such that updates to the SSL_CTX affect ALPN.
5529 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5534 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5535 methods are enabled and ssl2 is disabled the methods return NULL.
5539 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5541 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5542 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5543 provide any "EXPORT" or "LOW" strength ciphers.
5547 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5548 is by default disabled at build-time. Builds that are not configured with
5549 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5550 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5551 will need to explicitly call either of:
5553 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5555 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5557 as appropriate. Even if either of those is used, or the application
5558 explicitly uses the version-specific SSLv2_method() or its client and
5559 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5560 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5561 ciphers, and SSLv2 56-bit DES are no longer available.
5566 * Fix a double-free in DSA code
5568 A double free bug was discovered when OpenSSL parses malformed DSA private
5569 keys and could lead to a DoS attack or memory corruption for applications
5570 that receive DSA private keys from untrusted sources. This scenario is
5573 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5579 * Disable SRP fake user seed to address a server memory leak.
5581 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5583 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5584 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5585 was changed to ignore the "fake user" SRP seed, even if the seed
5588 Users should use SRP_VBASE_get1_by_user instead. Note that in
5589 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5590 also that even though configuring the SRP seed attempts to hide
5591 invalid usernames by continuing the handshake with fake
5592 credentials, this behaviour is not constant time and no strong
5593 guarantees are made that the handshake is indistinguishable from
5594 that of a valid user.
5599 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5601 In the BN_hex2bn function the number of hex digits is calculated using an
5602 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5603 large values of `i` this can result in `bn_expand` not allocating any
5604 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5605 field as NULL leading to a subsequent NULL ptr deref. For very large values
5606 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5607 In this case memory is allocated to the internal BIGNUM data field, but it
5608 is insufficiently sized leading to heap corruption. A similar issue exists
5609 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5610 is ever called by user applications with very large untrusted hex/dec data.
5611 This is anticipated to be a rare occurrence.
5613 All OpenSSL internal usage of these functions use data that is not expected
5614 to be untrusted, e.g. config file data or application command line
5615 arguments. If user developed applications generate config file data based
5616 on untrusted data then it is possible that this could also lead to security
5617 consequences. This is also anticipated to be rare.
5619 This issue was reported to OpenSSL by Guido Vranken.
5624 * Fix memory issues in `BIO_*printf` functions
5626 The internal `fmtstr` function used in processing a "%s" format string in
5627 the `BIO_*printf` functions could overflow while calculating the length of a
5628 string and cause an OOB read when printing very long strings.
5630 Additionally the internal `doapr_outch` function can attempt to write to an
5631 OOB memory location (at an offset from the NULL pointer) in the event of a
5632 memory allocation failure. In 1.0.2 and below this could be caused where
5633 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5634 could be in processing a very long "%s" format string. Memory leaks can
5637 The first issue may mask the second issue dependent on compiler behaviour.
5638 These problems could enable attacks where large amounts of untrusted data
5639 is passed to the `BIO_*printf` functions. If applications use these functions
5640 in this way then they could be vulnerable. OpenSSL itself uses these
5641 functions when printing out human-readable dumps of ASN.1 data. Therefore
5642 applications that print this data could be vulnerable if the data is from
5643 untrusted sources. OpenSSL command line applications could also be
5644 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5645 as command line arguments.
5647 Libssl is not considered directly vulnerable. Additionally certificates etc
5648 received via remote connections via libssl are also unlikely to be able to
5649 trigger these issues because of message size limits enforced within libssl.
5651 This issue was reported to OpenSSL Guido Vranken.
5656 * Side channel attack on modular exponentiation
5658 A side-channel attack was found which makes use of cache-bank conflicts on
5659 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5660 of RSA keys. The ability to exploit this issue is limited as it relies on
5661 an attacker who has control of code in a thread running on the same
5662 hyper-threaded core as the victim thread which is performing decryptions.
5664 This issue was reported to OpenSSL by Yuval Yarom, The University of
5665 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5666 Nadia Heninger, University of Pennsylvania with more information at
5667 <http://cachebleed.info>.
5672 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5673 if no keysize is specified with default_bits. This fixes an
5674 omission in an earlier change that changed all RSA/DSA key generation
5675 commands to use 2048 bits by default.
5679 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5681 * DH small subgroups
5683 Historically OpenSSL only ever generated DH parameters based on "safe"
5684 primes. More recently (in version 1.0.2) support was provided for
5685 generating X9.42 style parameter files such as those required for RFC 5114
5686 support. The primes used in such files may not be "safe". Where an
5687 application is using DH configured with parameters based on primes that are
5688 not "safe" then an attacker could use this fact to find a peer's private
5689 DH exponent. This attack requires that the attacker complete multiple
5690 handshakes in which the peer uses the same private DH exponent. For example
5691 this could be used to discover a TLS server's private DH exponent if it's
5692 reusing the private DH exponent or it's using a static DH ciphersuite.
5694 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5695 TLS. It is not on by default. If the option is not set then the server
5696 reuses the same private DH exponent for the life of the server process and
5697 would be vulnerable to this attack. It is believed that many popular
5698 applications do set this option and would therefore not be at risk.
5700 The fix for this issue adds an additional check where a "q" parameter is
5701 available (as is the case in X9.42 based parameters). This detects the
5702 only known attack, and is the only possible defense for static DH
5703 ciphersuites. This could have some performance impact.
5705 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5706 default and cannot be disabled. This could have some performance impact.
5708 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5713 * SSLv2 doesn't block disabled ciphers
5715 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5716 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5717 been disabled, provided that the SSLv2 protocol was not also disabled via
5720 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5721 and Sebastian Schinzel.
5726 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5728 * BN_mod_exp may produce incorrect results on x86_64
5730 There is a carry propagating bug in the x86_64 Montgomery squaring
5731 procedure. No EC algorithms are affected. Analysis suggests that attacks
5732 against RSA and DSA as a result of this defect would be very difficult to
5733 perform and are not believed likely. Attacks against DH are considered just
5734 feasible (although very difficult) because most of the work necessary to
5735 deduce information about a private key may be performed offline. The amount
5736 of resources required for such an attack would be very significant and
5737 likely only accessible to a limited number of attackers. An attacker would
5738 additionally need online access to an unpatched system using the target
5739 private key in a scenario with persistent DH parameters and a private
5740 key that is shared between multiple clients. For example this can occur by
5741 default in OpenSSL DHE based SSL/TLS ciphersuites.
5743 This issue was reported to OpenSSL by Hanno Böck.
5748 * Certificate verify crash with missing PSS parameter
5750 The signature verification routines will crash with a NULL pointer
5751 dereference if presented with an ASN.1 signature using the RSA PSS
5752 algorithm and absent mask generation function parameter. Since these
5753 routines are used to verify certificate signature algorithms this can be
5754 used to crash any certificate verification operation and exploited in a
5755 DoS attack. Any application which performs certificate verification is
5756 vulnerable including OpenSSL clients and servers which enable client
5759 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5764 * X509_ATTRIBUTE memory leak
5766 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5767 memory. This structure is used by the PKCS#7 and CMS routines so any
5768 application which reads PKCS#7 or CMS data from untrusted sources is
5769 affected. SSL/TLS is not affected.
5771 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5777 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5778 This changes the decoding behaviour for some invalid messages,
5779 though the change is mostly in the more lenient direction, and
5780 legacy behaviour is preserved as much as possible.
5784 * In DSA_generate_parameters_ex, if the provided seed is too short,
5787 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5789 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5791 * Alternate chains certificate forgery
5793 During certificate verification, OpenSSL will attempt to find an
5794 alternative certificate chain if the first attempt to build such a chain
5795 fails. An error in the implementation of this logic can mean that an
5796 attacker could cause certain checks on untrusted certificates to be
5797 bypassed, such as the CA flag, enabling them to use a valid leaf
5798 certificate to act as a CA and "issue" an invalid certificate.
5800 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5805 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5807 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5808 incompatibility in the handling of HMAC. The previous ABI has now been
5813 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5815 * Malformed ECParameters causes infinite loop
5817 When processing an ECParameters structure OpenSSL enters an infinite loop
5818 if the curve specified is over a specially malformed binary polynomial
5821 This can be used to perform denial of service against any
5822 system which processes public keys, certificate requests or
5823 certificates. This includes TLS clients and TLS servers with
5824 client authentication enabled.
5826 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5831 * Exploitable out-of-bounds read in X509_cmp_time
5833 X509_cmp_time does not properly check the length of the ASN1_TIME
5834 string and can read a few bytes out of bounds. In addition,
5835 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5838 An attacker can use this to craft malformed certificates and CRLs of
5839 various sizes and potentially cause a segmentation fault, resulting in
5840 a DoS on applications that verify certificates or CRLs. TLS clients
5841 that verify CRLs are affected. TLS clients and servers with client
5842 authentication enabled may be affected if they use custom verification
5845 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5846 independently by Hanno Böck.
5851 * PKCS7 crash with missing EnvelopedContent
5853 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5854 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5855 with missing content and trigger a NULL pointer dereference on parsing.
5857 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5858 structures from untrusted sources are affected. OpenSSL clients and
5859 servers are not affected.
5861 This issue was reported to OpenSSL by Michal Zalewski (Google).
5866 * CMS verify infinite loop with unknown hash function
5868 When verifying a signedData message the CMS code can enter an infinite loop
5869 if presented with an unknown hash function OID. This can be used to perform
5870 denial of service against any system which verifies signedData messages using
5872 This issue was reported to OpenSSL by Johannes Bauer.
5877 * Race condition handling NewSessionTicket
5879 If a NewSessionTicket is received by a multi-threaded client when attempting to
5880 reuse a previous ticket then a race condition can occur potentially leading to
5881 a double free of the ticket data.
5886 * Only support 256-bit or stronger elliptic curves with the
5887 'ecdh_auto' setting (server) or by default (client). Of supported
5888 curves, prefer P-256 (both).
5892 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5894 * ClientHello sigalgs DoS fix
5896 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5897 invalid signature algorithms extension a NULL pointer dereference will
5898 occur. This can be exploited in a DoS attack against the server.
5900 This issue was was reported to OpenSSL by David Ramos of Stanford
5904 *Stephen Henson and Matt Caswell*
5906 * Multiblock corrupted pointer fix
5908 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5909 feature only applies on 64 bit x86 architecture platforms that support AES
5910 NI instructions. A defect in the implementation of "multiblock" can cause
5911 OpenSSL's internal write buffer to become incorrectly set to NULL when
5912 using non-blocking IO. Typically, when the user application is using a
5913 socket BIO for writing, this will only result in a failed connection.
5914 However if some other BIO is used then it is likely that a segmentation
5915 fault will be triggered, thus enabling a potential DoS attack.
5917 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5922 * Segmentation fault in DTLSv1_listen fix
5924 The DTLSv1_listen function is intended to be stateless and processes the
5925 initial ClientHello from many peers. It is common for user code to loop
5926 over the call to DTLSv1_listen until a valid ClientHello is received with
5927 an associated cookie. A defect in the implementation of DTLSv1_listen means
5928 that state is preserved in the SSL object from one invocation to the next
5929 that can lead to a segmentation fault. Errors processing the initial
5930 ClientHello can trigger this scenario. An example of such an error could be
5931 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5934 This issue was reported to OpenSSL by Per Allansson.
5939 * Segmentation fault in ASN1_TYPE_cmp fix
5941 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5942 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5943 certificate signature algorithm consistency 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.
5951 * Segmentation fault for invalid PSS parameters fix
5953 The signature verification routines will crash with a NULL pointer
5954 dereference if presented with an ASN.1 signature using the RSA PSS
5955 algorithm and invalid parameters. Since these routines are used to verify
5956 certificate signature algorithms this can be used to crash any
5957 certificate verification operation and exploited in a DoS attack. Any
5958 application which performs certificate verification is vulnerable including
5959 OpenSSL clients and servers which enable client authentication.
5961 This issue was was reported to OpenSSL by Brian Carpenter.
5966 * ASN.1 structure reuse memory corruption fix
5968 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5969 memory corruption via an invalid write. Such reuse is and has been
5970 strongly discouraged and is believed to be rare.
5972 Applications that parse structures containing CHOICE or ANY DEFINED BY
5973 components may be affected. Certificate parsing (d2i_X509 and related
5974 functions) are however not affected. OpenSSL clients and servers are
5980 * PKCS7 NULL pointer dereferences fix
5982 The PKCS#7 parsing code does not handle missing outer ContentInfo
5983 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5984 missing content and trigger a NULL pointer dereference on parsing.
5986 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5987 otherwise parse PKCS#7 structures from untrusted sources are
5988 affected. OpenSSL clients and servers are not affected.
5990 This issue was reported to OpenSSL by Michal Zalewski (Google).
5995 * DoS via reachable assert in SSLv2 servers fix
5997 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5998 servers that both support SSLv2 and enable export cipher suites by sending
5999 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6001 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6002 (OpenSSL development team).
6007 * Empty CKE with client auth and DHE fix
6009 If client auth is used then a server can seg fault in the event of a DHE
6010 ciphersuite being selected and a zero length ClientKeyExchange message
6011 being sent by the client. This could be exploited in a DoS attack.
6016 * Handshake with unseeded PRNG fix
6018 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6019 with an unseeded PRNG. The conditions are:
6020 - The client is on a platform where the PRNG has not been seeded
6021 automatically, and the user has not seeded manually
6022 - A protocol specific client method version has been used (i.e. not
6023 SSL_client_methodv23)
6024 - A ciphersuite is used that does not require additional random data from
6025 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6027 If the handshake succeeds then the client random that has been used will
6028 have been generated from a PRNG with insufficient entropy and therefore the
6029 output may be predictable.
6031 For example using the following command with an unseeded openssl will
6032 succeed on an unpatched platform:
6034 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6039 * Use After Free following d2i_ECPrivatekey error fix
6041 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6042 could cause a use after free condition. This, in turn, could cause a double
6043 free in several private key parsing functions (such as d2i_PrivateKey
6044 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6045 for applications that receive EC private keys from untrusted
6046 sources. This scenario is considered rare.
6048 This issue was discovered by the BoringSSL project and fixed in their
6054 * X509_to_X509_REQ NULL pointer deref fix
6056 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6057 the certificate key is invalid. This function is rarely used in practice.
6059 This issue was discovered by Brian Carpenter.
6064 * Removed the export ciphers from the DEFAULT ciphers
6068 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6070 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6071 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6072 So far those who have to target multiple platforms would compromise
6073 and argue that binary targeting say ARMv5 would still execute on
6074 ARMv8. "Universal" build resolves this compromise by providing
6075 near-optimal performance even on newer platforms.
6079 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6080 (other platforms pending).
6082 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6084 * Add support for the SignedCertificateTimestampList certificate and
6085 OCSP response extensions from RFC6962.
6089 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6090 for corner cases. (Certain input points at infinity could lead to
6091 bogus results, with non-infinity inputs mapped to infinity too.)
6095 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6096 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6097 common cases are optimized and there still is room for further
6098 improvements. Vector Permutation AES for Altivec is also added.
6102 * Add support for little-endian ppc64 Linux target.
6104 *Marcelo Cerri (IBM)*
6106 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6107 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6108 are optimized and there still is room for further improvements.
6109 Both 32- and 64-bit modes are supported.
6111 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6113 * Improved ARMv7 NEON support.
6117 * Support for SPARC Architecture 2011 crypto extensions, first
6118 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6119 SHA256/512, MD5, GHASH and modular exponentiation.
6121 *Andy Polyakov, David Miller*
6123 * Accelerated modular exponentiation for Intel processors, a.k.a.
6126 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6128 * Support for new and upcoming Intel processors, including AVX2,
6129 BMI and SHA ISA extensions. This includes additional "stitched"
6130 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6133 This work was sponsored by Intel Corp.
6137 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6138 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6139 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6143 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6144 this fixes a limitation in previous versions of OpenSSL.
6148 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6149 MGF1 digest and OAEP label.
6153 * Add EVP support for key wrapping algorithms, to avoid problems with
6154 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6155 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6156 algorithms and include tests cases.
6160 * Add functions to allocate and set the fields of an ECDSA_METHOD
6163 *Douglas E. Engert, Steve Henson*
6165 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6166 difference in days and seconds between two tm or ASN1_TIME structures.
6170 * Add -rev test option to s_server to just reverse order of characters
6171 received by client and send back to server. Also prints an abbreviated
6172 summary of the connection parameters.
6176 * New option -brief for s_client and s_server to print out a brief summary
6177 of connection parameters.
6181 * Add callbacks for arbitrary TLS extensions.
6183 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6185 * New option -crl_download in several openssl utilities to download CRLs
6186 from CRLDP extension in certificates.
6190 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6194 * New function X509_CRL_diff to generate a delta CRL from the difference
6195 of two full CRLs. Add support to "crl" utility.
6199 * New functions to set lookup_crls function and to retrieve
6200 X509_STORE from X509_STORE_CTX.
6204 * Print out deprecated issuer and subject unique ID fields in
6209 * Extend OCSP I/O functions so they can be used for simple general purpose
6210 HTTP as well as OCSP. New wrapper function which can be used to download
6211 CRLs using the OCSP API.
6215 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6219 * `SSL_CONF*` functions. These provide a common framework for application
6220 configuration using configuration files or command lines.
6224 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6225 message callback and prints the results. Needs compile time option
6226 "enable-ssl-trace". New options to s_client and s_server to enable
6231 * New ctrl and macro to retrieve supported points extensions.
6232 Print out extension in s_server and s_client.
6236 * New functions to retrieve certificate signature and signature
6241 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6246 * New Suite B modes for TLS code. These use and enforce the requirements
6247 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6248 only use Suite B curves. The Suite B modes can be set by using the
6249 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6253 * New chain verification flags for Suite B levels of security. Check
6254 algorithms are acceptable when flags are set in X509_verify_cert.
6258 * Make tls1_check_chain return a set of flags indicating checks passed
6259 by a certificate chain. Add additional tests to handle client
6260 certificates: checks for matching certificate type and issuer name
6265 * If an attempt is made to use a signature algorithm not in the peer
6266 preference list abort the handshake. If client has no suitable
6267 signature algorithms in response to a certificate request do not
6268 use the certificate.
6272 * If server EC tmp key is not in client preference list abort handshake.
6276 * Add support for certificate stores in CERT structure. This makes it
6277 possible to have different stores per SSL structure or one store in
6278 the parent SSL_CTX. Include distinct stores for certificate chain
6279 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6280 to build and store a certificate chain in CERT structure: returning
6281 an error if the chain cannot be built: this will allow applications
6282 to test if a chain is correctly configured.
6284 Note: if the CERT based stores are not set then the parent SSL_CTX
6285 store is used to retain compatibility with existing behaviour.
6289 * New function ssl_set_client_disabled to set a ciphersuite disabled
6290 mask based on the current session, check mask when sending client
6291 hello and checking the requested ciphersuite.
6295 * New ctrls to retrieve and set certificate types in a certificate
6296 request message. Print out received values in s_client. If certificate
6297 types is not set with custom values set sensible values based on
6298 supported signature algorithms.
6302 * Support for distinct client and server supported signature algorithms.
6306 * Add certificate callback. If set this is called whenever a certificate
6307 is required by client or server. An application can decide which
6308 certificate chain to present based on arbitrary criteria: for example
6309 supported signature algorithms. Add very simple example to s_server.
6310 This fixes many of the problems and restrictions of the existing client
6311 certificate callback: for example you can now clear an existing
6312 certificate and specify the whole chain.
6316 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6317 the certificate can be used for (if anything). Set valid_flags field
6318 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6319 to have similar checks in it.
6321 Add new "cert_flags" field to CERT structure and include a "strict mode".
6322 This enforces some TLS certificate requirements (such as only permitting
6323 certificate signature algorithms contained in the supported algorithms
6324 extension) which some implementations ignore: this option should be used
6325 with caution as it could cause interoperability issues.
6329 * Update and tidy signature algorithm extension processing. Work out
6330 shared signature algorithms based on preferences and peer algorithms
6331 and print them out in s_client and s_server. Abort handshake if no
6332 shared signature algorithms.
6336 * Add new functions to allow customised supported signature algorithms
6337 for SSL and SSL_CTX structures. Add options to s_client and s_server
6342 * New function SSL_certs_clear() to delete all references to certificates
6343 from an SSL structure. Before this once a certificate had been added
6344 it couldn't be removed.
6348 * Integrate hostname, email address and IP address checking with certificate
6349 verification. New verify options supporting checking in openssl utility.
6353 * Fixes and wildcard matching support to hostname and email checking
6354 functions. Add manual page.
6356 *Florian Weimer (Red Hat Product Security Team)*
6358 * New functions to check a hostname email or IP address against a
6359 certificate. Add options x509 utility to print results of checks against
6364 * Fix OCSP checking.
6366 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6368 * Initial experimental support for explicitly trusted non-root CAs.
6369 OpenSSL still tries to build a complete chain to a root but if an
6370 intermediate CA has a trust setting included that is used. The first
6371 setting is used: whether to trust (e.g., -addtrust option to the x509
6376 * Add -trusted_first option which attempts to find certificates in the
6377 trusted store even if an untrusted chain is also supplied.
6381 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6382 platform support for Linux and Android.
6386 * Support for linux-x32, ILP32 environment in x86_64 framework.
6390 * Experimental multi-implementation support for FIPS capable OpenSSL.
6391 When in FIPS mode the approved implementations are used as normal,
6392 when not in FIPS mode the internal unapproved versions are used instead.
6393 This means that the FIPS capable OpenSSL isn't forced to use the
6394 (often lower performance) FIPS implementations outside FIPS mode.
6398 * Transparently support X9.42 DH parameters when calling
6399 PEM_read_bio_DHparameters. This means existing applications can handle
6400 the new parameter format automatically.
6404 * Initial experimental support for X9.42 DH parameter format: mainly
6405 to support use of 'q' parameter for RFC5114 parameters.
6409 * Add DH parameters from RFC5114 including test data to dhtest.
6413 * Support for automatic EC temporary key parameter selection. If enabled
6414 the most preferred EC parameters are automatically used instead of
6415 hardcoded fixed parameters. Now a server just has to call:
6416 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6417 support ECDH and use the most appropriate parameters.
6421 * Enhance and tidy EC curve and point format TLS extension code. Use
6422 static structures instead of allocation if default values are used.
6423 New ctrls to set curves we wish to support and to retrieve shared curves.
6424 Print out shared curves in s_server. New options to s_server and s_client
6425 to set list of supported curves.
6429 * New ctrls to retrieve supported signature algorithms and
6430 supported curve values as an array of NIDs. Extend openssl utility
6431 to print out received values.
6435 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6436 between NIDs and the more common NIST names such as "P-256". Enhance
6437 ecparam utility and ECC method to recognise the NIST names for curves.
6441 * Enhance SSL/TLS certificate chain handling to support different
6442 chains for each certificate instead of one chain in the parent SSL_CTX.
6446 * Support for fixed DH ciphersuite client authentication: where both
6447 server and client use DH certificates with common parameters.
6451 * Support for fixed DH ciphersuites: those requiring DH server
6456 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6458 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6459 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6460 X509_CINF_get_signature were reverted post internal team review.
6465 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6467 * OCSP Status Request extension unbounded memory growth
6469 A malicious client can send an excessively large OCSP Status Request
6470 extension. If that client continually requests renegotiation, sending a
6471 large OCSP Status Request extension each time, then there will be unbounded
6472 memory growth on the server. This will eventually lead to a Denial Of
6473 Service attack through memory exhaustion. Servers with a default
6474 configuration are vulnerable even if they do not support OCSP. Builds using
6475 the "no-ocsp" build time option are not affected.
6477 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6482 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6485 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6491 * OOB write in MDC2_Update()
6493 An overflow can occur in MDC2_Update() either if called directly or
6494 through the EVP_DigestUpdate() function using MDC2. If an attacker
6495 is able to supply very large amounts of input data after a previous
6496 call to EVP_EncryptUpdate() with a partial block then a length check
6497 can overflow resulting in a heap corruption.
6499 The amount of data needed is comparable to SIZE_MAX which is impractical
6502 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6507 * Malformed SHA512 ticket DoS
6509 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6510 DoS attack where a malformed ticket will result in an OOB read which will
6513 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6514 a custom server callback and ticket lookup mechanism.
6516 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6521 * OOB write in BN_bn2dec()
6523 The function BN_bn2dec() does not check the return value of BN_div_word().
6524 This can cause an OOB write if an application uses this function with an
6525 overly large BIGNUM. This could be a problem if an overly large certificate
6526 or CRL is printed out from an untrusted source. TLS is not affected because
6527 record limits will reject an oversized certificate before it is parsed.
6529 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6534 * OOB read in TS_OBJ_print_bio()
6536 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6537 the total length the OID text representation would use and not the amount
6538 of data written. This will result in OOB reads when large OIDs are
6541 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6546 * Pointer arithmetic undefined behaviour
6548 Avoid some undefined pointer arithmetic
6550 A common idiom in the codebase is to check limits in the following manner:
6553 Where "p" points to some malloc'd data of SIZE bytes and
6556 "len" here could be from some externally supplied data (e.g. from a TLS
6559 The rules of C pointer arithmetic are such that "p + len" is only well
6560 defined where len <= SIZE. Therefore, the above idiom is actually
6561 undefined behaviour.
6563 For example this could cause problems if some malloc implementation
6564 provides an address for "p" such that "p + len" actually overflows for
6565 values of len that are too big and therefore p + len < limit.
6567 This issue was reported to OpenSSL by Guido Vranken
6572 * Constant time flag not preserved in DSA signing
6574 Operations in the DSA signing algorithm should run in constant time in
6575 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6576 implementation means that a non-constant time codepath is followed for
6577 certain operations. This has been demonstrated through a cache-timing
6578 attack to be sufficient for an attacker to recover the private DSA key.
6580 This issue was reported by César Pereida (Aalto University), Billy Brumley
6581 (Tampere University of Technology), and Yuval Yarom (The University of
6582 Adelaide and NICTA).
6587 * DTLS buffered message DoS
6589 In a DTLS connection where handshake messages are delivered out-of-order
6590 those messages that OpenSSL is not yet ready to process will be buffered
6591 for later use. Under certain circumstances, a flaw in the logic means that
6592 those messages do not get removed from the buffer even though the handshake
6593 has been completed. An attacker could force up to approx. 15 messages to
6594 remain in the buffer when they are no longer required. These messages will
6595 be cleared when the DTLS connection is closed. The default maximum size for
6596 a message is 100k. Therefore, the attacker could force an additional 1500k
6597 to be consumed per connection. By opening many simultaneous connections an
6598 attacker could cause a DoS attack through memory exhaustion.
6600 This issue was reported to OpenSSL by Quan Luo.
6605 * DTLS replay protection DoS
6607 A flaw in the DTLS replay attack protection mechanism means that records
6608 that arrive for future epochs update the replay protection "window" before
6609 the MAC for the record has been validated. This could be exploited by an
6610 attacker by sending a record for the next epoch (which does not have to
6611 decrypt or have a valid MAC), with a very large sequence number. This means
6612 that all subsequent legitimate packets are dropped causing a denial of
6613 service for a specific DTLS connection.
6615 This issue was reported to OpenSSL by the OCAP audit team.
6620 * Certificate message OOB reads
6622 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6623 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6624 theoretical DoS risk but this has not been observed in practice on common
6627 The messages affected are client certificate, client certificate request
6628 and server certificate. As a result the attack can only be performed
6629 against a client or a server which enables client authentication.
6631 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6636 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6638 * Prevent padding oracle in AES-NI CBC MAC check
6640 A MITM attacker can use a padding oracle attack to decrypt traffic
6641 when the connection uses an AES CBC cipher and the server support
6644 This issue was introduced as part of the fix for Lucky 13 padding
6645 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6646 constant time by making sure that always the same bytes are read and
6647 compared against either the MAC or padding bytes. But it no longer
6648 checked that there was enough data to have both the MAC and padding
6651 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6656 * Fix EVP_EncodeUpdate overflow
6658 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6659 Base64 encoding of binary data. If an attacker is able to supply very large
6660 amounts of input data then a length check can overflow resulting in a heap
6663 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6664 the `PEM_write_bio*` family of functions. These are mainly used within the
6665 OpenSSL command line applications, so any application which processes data
6666 from an untrusted source and outputs it as a PEM file should be considered
6667 vulnerable to this issue. User applications that call these APIs directly
6668 with large amounts of untrusted data may also be vulnerable.
6670 This issue was reported by Guido Vranken.
6675 * Fix EVP_EncryptUpdate overflow
6677 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6678 is able to supply very large amounts of input data after a previous call to
6679 EVP_EncryptUpdate() with a partial block then a length check can overflow
6680 resulting in a heap corruption. Following an analysis of all OpenSSL
6681 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6682 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6683 the first called function after an EVP_EncryptInit(), and therefore that
6684 specific call must be safe. The second form is where the length passed to
6685 EVP_EncryptUpdate() can be seen from the code to be some small value and
6686 therefore there is no possibility of an overflow. Since all instances are
6687 one of these two forms, it is believed that there can be no overflows in
6688 internal code due to this problem. It should be noted that
6689 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6690 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6691 of these calls have also been analysed too and it is believed there are no
6692 instances in internal usage where an overflow could occur.
6694 This issue was reported by Guido Vranken.
6699 * Prevent ASN.1 BIO excessive memory allocation
6701 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6702 a short invalid encoding can casuse allocation of large amounts of memory
6703 potentially consuming excessive resources or exhausting memory.
6705 Any application parsing untrusted data through d2i BIO functions is
6706 affected. The memory based functions such as d2i_X509() are *not* affected.
6707 Since the memory based functions are used by the TLS library, TLS
6708 applications are not affected.
6710 This issue was reported by Brian Carpenter.
6717 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6718 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6719 in arbitrary stack data being returned in the buffer.
6721 This issue was reported by Guido Vranken.
6726 * Modify behavior of ALPN to invoke callback after SNI/servername
6727 callback, such that updates to the SSL_CTX affect ALPN.
6731 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6736 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6737 methods are enabled and ssl2 is disabled the methods return NULL.
6741 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6743 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6744 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6745 provide any "EXPORT" or "LOW" strength ciphers.
6749 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6750 is by default disabled at build-time. Builds that are not configured with
6751 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6752 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6753 will need to explicitly call either of:
6755 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6757 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6759 as appropriate. Even if either of those is used, or the application
6760 explicitly uses the version-specific SSLv2_method() or its client and
6761 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6762 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6763 ciphers, and SSLv2 56-bit DES are no longer available.
6768 * Fix a double-free in DSA code
6770 A double free bug was discovered when OpenSSL parses malformed DSA private
6771 keys and could lead to a DoS attack or memory corruption for applications
6772 that receive DSA private keys from untrusted sources. This scenario is
6775 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6781 * Disable SRP fake user seed to address a server memory leak.
6783 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6785 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6786 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6787 was changed to ignore the "fake user" SRP seed, even if the seed
6790 Users should use SRP_VBASE_get1_by_user instead. Note that in
6791 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6792 also that even though configuring the SRP seed attempts to hide
6793 invalid usernames by continuing the handshake with fake
6794 credentials, this behaviour is not constant time and no strong
6795 guarantees are made that the handshake is indistinguishable from
6796 that of a valid user.
6801 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6803 In the BN_hex2bn function the number of hex digits is calculated using an
6804 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6805 large values of `i` this can result in `bn_expand` not allocating any
6806 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6807 field as NULL leading to a subsequent NULL ptr deref. For very large values
6808 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6809 In this case memory is allocated to the internal BIGNUM data field, but it
6810 is insufficiently sized leading to heap corruption. A similar issue exists
6811 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6812 is ever called by user applications with very large untrusted hex/dec data.
6813 This is anticipated to be a rare occurrence.
6815 All OpenSSL internal usage of these functions use data that is not expected
6816 to be untrusted, e.g. config file data or application command line
6817 arguments. If user developed applications generate config file data based
6818 on untrusted data then it is possible that this could also lead to security
6819 consequences. This is also anticipated to be rare.
6821 This issue was reported to OpenSSL by Guido Vranken.
6826 * Fix memory issues in `BIO_*printf` functions
6828 The internal `fmtstr` function used in processing a "%s" format string in
6829 the `BIO_*printf` functions could overflow while calculating the length of a
6830 string and cause an OOB read when printing very long strings.
6832 Additionally the internal `doapr_outch` function can attempt to write to an
6833 OOB memory location (at an offset from the NULL pointer) in the event of a
6834 memory allocation failure. In 1.0.2 and below this could be caused where
6835 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6836 could be in processing a very long "%s" format string. Memory leaks can
6839 The first issue may mask the second issue dependent on compiler behaviour.
6840 These problems could enable attacks where large amounts of untrusted data
6841 is passed to the `BIO_*printf` functions. If applications use these functions
6842 in this way then they could be vulnerable. OpenSSL itself uses these
6843 functions when printing out human-readable dumps of ASN.1 data. Therefore
6844 applications that print this data could be vulnerable if the data is from
6845 untrusted sources. OpenSSL command line applications could also be
6846 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6847 as command line arguments.
6849 Libssl is not considered directly vulnerable. Additionally certificates etc
6850 received via remote connections via libssl are also unlikely to be able to
6851 trigger these issues because of message size limits enforced within libssl.
6853 This issue was reported to OpenSSL Guido Vranken.
6858 * Side channel attack on modular exponentiation
6860 A side-channel attack was found which makes use of cache-bank conflicts on
6861 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6862 of RSA keys. The ability to exploit this issue is limited as it relies on
6863 an attacker who has control of code in a thread running on the same
6864 hyper-threaded core as the victim thread which is performing decryptions.
6866 This issue was reported to OpenSSL by Yuval Yarom, The University of
6867 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6868 Nadia Heninger, University of Pennsylvania with more information at
6869 <http://cachebleed.info>.
6874 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6875 if no keysize is specified with default_bits. This fixes an
6876 omission in an earlier change that changed all RSA/DSA key generation
6877 commands to use 2048 bits by default.
6881 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6883 * Protection for DH small subgroup attacks
6885 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6886 switched on by default and cannot be disabled. This could have some
6891 * SSLv2 doesn't block disabled ciphers
6893 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6894 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6895 been disabled, provided that the SSLv2 protocol was not also disabled via
6898 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6899 and Sebastian Schinzel.
6904 * Reject DH handshakes with parameters shorter than 1024 bits.
6908 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6910 * Certificate verify crash with missing PSS parameter
6912 The signature verification routines will crash with a NULL pointer
6913 dereference if presented with an ASN.1 signature using the RSA PSS
6914 algorithm and absent mask generation function parameter. Since these
6915 routines are used to verify certificate signature algorithms this can be
6916 used to crash any certificate verification operation and exploited in a
6917 DoS attack. Any application which performs certificate verification is
6918 vulnerable including OpenSSL clients and servers which enable client
6921 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6926 * X509_ATTRIBUTE memory leak
6928 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6929 memory. This structure is used by the PKCS#7 and CMS routines so any
6930 application which reads PKCS#7 or CMS data from untrusted sources is
6931 affected. SSL/TLS is not affected.
6933 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6939 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6940 This changes the decoding behaviour for some invalid messages,
6941 though the change is mostly in the more lenient direction, and
6942 legacy behaviour is preserved as much as possible.
6946 * In DSA_generate_parameters_ex, if the provided seed is too short,
6947 use a random seed, as already documented.
6949 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6951 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6953 * Alternate chains certificate forgery
6955 During certificate verfification, OpenSSL will attempt to find an
6956 alternative certificate chain if the first attempt to build such a chain
6957 fails. An error in the implementation of this logic can mean that an
6958 attacker could cause certain checks on untrusted certificates to be
6959 bypassed, such as the CA flag, enabling them to use a valid leaf
6960 certificate to act as a CA and "issue" an invalid certificate.
6962 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6968 * Race condition handling PSK identify hint
6970 If PSK identity hints are received by a multi-threaded client then
6971 the values are wrongly updated in the parent SSL_CTX structure. This can
6972 result in a race condition potentially leading to a double free of the
6978 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6980 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6981 incompatibility in the handling of HMAC. The previous ABI has now been
6984 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6986 * Malformed ECParameters causes infinite loop
6988 When processing an ECParameters structure OpenSSL enters an infinite loop
6989 if the curve specified is over a specially malformed binary polynomial
6992 This can be used to perform denial of service against any
6993 system which processes public keys, certificate requests or
6994 certificates. This includes TLS clients and TLS servers with
6995 client authentication enabled.
6997 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7002 * Exploitable out-of-bounds read in X509_cmp_time
7004 X509_cmp_time does not properly check the length of the ASN1_TIME
7005 string and can read a few bytes out of bounds. In addition,
7006 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7009 An attacker can use this to craft malformed certificates and CRLs of
7010 various sizes and potentially cause a segmentation fault, resulting in
7011 a DoS on applications that verify certificates or CRLs. TLS clients
7012 that verify CRLs are affected. TLS clients and servers with client
7013 authentication enabled may be affected if they use custom verification
7016 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7017 independently by Hanno Böck.
7022 * PKCS7 crash with missing EnvelopedContent
7024 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7025 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7026 with missing content and trigger a NULL pointer dereference on parsing.
7028 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7029 structures from untrusted sources are affected. OpenSSL clients and
7030 servers are not affected.
7032 This issue was reported to OpenSSL by Michal Zalewski (Google).
7037 * CMS verify infinite loop with unknown hash function
7039 When verifying a signedData message the CMS code can enter an infinite loop
7040 if presented with an unknown hash function OID. This can be used to perform
7041 denial of service against any system which verifies signedData messages using
7043 This issue was reported to OpenSSL by Johannes Bauer.
7048 * Race condition handling NewSessionTicket
7050 If a NewSessionTicket is received by a multi-threaded client when attempting to
7051 reuse a previous ticket then a race condition can occur potentially leading to
7052 a double free of the ticket data.
7057 * Reject DH handshakes with parameters shorter than 768 bits.
7059 *Kurt Roeckx and Emilia Kasper*
7061 * dhparam: generate 2048-bit parameters by default.
7063 *Kurt Roeckx and Emilia Kasper*
7065 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7067 * Segmentation fault in ASN1_TYPE_cmp fix
7069 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7070 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7071 certificate signature algorithm consistency this can be used to crash any
7072 certificate verification operation and exploited in a DoS attack. Any
7073 application which performs certificate verification is vulnerable including
7074 OpenSSL clients and servers which enable client authentication.
7079 * ASN.1 structure reuse memory corruption fix
7081 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7082 memory corruption via an invalid write. Such reuse is and has been
7083 strongly discouraged and is believed to be rare.
7085 Applications that parse structures containing CHOICE or ANY DEFINED BY
7086 components may be affected. Certificate parsing (d2i_X509 and related
7087 functions) are however not affected. OpenSSL clients and servers are
7093 * PKCS7 NULL pointer dereferences fix
7095 The PKCS#7 parsing code does not handle missing outer ContentInfo
7096 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7097 missing content and trigger a NULL pointer dereference on parsing.
7099 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7100 otherwise parse PKCS#7 structures from untrusted sources are
7101 affected. OpenSSL clients and servers are not affected.
7103 This issue was reported to OpenSSL by Michal Zalewski (Google).
7108 * DoS via reachable assert in SSLv2 servers fix
7110 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7111 servers that both support SSLv2 and enable export cipher suites by sending
7112 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7114 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7115 (OpenSSL development team).
7120 * Use After Free following d2i_ECPrivatekey error fix
7122 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7123 could cause a use after free condition. This, in turn, could cause a double
7124 free in several private key parsing functions (such as d2i_PrivateKey
7125 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7126 for applications that receive EC private keys from untrusted
7127 sources. This scenario is considered rare.
7129 This issue was discovered by the BoringSSL project and fixed in their
7135 * X509_to_X509_REQ NULL pointer deref fix
7137 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7138 the certificate key is invalid. This function is rarely used in practice.
7140 This issue was discovered by Brian Carpenter.
7145 * Removed the export ciphers from the DEFAULT ciphers
7149 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7151 * Build fixes for the Windows and OpenVMS platforms
7153 *Matt Caswell and Richard Levitte*
7155 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7157 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7158 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7159 dereference. This could lead to a Denial Of Service attack. Thanks to
7160 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7165 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7166 dtls1_buffer_record function under certain conditions. In particular this
7167 could occur if an attacker sent repeated DTLS records with the same
7168 sequence number but for the next epoch. The memory leak could be exploited
7169 by an attacker in a Denial of Service attack through memory exhaustion.
7170 Thanks to Chris Mueller for reporting this issue.
7175 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7176 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7177 method would be set to NULL which could later result in a NULL pointer
7178 dereference. Thanks to Frank Schmirler for reporting this issue.
7183 * Abort handshake if server key exchange message is omitted for ephemeral
7186 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7187 reporting this issue.
7192 * Remove non-export ephemeral RSA code on client and server. This code
7193 violated the TLS standard by allowing the use of temporary RSA keys in
7194 non-export ciphersuites and could be used by a server to effectively
7195 downgrade the RSA key length used to a value smaller than the server
7196 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7197 INRIA or reporting this issue.
7202 * Fixed issue where DH client certificates are accepted without verification.
7203 An OpenSSL server will accept a DH certificate for client authentication
7204 without the certificate verify message. This effectively allows a client to
7205 authenticate without the use of a private key. This only affects servers
7206 which trust a client certificate authority which issues certificates
7207 containing DH keys: these are extremely rare and hardly ever encountered.
7208 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7214 * Ensure that the session ID context of an SSL is updated when its
7215 SSL_CTX is updated via SSL_set_SSL_CTX.
7217 The session ID context is typically set from the parent SSL_CTX,
7218 and can vary with the CTX.
7222 * Fix various certificate fingerprint issues.
7224 By using non-DER or invalid encodings outside the signed portion of a
7225 certificate the fingerprint can be changed without breaking the signature.
7226 Although no details of the signed portion of the certificate can be changed
7227 this can cause problems with some applications: e.g. those using the
7228 certificate fingerprint for blacklists.
7230 1. Reject signatures with non zero unused bits.
7232 If the BIT STRING containing the signature has non zero unused bits reject
7233 the signature. All current signature algorithms require zero unused bits.
7235 2. Check certificate algorithm consistency.
7237 Check the AlgorithmIdentifier inside TBS matches the one in the
7238 certificate signature. NB: this will result in signature failure
7239 errors for some broken certificates.
7241 Thanks to Konrad Kraszewski from Google for reporting this issue.
7243 3. Check DSA/ECDSA signatures use DER.
7245 Re-encode DSA/ECDSA signatures and compare with the original received
7246 signature. Return an error if there is a mismatch.
7248 This will reject various cases including garbage after signature
7249 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7250 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7251 (negative or with leading zeroes).
7253 Further analysis was conducted and fixes were developed by Stephen Henson
7254 of the OpenSSL core team.
7260 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7261 results on some platforms, including x86_64. This bug occurs at random
7262 with a very low probability, and is not known to be exploitable in any
7263 way, though its exact impact is difficult to determine. Thanks to Pieter
7264 Wuille (Blockstream) who reported this issue and also suggested an initial
7265 fix. Further analysis was conducted by the OpenSSL development team and
7266 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7267 the OpenSSL core team.
7272 * Do not resume sessions on the server if the negotiated protocol
7273 version does not match the session's version. Resuming with a different
7274 version, while not strictly forbidden by the RFC, is of questionable
7275 sanity and breaks all known clients.
7277 *David Benjamin, Emilia Käsper*
7279 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7280 early CCS messages during renegotiation. (Note that because
7281 renegotiation is encrypted, this early CCS was not exploitable.)
7285 * Tighten client-side session ticket handling during renegotiation:
7286 ensure that the client only accepts a session ticket if the server sends
7287 the extension anew in the ServerHello. Previously, a TLS client would
7288 reuse the old extension state and thus accept a session ticket if one was
7289 announced in the initial ServerHello.
7291 Similarly, ensure that the client requires a session ticket if one
7292 was advertised in the ServerHello. Previously, a TLS client would
7293 ignore a missing NewSessionTicket message.
7297 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7301 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7302 sends a carefully crafted handshake message, to cause OpenSSL to fail
7303 to free up to 64k of memory causing a memory leak. This could be
7304 exploited in a Denial Of Service attack. This issue affects OpenSSL
7305 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7306 whether SRTP is used or configured. Implementations of OpenSSL that
7307 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7309 The fix was developed by the OpenSSL team.
7314 * Session Ticket Memory Leak.
7316 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7317 integrity of that ticket is first verified. In the event of a session
7318 ticket integrity check failing, OpenSSL will fail to free memory
7319 causing a memory leak. By sending a large number of invalid session
7320 tickets an attacker could exploit this issue in a Denial Of Service
7326 * Build option no-ssl3 is incomplete.
7328 When OpenSSL is configured with "no-ssl3" as a build option, servers
7329 could accept and complete an SSL 3.0 handshake, and clients could be
7330 configured to send them.
7333 *Akamai and the OpenSSL team*
7335 * Add support for TLS_FALLBACK_SCSV.
7336 Client applications doing fallback retries should call
7337 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7340 *Adam Langley, Bodo Moeller*
7342 * Add additional DigestInfo checks.
7344 Re-encode DigestInto in DER and check against the original when
7345 verifying RSA signature: this will reject any improperly encoded
7346 DigestInfo structures.
7348 Note: this is a precautionary measure and no attacks are currently known.
7352 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7354 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7355 SRP code can be overrun an internal buffer. Add sanity check that
7356 g, A, B < N to SRP code.
7358 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7359 Group for discovering this issue.
7364 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7365 TLS 1.0 instead of higher protocol versions when the ClientHello message
7366 is badly fragmented. This allows a man-in-the-middle attacker to force a
7367 downgrade to TLS 1.0 even if both the server and the client support a
7368 higher protocol version, by modifying the client's TLS records.
7370 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7371 researching this issue.
7376 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7377 to a denial of service attack. A malicious server can crash the client
7378 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7379 ciphersuite and sending carefully crafted handshake messages.
7381 Thanks to Felix Gröbert (Google) for discovering and researching this
7387 * By sending carefully crafted DTLS packets an attacker could cause openssl
7388 to leak memory. This can be exploited through a Denial of Service attack.
7389 Thanks to Adam Langley for discovering and researching this issue.
7394 * An attacker can force openssl to consume large amounts of memory whilst
7395 processing DTLS handshake messages. This can be exploited through a
7396 Denial of Service attack.
7397 Thanks to Adam Langley for discovering and researching this issue.
7402 * An attacker can force an error condition which causes openssl to crash
7403 whilst processing DTLS packets due to memory being freed twice. This
7404 can be exploited through a Denial of Service attack.
7405 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7411 * If a multithreaded client connects to a malicious server using a resumed
7412 session and the server sends an ec point format extension it could write
7413 up to 255 bytes to freed memory.
7415 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7421 * A malicious server can crash an OpenSSL client with a null pointer
7422 dereference (read) by specifying an SRP ciphersuite even though it was not
7423 properly negotiated with the client. This can be exploited through a
7424 Denial of Service attack.
7426 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7427 discovering and researching this issue.
7432 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7433 X509_name_oneline, X509_name_print_ex et al. to leak some information
7434 from the stack. Applications may be affected if they echo pretty printing
7435 output to the attacker.
7437 Thanks to Ivan Fratric (Google) for discovering this issue.
7440 *Emilia Käsper, and Steve Henson*
7442 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7443 for corner cases. (Certain input points at infinity could lead to
7444 bogus results, with non-infinity inputs mapped to infinity too.)
7448 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7450 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7451 handshake can force the use of weak keying material in OpenSSL
7452 SSL/TLS clients and servers.
7454 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7455 researching this issue. ([CVE-2014-0224])
7457 *KIKUCHI Masashi, Steve Henson*
7459 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7460 OpenSSL DTLS client the code can be made to recurse eventually crashing
7463 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7466 *Imre Rad, Steve Henson*
7468 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7469 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7470 client or server. This is potentially exploitable to run arbitrary
7471 code on a vulnerable client or server.
7473 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7475 *Jüri Aedla, Steve Henson*
7477 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7478 are subject to a denial of service attack.
7480 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7481 this issue. ([CVE-2014-3470])
7483 *Felix Gröbert, Ivan Fratric, Steve Henson*
7485 * Harmonize version and its documentation. -f flag is used to display
7488 *mancha <mancha1@zoho.com>*
7490 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7491 in i2d_ECPrivateKey.
7493 *mancha <mancha1@zoho.com>*
7495 * Fix some double frees. These are not thought to be exploitable.
7497 *mancha <mancha1@zoho.com>*
7499 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7501 * A missing bounds check in the handling of the TLS heartbeat extension
7502 can be used to reveal up to 64k of memory to a connected client or
7505 Thanks for Neel Mehta of Google Security for discovering this bug and to
7506 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7507 preparing the fix ([CVE-2014-0160])
7509 *Adam Langley, Bodo Moeller*
7511 * Fix for the attack described in the paper "Recovering OpenSSL
7512 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7513 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7514 <http://eprint.iacr.org/2014/140>
7516 Thanks to Yuval Yarom and Naomi Benger for discovering this
7517 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7519 *Yuval Yarom and Naomi Benger*
7521 * TLS pad extension: draft-agl-tls-padding-03
7523 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7524 TLS client Hello record length value would otherwise be > 255 and
7525 less that 512 pad with a dummy extension containing zeroes so it
7526 is at least 512 bytes long.
7528 *Adam Langley, Steve Henson*
7530 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7532 * Fix for TLS record tampering bug. A carefully crafted invalid
7533 handshake could crash OpenSSL with a NULL pointer exception.
7534 Thanks to Anton Johansson for reporting this issues.
7537 * Keep original DTLS digest and encryption contexts in retransmission
7538 structures so we can use the previous session parameters if they need
7539 to be resent. ([CVE-2013-6450])
7543 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7544 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7545 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7546 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7547 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7548 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7550 *Rob Stradling, Adam Langley*
7552 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7554 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7555 supporting platforms or when small records were transferred.
7557 *Andy Polyakov, Steve Henson*
7559 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7561 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7563 This addresses the flaw in CBC record processing discovered by
7564 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7565 at: <http://www.isg.rhul.ac.uk/tls/>
7567 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7568 Security Group at Royal Holloway, University of London
7569 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7570 Emilia Käsper for the initial patch.
7573 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7575 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7576 ciphersuites which can be exploited in a denial of service attack.
7577 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7578 and detecting this bug and to Wolfgang Ettlinger
7579 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7584 * Return an error when checking OCSP signatures when key is NULL.
7585 This fixes a DoS attack. ([CVE-2013-0166])
7589 * Make openssl verify return errors.
7591 *Chris Palmer <palmer@google.com> and Ben Laurie*
7593 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7594 the right response is stapled. Also change SSL_get_certificate()
7595 so it returns the certificate actually sent.
7596 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7598 *Rob Stradling <rob.stradling@comodo.com>*
7600 * Fix possible deadlock when decoding public keys.
7604 * Don't use TLS 1.0 record version number in initial client hello
7609 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7611 * Sanity check record length before skipping explicit IV in TLS
7612 1.2, 1.1 and DTLS to fix DoS attack.
7614 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7615 fuzzing as a service testing platform.
7620 * Initialise tkeylen properly when encrypting CMS messages.
7621 Thanks to Solar Designer of Openwall for reporting this issue.
7625 * In FIPS mode don't try to use composite ciphers as they are not
7630 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7632 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7633 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7634 mean any application compiled against OpenSSL 1.0.0 headers setting
7635 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7636 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7637 0x10000000L Any application which was previously compiled against
7638 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7639 will need to be recompiled as a result. Letting be results in
7640 inability to disable specifically TLS 1.1 and in client context,
7641 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7645 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7646 disable just protocol X, but all protocols above X *if* there are
7647 protocols *below* X still enabled. In more practical terms it means
7648 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7649 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7650 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7655 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7657 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7658 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7659 in CRYPTO_realloc_clean.
7661 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7662 issue and to Adam Langley <agl@chromium.org> for fixing it.
7665 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7667 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7671 * Workarounds for some broken servers that "hang" if a client hello
7672 record length exceeds 255 bytes.
7674 1. Do not use record version number > TLS 1.0 in initial client
7675 hello: some (but not all) hanging servers will now work.
7676 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7677 the number of ciphers sent in the client hello. This should be
7678 set to an even number, such as 50, for example by passing:
7679 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7680 Most broken servers should now work.
7681 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7682 TLS 1.2 client support entirely.
7686 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7690 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7692 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7693 STRING form instead of a DigestInfo.
7697 * The format used for MDC2 RSA signatures is inconsistent between EVP
7698 and the RSA_sign/RSA_verify functions. This was made more apparent when
7699 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7700 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7701 the correct format in RSA_verify so both forms transparently work.
7705 * Some servers which support TLS 1.0 can choke if we initially indicate
7706 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7707 encrypted premaster secret. As a workaround use the maximum permitted
7708 client version in client hello, this should keep such servers happy
7709 and still work with previous versions of OpenSSL.
7713 * Add support for TLS/DTLS heartbeats.
7715 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7717 * Add support for SCTP.
7719 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7721 * Improved PRNG seeding for VOS.
7723 *Paul Green <Paul.Green@stratus.com>*
7725 * Extensive assembler packs updates, most notably:
7727 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7728 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7729 - x86_64: bit-sliced AES implementation;
7730 - ARM: NEON support, contemporary platforms optimizations;
7731 - s390x: z196 support;
7732 - `*`: GHASH and GF(2^m) multiplication implementations;
7736 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7737 (removal of unnecessary code)
7739 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7741 * Add TLS key material exporter from RFC 5705.
7745 * Add DTLS-SRTP negotiation from RFC 5764.
7749 * Add Next Protocol Negotiation,
7750 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7751 disabled with a no-npn flag to config or Configure. Code donated
7754 *Adam Langley <agl@google.com> and Ben Laurie*
7756 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7757 NIST-P256, NIST-P521, with constant-time single point multiplication on
7758 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7759 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7760 Code made available under Apache License version 2.0.
7762 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7763 line to include this in your build of OpenSSL, and run "make depend" (or
7764 "make update"). This enables the following EC_METHODs:
7766 EC_GFp_nistp224_method()
7767 EC_GFp_nistp256_method()
7768 EC_GFp_nistp521_method()
7770 EC_GROUP_new_by_curve_name() will automatically use these (while
7771 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7774 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7776 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7777 all platforms. Move ssize_t definition from e_os.h to the public
7778 header file e_os2.h as it now appears in public header file cms.h
7782 * New -sigopt option to the ca, req and x509 utilities. Additional
7783 signature parameters can be passed using this option and in
7788 * Add RSA PSS signing function. This will generate and set the
7789 appropriate AlgorithmIdentifiers for PSS based on those in the
7790 corresponding EVP_MD_CTX structure. No application support yet.
7794 * Support for companion algorithm specific ASN1 signing routines.
7795 New function ASN1_item_sign_ctx() signs a pre-initialised
7796 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7797 the appropriate parameters.
7801 * Add new algorithm specific ASN1 verification initialisation function
7802 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7803 handling will be the same no matter what EVP_PKEY_METHOD is used.
7804 Add a PSS handler to support verification of PSS signatures: checked
7805 against a number of sample certificates.
7809 * Add signature printing for PSS. Add PSS OIDs.
7811 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7813 * Add algorithm specific signature printing. An individual ASN1 method
7814 can now print out signatures instead of the standard hex dump.
7816 More complex signatures (e.g. PSS) can print out more meaningful
7817 information. Include DSA version that prints out the signature
7822 * Password based recipient info support for CMS library: implementing
7827 * Split password based encryption into PBES2 and PBKDF2 functions. This
7828 neatly separates the code into cipher and PBE sections and is required
7829 for some algorithms that split PBES2 into separate pieces (such as
7830 password based CMS).
7834 * Session-handling fixes:
7835 - Fix handling of connections that are resuming with a session ID,
7836 but also support Session Tickets.
7837 - Fix a bug that suppressed issuing of a new ticket if the client
7838 presented a ticket with an expired session.
7839 - Try to set the ticket lifetime hint to something reasonable.
7840 - Make tickets shorter by excluding irrelevant information.
7841 - On the client side, don't ignore renewed tickets.
7843 *Adam Langley, Bodo Moeller (Google)*
7845 * Fix PSK session representation.
7849 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7851 This work was sponsored by Intel.
7855 * Add GCM support to TLS library. Some custom code is needed to split
7856 the IV between the fixed (from PRF) and explicit (from TLS record)
7857 portions. This adds all GCM ciphersuites supported by RFC5288 and
7858 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7859 add a special AESGCM string for GCM only.
7863 * Expand range of ctrls for AES GCM. Permit setting invocation
7864 field on decrypt and retrieval of invocation field only on encrypt.
7868 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7869 As required by RFC5289 these ciphersuites cannot be used if for
7870 versions of TLS earlier than 1.2.
7874 * For FIPS capable OpenSSL interpret a NULL default public key method
7875 as unset and return the appropriate default but do *not* set the default.
7876 This means we can return the appropriate method in applications that
7877 switch between FIPS and non-FIPS modes.
7881 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7882 ENGINE is used then we cannot handle that in the FIPS module so we
7883 keep original code iff non-FIPS operations are allowed.
7887 * Add -attime option to openssl utilities.
7889 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7891 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7895 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7896 FIPS EC methods unconditionally for now.
7900 * New build option no-ec2m to disable characteristic 2 code.
7904 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7905 all cases can be covered as some introduce binary incompatibilities.
7909 * Redirect RSA operations to FIPS module including keygen,
7910 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7914 * Add similar low-level API blocking to ciphers.
7918 * low-level digest APIs are not approved in FIPS mode: any attempt
7919 to use these will cause a fatal error. Applications that *really* want
7920 to use them can use the `private_*` version instead.
7924 * Redirect cipher operations to FIPS module for FIPS builds.
7928 * Redirect digest operations to FIPS module for FIPS builds.
7932 * Update build system to add "fips" flag which will link in fipscanister.o
7933 for static and shared library builds embedding a signature if needed.
7937 * Output TLS supported curves in preference order instead of numerical
7938 order. This is currently hardcoded for the highest order curves first.
7939 This should be configurable so applications can judge speed vs strength.
7943 * Add TLS v1.2 server support for client authentication.
7947 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7952 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7953 FIPS modules versions.
7957 * Add TLS v1.2 client side support for client authentication. Keep cache
7958 of handshake records longer as we don't know the hash algorithm to use
7959 until after the certificate request message is received.
7963 * Initial TLS v1.2 client support. Add a default signature algorithms
7964 extension including all the algorithms we support. Parse new signature
7965 format in client key exchange. Relax some ECC signing restrictions for
7966 TLS v1.2 as indicated in RFC5246.
7970 * Add server support for TLS v1.2 signature algorithms extension. Switch
7971 to new signature format when needed using client digest preference.
7972 All server ciphersuites should now work correctly in TLS v1.2. No client
7973 support yet and no support for client certificates.
7977 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7978 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7979 ciphersuites. At present only RSA key exchange ciphersuites work with
7980 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7981 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7982 and version checking.
7986 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7987 with this defined it will not be affected by any changes to ssl internal
7988 structures. Add several utility functions to allow openssl application
7989 to work with OPENSSL_NO_SSL_INTERN defined.
7993 * A long standing patch to add support for SRP from EdelWeb (Peter
7994 Sylvester and Christophe Renou) was integrated.
7995 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7996 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7999 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8003 * Permit abbreviated handshakes when renegotiating using the function
8004 SSL_renegotiate_abbreviated().
8006 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8008 * Add call to ENGINE_register_all_complete() to
8009 ENGINE_load_builtin_engines(), so some implementations get used
8010 automatically instead of needing explicit application support.
8014 * Add support for TLS key exporter as described in RFC5705.
8016 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8018 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8019 a few changes are required:
8021 Add SSL_OP_NO_TLSv1_1 flag.
8022 Add TLSv1_1 methods.
8023 Update version checking logic to handle version 1.1.
8024 Add explicit IV handling (ported from DTLS code).
8025 Add command line options to s_client/s_server.
8032 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8034 * X509_ATTRIBUTE memory leak
8036 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8037 memory. This structure is used by the PKCS#7 and CMS routines so any
8038 application which reads PKCS#7 or CMS data from untrusted sources is
8039 affected. SSL/TLS is not affected.
8041 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8047 * Race condition handling PSK identify hint
8049 If PSK identity hints are received by a multi-threaded client then
8050 the values are wrongly updated in the parent SSL_CTX structure. This can
8051 result in a race condition potentially leading to a double free of the
8057 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8059 * Malformed ECParameters causes infinite loop
8061 When processing an ECParameters structure OpenSSL enters an infinite loop
8062 if the curve specified is over a specially malformed binary polynomial
8065 This can be used to perform denial of service against any
8066 system which processes public keys, certificate requests or
8067 certificates. This includes TLS clients and TLS servers with
8068 client authentication enabled.
8070 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8075 * Exploitable out-of-bounds read in X509_cmp_time
8077 X509_cmp_time does not properly check the length of the ASN1_TIME
8078 string and can read a few bytes out of bounds. In addition,
8079 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8082 An attacker can use this to craft malformed certificates and CRLs of
8083 various sizes and potentially cause a segmentation fault, resulting in
8084 a DoS on applications that verify certificates or CRLs. TLS clients
8085 that verify CRLs are affected. TLS clients and servers with client
8086 authentication enabled may be affected if they use custom verification
8089 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8090 independently by Hanno Böck.
8095 * PKCS7 crash with missing EnvelopedContent
8097 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8098 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8099 with missing content and trigger a NULL pointer dereference on parsing.
8101 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8102 structures from untrusted sources are affected. OpenSSL clients and
8103 servers are not affected.
8105 This issue was reported to OpenSSL by Michal Zalewski (Google).
8110 * CMS verify infinite loop with unknown hash function
8112 When verifying a signedData message the CMS code can enter an infinite loop
8113 if presented with an unknown hash function OID. This can be used to perform
8114 denial of service against any system which verifies signedData messages using
8116 This issue was reported to OpenSSL by Johannes Bauer.
8121 * Race condition handling NewSessionTicket
8123 If a NewSessionTicket is received by a multi-threaded client when attempting to
8124 reuse a previous ticket then a race condition can occur potentially leading to
8125 a double free of the ticket data.
8130 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8132 * Segmentation fault in ASN1_TYPE_cmp fix
8134 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8135 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8136 certificate signature algorithm consistency this can be used to crash any
8137 certificate verification operation and exploited in a DoS attack. Any
8138 application which performs certificate verification is vulnerable including
8139 OpenSSL clients and servers which enable client authentication.
8144 * ASN.1 structure reuse memory corruption fix
8146 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8147 memory corruption via an invalid write. Such reuse is and has been
8148 strongly discouraged and is believed to be rare.
8150 Applications that parse structures containing CHOICE or ANY DEFINED BY
8151 components may be affected. Certificate parsing (d2i_X509 and related
8152 functions) are however not affected. OpenSSL clients and servers are
8158 * PKCS7 NULL pointer dereferences fix
8160 The PKCS#7 parsing code does not handle missing outer ContentInfo
8161 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8162 missing content and trigger a NULL pointer dereference on parsing.
8164 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8165 otherwise parse PKCS#7 structures from untrusted sources are
8166 affected. OpenSSL clients and servers are not affected.
8168 This issue was reported to OpenSSL by Michal Zalewski (Google).
8173 * DoS via reachable assert in SSLv2 servers fix
8175 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8176 servers that both support SSLv2 and enable export cipher suites by sending
8177 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8179 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8180 (OpenSSL development team).
8185 * Use After Free following d2i_ECPrivatekey error fix
8187 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8188 could cause a use after free condition. This, in turn, could cause a double
8189 free in several private key parsing functions (such as d2i_PrivateKey
8190 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8191 for applications that receive EC private keys from untrusted
8192 sources. This scenario is considered rare.
8194 This issue was discovered by the BoringSSL project and fixed in their
8200 * X509_to_X509_REQ NULL pointer deref fix
8202 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8203 the certificate key is invalid. This function is rarely used in practice.
8205 This issue was discovered by Brian Carpenter.
8210 * Removed the export ciphers from the DEFAULT ciphers
8214 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8216 * Build fixes for the Windows and OpenVMS platforms
8218 *Matt Caswell and Richard Levitte*
8220 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8222 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8223 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8224 dereference. This could lead to a Denial Of Service attack. Thanks to
8225 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8230 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8231 dtls1_buffer_record function under certain conditions. In particular this
8232 could occur if an attacker sent repeated DTLS records with the same
8233 sequence number but for the next epoch. The memory leak could be exploited
8234 by an attacker in a Denial of Service attack through memory exhaustion.
8235 Thanks to Chris Mueller for reporting this issue.
8240 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8241 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8242 method would be set to NULL which could later result in a NULL pointer
8243 dereference. Thanks to Frank Schmirler for reporting this issue.
8248 * Abort handshake if server key exchange message is omitted for ephemeral
8251 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8252 reporting this issue.
8257 * Remove non-export ephemeral RSA code on client and server. This code
8258 violated the TLS standard by allowing the use of temporary RSA keys in
8259 non-export ciphersuites and could be used by a server to effectively
8260 downgrade the RSA key length used to a value smaller than the server
8261 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8262 INRIA or reporting this issue.
8267 * Fixed issue where DH client certificates are accepted without verification.
8268 An OpenSSL server will accept a DH certificate for client authentication
8269 without the certificate verify message. This effectively allows a client to
8270 authenticate without the use of a private key. This only affects servers
8271 which trust a client certificate authority which issues certificates
8272 containing DH keys: these are extremely rare and hardly ever encountered.
8273 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8279 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8280 results on some platforms, including x86_64. This bug occurs at random
8281 with a very low probability, and is not known to be exploitable in any
8282 way, though its exact impact is difficult to determine. Thanks to Pieter
8283 Wuille (Blockstream) who reported this issue and also suggested an initial
8284 fix. Further analysis was conducted by the OpenSSL development team and
8285 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8286 the OpenSSL core team.
8291 * Fix various certificate fingerprint issues.
8293 By using non-DER or invalid encodings outside the signed portion of a
8294 certificate the fingerprint can be changed without breaking the signature.
8295 Although no details of the signed portion of the certificate can be changed
8296 this can cause problems with some applications: e.g. those using the
8297 certificate fingerprint for blacklists.
8299 1. Reject signatures with non zero unused bits.
8301 If the BIT STRING containing the signature has non zero unused bits reject
8302 the signature. All current signature algorithms require zero unused bits.
8304 2. Check certificate algorithm consistency.
8306 Check the AlgorithmIdentifier inside TBS matches the one in the
8307 certificate signature. NB: this will result in signature failure
8308 errors for some broken certificates.
8310 Thanks to Konrad Kraszewski from Google for reporting this issue.
8312 3. Check DSA/ECDSA signatures use DER.
8314 Reencode DSA/ECDSA signatures and compare with the original received
8315 signature. Return an error if there is a mismatch.
8317 This will reject various cases including garbage after signature
8318 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8319 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8320 (negative or with leading zeroes).
8322 Further analysis was conducted and fixes were developed by Stephen Henson
8323 of the OpenSSL core team.
8329 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8331 * Session Ticket Memory Leak.
8333 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8334 integrity of that ticket is first verified. In the event of a session
8335 ticket integrity check failing, OpenSSL will fail to free memory
8336 causing a memory leak. By sending a large number of invalid session
8337 tickets an attacker could exploit this issue in a Denial Of Service
8343 * Build option no-ssl3 is incomplete.
8345 When OpenSSL is configured with "no-ssl3" as a build option, servers
8346 could accept and complete an SSL 3.0 handshake, and clients could be
8347 configured to send them.
8350 *Akamai and the OpenSSL team*
8352 * Add support for TLS_FALLBACK_SCSV.
8353 Client applications doing fallback retries should call
8354 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8357 *Adam Langley, Bodo Moeller*
8359 * Add additional DigestInfo checks.
8361 Reencode DigestInto in DER and check against the original when
8362 verifying RSA signature: this will reject any improperly encoded
8363 DigestInfo structures.
8365 Note: this is a precautionary measure and no attacks are currently known.
8369 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8371 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8372 to a denial of service attack. A malicious server can crash the client
8373 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8374 ciphersuite and sending carefully crafted handshake messages.
8376 Thanks to Felix Gröbert (Google) for discovering and researching this
8382 * By sending carefully crafted DTLS packets an attacker could cause openssl
8383 to leak memory. This can be exploited through a Denial of Service attack.
8384 Thanks to Adam Langley for discovering and researching this issue.
8389 * An attacker can force openssl to consume large amounts of memory whilst
8390 processing DTLS handshake messages. This can be exploited through a
8391 Denial of Service attack.
8392 Thanks to Adam Langley for discovering and researching this issue.
8397 * An attacker can force an error condition which causes openssl to crash
8398 whilst processing DTLS packets due to memory being freed twice. This
8399 can be exploited through a Denial of Service attack.
8400 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8406 * If a multithreaded client connects to a malicious server using a resumed
8407 session and the server sends an ec point format extension it could write
8408 up to 255 bytes to freed memory.
8410 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8416 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8417 X509_name_oneline, X509_name_print_ex et al. to leak some information
8418 from the stack. Applications may be affected if they echo pretty printing
8419 output to the attacker.
8421 Thanks to Ivan Fratric (Google) for discovering this issue.
8424 *Emilia Käsper, and Steve Henson*
8426 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8427 for corner cases. (Certain input points at infinity could lead to
8428 bogus results, with non-infinity inputs mapped to infinity too.)
8432 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8434 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8435 handshake can force the use of weak keying material in OpenSSL
8436 SSL/TLS clients and servers.
8438 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8439 researching this issue. ([CVE-2014-0224])
8441 *KIKUCHI Masashi, Steve Henson*
8443 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8444 OpenSSL DTLS client the code can be made to recurse eventually crashing
8447 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8450 *Imre Rad, Steve Henson*
8452 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8453 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8454 client or server. This is potentially exploitable to run arbitrary
8455 code on a vulnerable client or server.
8457 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8459 *Jüri Aedla, Steve Henson*
8461 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8462 are subject to a denial of service attack.
8464 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8465 this issue. ([CVE-2014-3470])
8467 *Felix Gröbert, Ivan Fratric, Steve Henson*
8469 * Harmonize version and its documentation. -f flag is used to display
8472 *mancha <mancha1@zoho.com>*
8474 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8475 in i2d_ECPrivateKey.
8477 *mancha <mancha1@zoho.com>*
8479 * Fix some double frees. These are not thought to be exploitable.
8481 *mancha <mancha1@zoho.com>*
8483 * Fix for the attack described in the paper "Recovering OpenSSL
8484 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8485 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8486 <http://eprint.iacr.org/2014/140>
8488 Thanks to Yuval Yarom and Naomi Benger for discovering this
8489 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8491 *Yuval Yarom and Naomi Benger*
8493 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8495 * Keep original DTLS digest and encryption contexts in retransmission
8496 structures so we can use the previous session parameters if they need
8497 to be resent. ([CVE-2013-6450])
8501 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8502 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8503 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8504 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8505 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8506 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8508 *Rob Stradling, Adam Langley*
8510 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8512 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8514 This addresses the flaw in CBC record processing discovered by
8515 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8516 at: <http://www.isg.rhul.ac.uk/tls/>
8518 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8519 Security Group at Royal Holloway, University of London
8520 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8521 Emilia Käsper for the initial patch.
8524 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8526 * Return an error when checking OCSP signatures when key is NULL.
8527 This fixes a DoS attack. ([CVE-2013-0166])
8531 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8532 the right response is stapled. Also change SSL_get_certificate()
8533 so it returns the certificate actually sent.
8534 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8535 (This is a backport)
8537 *Rob Stradling <rob.stradling@comodo.com>*
8539 * Fix possible deadlock when decoding public keys.
8543 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8545 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8548 * Sanity check record length before skipping explicit IV in DTLS
8551 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8552 fuzzing as a service testing platform.
8557 * Initialise tkeylen properly when encrypting CMS messages.
8558 Thanks to Solar Designer of Openwall for reporting this issue.
8562 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8564 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8565 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8566 in CRYPTO_realloc_clean.
8568 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8569 issue and to Adam Langley <agl@chromium.org> for fixing it.
8572 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8574 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8576 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8577 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8578 content decryption and always return the same error. Note: this attack
8579 needs on average 2^20 messages so it only affects automated senders. The
8580 old behaviour can be re-enabled in the CMS code by setting the
8581 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8582 an MMA defence is not necessary.
8583 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8584 this issue. ([CVE-2012-0884])
8588 * Fix CVE-2011-4619: make sure we really are receiving a
8589 client hello before rejecting multiple SGC restarts. Thanks to
8590 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8594 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8596 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8597 Thanks to Antonio Martin, Enterprise Secure Access Research and
8598 Development, Cisco Systems, Inc. for discovering this bug and
8599 preparing a fix. ([CVE-2012-0050])
8603 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8605 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8606 of the Vaudenay padding oracle attack on CBC mode encryption
8607 which enables an efficient plaintext recovery attack against
8608 the OpenSSL implementation of DTLS. Their attack exploits timing
8609 differences arising during decryption processing. A research
8610 paper describing this attack can be found at:
8611 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8612 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8613 Security Group at Royal Holloway, University of London
8614 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8615 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8616 for preparing the fix. ([CVE-2011-4108])
8618 *Robin Seggelmann, Michael Tuexen*
8620 * Clear bytes used for block padding of SSL 3.0 records.
8623 *Adam Langley (Google)*
8625 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8626 Kadianakis <desnacked@gmail.com> for discovering this issue and
8627 Adam Langley for preparing the fix. ([CVE-2011-4619])
8629 *Adam Langley (Google)*
8631 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8633 *Andrey Kulikov <amdeich@gmail.com>*
8635 * Prevent malformed RFC3779 data triggering an assertion failure.
8636 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8637 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8639 *Rob Austein <sra@hactrn.net>*
8641 * Improved PRNG seeding for VOS.
8643 *Paul Green <Paul.Green@stratus.com>*
8645 * Fix ssl_ciph.c set-up race.
8647 *Adam Langley (Google)*
8649 * Fix spurious failures in ecdsatest.c.
8651 *Emilia Käsper (Google)*
8653 * Fix the BIO_f_buffer() implementation (which was mixing different
8654 interpretations of the `..._len` fields).
8656 *Adam Langley (Google)*
8658 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8659 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8660 threads won't reuse the same blinding coefficients.
8662 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8663 lock to call BN_BLINDING_invert_ex, and avoids one use of
8664 BN_BLINDING_update for each BN_BLINDING structure (previously,
8665 the last update always remained unused).
8667 *Emilia Käsper (Google)*
8669 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8671 *Bob Buckholz (Google)*
8673 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8675 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8676 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8678 *Kaspar Brand <ossl@velox.ch>*
8680 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8681 for multi-threaded use of ECDH. ([CVE-2011-3210])
8683 *Adam Langley (Google)*
8685 * Fix x509_name_ex_d2i memory leak on bad inputs.
8689 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8690 signature public key algorithm by using OID xref utilities instead.
8691 Before this you could only use some ECC ciphersuites with SHA1 only.
8695 * Add protection against ECDSA timing attacks as mentioned in the paper
8696 by Billy Bob Brumley and Nicola Tuveri, see:
8697 <http://eprint.iacr.org/2011/232.pdf>
8699 *Billy Bob Brumley and Nicola Tuveri*
8701 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8703 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8705 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8707 * Fix bug in string printing code: if *any* escaping is enabled we must
8708 escape the escape character (backslash) or the resulting string is
8713 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8715 * Disable code workaround for ancient and obsolete Netscape browsers
8716 and servers: an attacker can use it in a ciphersuite downgrade attack.
8717 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8721 * Fixed J-PAKE implementation error, originally discovered by
8722 Sebastien Martini, further info and confirmation from Stefan
8723 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8727 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8729 * Fix extension code to avoid race conditions which can result in a buffer
8730 overrun vulnerability: resumed sessions must not be modified as they can
8731 be shared by multiple threads. CVE-2010-3864
8735 * Fix WIN32 build system to correctly link an ENGINE directory into
8740 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8742 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8745 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8747 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8749 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8750 context. The operation can be customised via the ctrl mechanism in
8751 case ENGINEs want to include additional functionality.
8755 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8759 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8760 output hashes compatible with older versions of OpenSSL.
8762 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8764 * Fix compression algorithm handling: if resuming a session use the
8765 compression algorithm of the resumed session instead of determining
8766 it from client hello again. Don't allow server to change algorithm.
8770 * Add load_crls() function to commands tidying load_certs() too. Add option
8771 to verify utility to allow additional CRLs to be included.
8775 * Update OCSP request code to permit adding custom headers to the request:
8776 some responders need this.
8780 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8783 *Julia Lawall <julia@diku.dk>*
8785 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8786 needlessly dereferenced structures, used obsolete functions and
8787 didn't handle all updated verify codes correctly.
8791 * Disable MD2 in the default configuration.
8795 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8796 indicate the initial BIO being pushed or popped. This makes it possible
8797 to determine whether the BIO is the one explicitly called or as a result
8798 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8799 it handles reference counts correctly and doesn't zero out the I/O bio
8800 when it is not being explicitly popped. WARNING: applications which
8801 included workarounds for the old buggy behaviour will need to be modified
8802 or they could free up already freed BIOs.
8806 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8807 renaming to all platforms (within the 0.9.8 branch, this was
8808 done conditionally on Netware platforms to avoid a name clash).
8810 *Guenter <lists@gknw.net>*
8812 * Add ECDHE and PSK support to DTLS.
8814 *Michael Tuexen <tuexen@fh-muenster.de>*
8816 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8821 * Add "missing" function EVP_MD_flags() (without this the only way to
8822 retrieve a digest flags is by accessing the structure directly. Update
8823 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8824 or cipher is registered as in the "from" argument. Print out all
8825 registered digests in the dgst usage message instead of manually
8826 attempting to work them out.
8830 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8831 this allows the use of compression and extensions. Change default cipher
8832 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8833 by default unless an application cipher string requests it.
8837 * Alter match criteria in PKCS12_parse(). It used to try to use local
8838 key ids to find matching certificates and keys but some PKCS#12 files
8839 don't follow the (somewhat unwritten) rules and this strategy fails.
8840 Now just gather all certificates together and the first private key
8841 then look for the first certificate that matches the key.
8845 * Support use of registered digest and cipher names for dgst and cipher
8846 commands instead of having to add each one as a special case. So now
8853 openssl dgst -sha256 foo
8855 and this works for ENGINE based algorithms too.
8859 * Update Gost ENGINE to support parameter files.
8861 *Victor B. Wagner <vitus@cryptocom.ru>*
8863 * Support GeneralizedTime in ca utility.
8865 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8867 * Enhance the hash format used for certificate directory links. The new
8868 form uses the canonical encoding (meaning equivalent names will work
8869 even if they aren't identical) and uses SHA1 instead of MD5. This form
8870 is incompatible with the older format and as a result c_rehash should
8871 be used to rebuild symbolic links.
8875 * Make PKCS#8 the default write format for private keys, replacing the
8876 traditional format. This form is standardised, more secure and doesn't
8877 include an implicit MD5 dependency.
8881 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8882 committed to OpenSSL should pass this lot as a minimum.
8886 * Add session ticket override functionality for use by EAP-FAST.
8888 *Jouni Malinen <j@w1.fi>*
8890 * Modify HMAC functions to return a value. Since these can be implemented
8891 in an ENGINE errors can occur.
8895 * Type-checked OBJ_bsearch_ex.
8899 * Type-checked OBJ_bsearch. Also some constification necessitated
8900 by type-checking. Still to come: TXT_DB, bsearch(?),
8901 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8906 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8907 seconds to a tm structure directly, instead of going through OS
8908 specific date routines. This avoids any issues with OS routines such
8909 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8910 and X509_time_adj_ex() to cover the extended range. The existing
8911 X509_time_adj() is still usable and will no longer have any date issues.
8915 * Delta CRL support. New use deltas option which will attempt to locate
8916 and search any appropriate delta CRLs available.
8918 This work was sponsored by Google.
8922 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8923 code and add additional score elements. Validate alternate CRL paths
8924 as part of the CRL checking and indicate a new error "CRL path validation
8925 error" in this case. Applications wanting additional details can use
8926 the verify callback and check the new "parent" field. If this is not
8927 NULL CRL path validation is taking place. Existing applications won't
8928 see this because it requires extended CRL support which is off by
8931 This work was sponsored by Google.
8935 * Support for freshest CRL extension.
8937 This work was sponsored by Google.
8941 * Initial indirect CRL support. Currently only supported in the CRLs
8942 passed directly and not via lookup. Process certificate issuer
8943 CRL entry extension and lookup CRL entries by bother issuer name
8944 and serial number. Check and process CRL issuer entry in IDP extension.
8946 This work was sponsored by Google.
8950 * Add support for distinct certificate and CRL paths. The CRL issuer
8951 certificate is validated separately in this case. Only enabled if
8952 an extended CRL support flag is set: this flag will enable additional
8953 CRL functionality in future.
8955 This work was sponsored by Google.
8959 * Add support for policy mappings extension.
8961 This work was sponsored by Google.
8965 * Fixes to pathlength constraint, self issued certificate handling,
8966 policy processing to align with RFC3280 and PKITS tests.
8968 This work was sponsored by Google.
8972 * Support for name constraints certificate extension. DN, email, DNS
8973 and URI types are currently supported.
8975 This work was sponsored by Google.
8979 * To cater for systems that provide a pointer-based thread ID rather
8980 than numeric, deprecate the current numeric thread ID mechanism and
8981 replace it with a structure and associated callback type. This
8982 mechanism allows a numeric "hash" to be extracted from a thread ID in
8983 either case, and on platforms where pointers are larger than 'long',
8984 mixing is done to help ensure the numeric 'hash' is usable even if it
8985 can't be guaranteed unique. The default mechanism is to use "&errno"
8986 as a pointer-based thread ID to distinguish between threads.
8988 Applications that want to provide their own thread IDs should now use
8989 CRYPTO_THREADID_set_callback() to register a callback that will call
8990 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8992 Note that ERR_remove_state() is now deprecated, because it is tied
8993 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8994 to free the current thread's error state should be replaced by
8995 ERR_remove_thread_state(NULL).
8997 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8998 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8999 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9000 application was previously providing a numeric thread callback that
9001 was inappropriate for distinguishing threads, then uniqueness might
9002 have been obtained with &errno that happened immediately in the
9003 intermediate development versions of OpenSSL; this is no longer the
9004 case, the numeric thread callback will now override the automatic use
9007 *Geoff Thorpe, with help from Bodo Moeller*
9009 * Initial support for different CRL issuing certificates. This covers a
9010 simple case where the self issued certificates in the chain exist and
9011 the real CRL issuer is higher in the existing chain.
9013 This work was sponsored by Google.
9017 * Removed effectively defunct crypto/store from the build.
9021 * Revamp of STACK to provide stronger type-checking. Still to come:
9022 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9023 ASN1_STRING, CONF_VALUE.
9027 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9028 RAM on SSL connections. This option can save about 34k per idle SSL.
9032 * Revamp of LHASH to provide stronger type-checking. Still to come:
9033 STACK, TXT_DB, bsearch, qsort.
9037 * Initial support for Cryptographic Message Syntax (aka CMS) based
9038 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9039 support for data, signedData, compressedData, digestedData and
9040 encryptedData, envelopedData types included. Scripts to check against
9041 RFC4134 examples draft and interop and consistency checks of many
9042 content types and variants.
9046 * Add options to enc utility to support use of zlib compression BIO.
9050 * Extend mk1mf to support importing of options and assembly language
9051 files from Configure script, currently only included in VC-WIN32.
9052 The assembly language rules can now optionally generate the source
9053 files from the associated perl scripts.
9057 * Implement remaining functionality needed to support GOST ciphersuites.
9058 Interop testing has been performed using CryptoPro implementations.
9060 *Victor B. Wagner <vitus@cryptocom.ru>*
9062 * s390x assembler pack.
9066 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9071 * Implement Opaque PRF Input TLS extension as specified in
9072 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9073 official specification yet and no extension type assignment by
9074 IANA exists, this extension (for now) will have to be explicitly
9075 enabled when building OpenSSL by providing the extension number
9076 to use. For example, specify an option
9078 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9080 to the "config" or "Configure" script to enable the extension,
9081 assuming extension number 0x9527 (which is a completely arbitrary
9082 and unofficial assignment based on the MD5 hash of the Internet
9083 Draft). Note that by doing so, you potentially lose
9084 interoperability with other TLS implementations since these might
9085 be using the same extension number for other purposes.
9087 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9088 opaque PRF input value to use in the handshake. This will create
9089 an internal copy of the length-'len' string at 'src', and will
9090 return non-zero for success.
9092 To get more control and flexibility, provide a callback function
9095 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9096 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9100 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9103 Callback function 'cb' will be called in handshakes, and is
9104 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9105 Argument 'arg' is for application purposes (the value as given to
9106 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9107 be provided to the callback function). The callback function
9108 has to return non-zero to report success: usually 1 to use opaque
9109 PRF input just if possible, or 2 to enforce use of the opaque PRF
9110 input. In the latter case, the library will abort the handshake
9111 if opaque PRF input is not successfully negotiated.
9113 Arguments 'peerinput' and 'len' given to the callback function
9114 will always be NULL and 0 in the case of a client. A server will
9115 see the client's opaque PRF input through these variables if
9116 available (NULL and 0 otherwise). Note that if the server
9117 provides an opaque PRF input, the length must be the same as the
9118 length of the client's opaque PRF input.
9120 Note that the callback function will only be called when creating
9121 a new session (session resumption can resume whatever was
9122 previously negotiated), and will not be called in SSL 2.0
9123 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9124 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9125 for applications that need to enforce opaque PRF input.
9129 * Update ssl code to support digests other than SHA1+MD5 for handshake
9132 *Victor B. Wagner <vitus@cryptocom.ru>*
9134 * Add RFC4507 support to OpenSSL. This includes the corrections in
9135 RFC4507bis. The encrypted ticket format is an encrypted encoded
9136 SSL_SESSION structure, that way new session features are automatically
9139 If a client application caches session in an SSL_SESSION structure
9140 support is transparent because tickets are now stored in the encoded
9143 The SSL_CTX structure automatically generates keys for ticket
9144 protection in servers so again support should be possible
9145 with no application modification.
9147 If a client or server wishes to disable RFC4507 support then the option
9148 SSL_OP_NO_TICKET can be set.
9150 Add a TLS extension debugging callback to allow the contents of any client
9151 or server extensions to be examined.
9153 This work was sponsored by Google.
9157 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9158 OpenSSL should now compile cleanly on gcc 4.2
9160 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9162 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9163 support including streaming MAC support: this is required for GOST
9164 ciphersuite support.
9166 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9168 * Add option -stream to use PKCS#7 streaming in smime utility. New
9169 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9170 to output in BER and PEM format.
9174 * Experimental support for use of HMAC via EVP_PKEY interface. This
9175 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9176 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9177 ENGINE support for HMAC keys which are unextractable. New -mac and
9178 -macopt options to dgst utility.
9182 * New option -sigopt to dgst utility. Update dgst to use
9183 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9184 alternative signing parameters such as X9.31 or PSS in the dgst
9189 * Change ssl_cipher_apply_rule(), the internal function that does
9190 the work each time a ciphersuite string requests enabling
9191 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9192 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9193 the order of disabled ciphersuites such that those ciphersuites
9194 that most recently went from enabled to disabled not only stay
9195 in order with respect to each other, but also have higher priority
9196 than other disabled ciphersuites the next time ciphersuites are
9199 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9200 the same ciphersuites as with "HIGH" alone, but in a specific
9201 order where the PSK ciphersuites come first (since they are the
9202 most recently disabled ciphersuites when "HIGH" is parsed).
9204 Also, change ssl_create_cipher_list() (using this new
9205 functionality) such that between otherwise identical
9206 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9211 * Change ssl_create_cipher_list() so that it automatically
9212 arranges the ciphersuites in reasonable order before starting
9213 to process the rule string. Thus, the definition for "DEFAULT"
9214 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9215 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9216 This makes it much easier to arrive at a reasonable default order
9217 in applications for which anonymous ciphers are OK (meaning
9218 that you can't actually use DEFAULT).
9220 *Bodo Moeller; suggested by Victor Duchovni*
9222 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9223 processing) into multiple integers instead of setting
9224 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9225 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9226 (These masks as well as the individual bit definitions are hidden
9227 away into the non-exported interface ssl/ssl_locl.h, so this
9228 change to the definition of the SSL_CIPHER structure shouldn't
9229 affect applications.) This give us more bits for each of these
9230 categories, so there is no longer a need to coagulate AES128 and
9231 AES256 into a single algorithm bit, and to coagulate Camellia128
9232 and Camellia256 into a single algorithm bit, which has led to all
9235 Thus, among other things, the kludge introduced in 0.9.7m and
9236 0.9.8e for masking out AES256 independently of AES128 or masking
9237 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9239 With the change, we also introduce new ciphersuite aliases that
9240 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9245 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9246 Use the leftmost N bytes of the signature input if the input is
9247 larger than the prime q (with N being the size in bytes of q).
9251 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9252 it yet and it is largely untested.
9256 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9260 * Initial incomplete changes to avoid need for function casts in OpenSSL
9261 some compilers (gcc 4.2 and later) reject their use. Safestack is
9262 reimplemented. Update ASN1 to avoid use of legacy functions.
9266 * Win32/64 targets are linked with Winsock2.
9270 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9271 to external functions. This can be used to increase CRL handling
9272 efficiency especially when CRLs are very large by (for example) storing
9273 the CRL revoked certificates in a database.
9277 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9278 new CRLs added to a directory can be used. New command line option
9279 -verify_return_error to s_client and s_server. This causes real errors
9280 to be returned by the verify callback instead of carrying on no matter
9281 what. This reflects the way a "real world" verify callback would behave.
9285 * GOST engine, supporting several GOST algorithms and public key formats.
9286 Kindly donated by Cryptocom.
9290 * Partial support for Issuing Distribution Point CRL extension. CRLs
9291 partitioned by DP are handled but no indirect CRL or reason partitioning
9292 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9293 selected via a scoring technique which handles IDP and AKID in CRLs.
9297 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9298 will ultimately be used for all verify operations: this will remove the
9299 X509_STORE dependency on certificate verification and allow alternative
9300 lookup methods. X509_STORE based implementations of these two callbacks.
9304 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9305 Modify get_crl() to find a valid (unexpired) CRL if possible.
9309 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9310 this would be called X509_CRL_cmp() but that name is already used by
9311 a function that just compares CRL issuer names. Cache several CRL
9312 extensions in X509_CRL structure and cache CRLDP in X509.
9316 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9317 this maps equivalent X509_NAME structures into a consistent structure.
9318 Name comparison can then be performed rapidly using memcmp().
9322 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9327 * Allow digests to supply their own micalg string for S/MIME type using
9328 the ctrl EVP_MD_CTRL_MICALG.
9332 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9333 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9334 ctrl. It can then customise the structure before and/or after signing
9339 * New function OBJ_add_sigid() to allow application defined signature OIDs
9340 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9341 to free up any added signature OIDs.
9345 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9346 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9347 digest and cipher tables. New options added to openssl utility:
9348 list-message-digest-algorithms and list-cipher-algorithms.
9352 * Change the array representation of binary polynomials: the list
9353 of degrees of non-zero coefficients is now terminated with -1.
9354 Previously it was terminated with 0, which was also part of the
9355 value; thus, the array representation was not applicable to
9356 polynomials where t^0 has coefficient zero. This change makes
9357 the array representation useful in a more general context.
9361 * Various modifications and fixes to SSL/TLS cipher string
9362 handling. For ECC, the code now distinguishes between fixed ECDH
9363 with RSA certificates on the one hand and with ECDSA certificates
9364 on the other hand, since these are separate ciphersuites. The
9365 unused code for Fortezza ciphersuites has been removed.
9367 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9368 (not "ECDHE"). For consistency with the code for DH
9369 certificates, use of ECDH certificates is now considered ECDH
9370 authentication, not RSA or ECDSA authentication (the latter is
9371 merely the CA's signing algorithm and not actively used in the
9374 The temporary ciphersuite alias "ECCdraft" is no longer
9375 available, and ECC ciphersuites are no longer excluded from "ALL"
9376 and "DEFAULT". The following aliases now exist for RFC 4492
9377 ciphersuites, most of these by analogy with the DH case:
9379 kECDHr - ECDH cert, signed with RSA
9380 kECDHe - ECDH cert, signed with ECDSA
9381 kECDH - ECDH cert (signed with either RSA or ECDSA)
9382 kEECDH - ephemeral ECDH
9383 ECDH - ECDH cert or ephemeral ECDH
9389 AECDH - anonymous ECDH
9390 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9394 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9395 Use correct micalg parameters depending on digest(s) in signed message.
9399 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9400 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9404 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9405 an engine to register a method. Add ENGINE lookups for methods and
9406 functional reference processing.
9410 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9411 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9416 * New -resign option to smime utility. This adds one or more signers
9417 to an existing PKCS#7 signedData structure. Also -md option to use an
9418 alternative message digest algorithm for signing.
9422 * Tidy up PKCS#7 routines and add new functions to make it easier to
9423 create PKCS7 structures containing multiple signers. Update smime
9424 application to support multiple signers.
9428 * New -macalg option to pkcs12 utility to allow setting of an alternative
9433 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9434 Reorganize PBE internals to lookup from a static table using NIDs,
9435 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9436 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9437 PRF which will be automatically used with PBES2.
9441 * Replace the algorithm specific calls to generate keys in "req" with the
9446 * Update PKCS#7 enveloped data routines to use new API. This is now
9447 supported by any public key method supporting the encrypt operation. A
9448 ctrl is added to allow the public key algorithm to examine or modify
9449 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9454 * Add a ctrl to asn1 method to allow a public key algorithm to express
9455 a default digest type to use. In most cases this will be SHA1 but some
9456 algorithms (such as GOST) need to specify an alternative digest. The
9457 return value indicates how strong the preference is 1 means optional and
9458 2 is mandatory (that is it is the only supported type). Modify
9459 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9460 use the default md. Update openssl utilities to use the default digest
9461 type for signing if it is not explicitly indicated.
9465 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9466 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9467 signing method from the key type. This effectively removes the link
9468 between digests and public key types.
9472 * Add an OID cross reference table and utility functions. Its purpose is to
9473 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9474 rsaEncryption. This will allow some of the algorithm specific hackery
9475 needed to use the correct OID to be removed.
9479 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9480 structures for PKCS7_sign(). They are now set up by the relevant public
9485 * Add provisional EC pkey method with support for ECDSA and ECDH.
9489 * Add support for key derivation (agreement) in the API, DH method and
9494 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9495 public and private key formats. As a side effect these add additional
9496 command line functionality not previously available: DSA signatures can be
9497 generated and verified using pkeyutl and DH key support and generation in
9504 *Oliver Tappe <zooey@hirschkaefer.de>*
9506 * New make target "install_html_docs" installs HTML renditions of the
9509 *Oliver Tappe <zooey@hirschkaefer.de>*
9511 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9512 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9513 support key and parameter generation and add initial key generation
9514 functionality for RSA.
9518 * Add functions for main EVP_PKEY_method operations. The undocumented
9519 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9520 `EVP_PKEY_{encrypt,decrypt}_old`.
9524 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9525 key API, doesn't do much yet.
9529 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9530 public key algorithms. New option to openssl utility:
9531 "list-public-key-algorithms" to print out info.
9535 * Implement the Supported Elliptic Curves Extension for
9536 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9540 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9541 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9545 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9546 utilities such as rsa, dsa, dsaparam etc except they process any key
9551 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9552 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9553 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9558 * Initial support for pluggable public key ASN1.
9559 De-spaghettify the public key ASN1 handling. Move public and private
9560 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9561 algorithm specific handling to a single module within the relevant
9562 algorithm directory. Add functions to allow (near) opaque processing
9563 of public and private key structures.
9567 * Implement the Supported Point Formats Extension for
9568 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9572 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9573 for the psk identity [hint] and the psk callback functions to the
9574 SSL_SESSION, SSL and SSL_CTX structure.
9577 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9581 SSL_CTX_use_psk_identity_hint
9582 SSL_get_psk_identity_hint
9583 SSL_get_psk_identity
9584 SSL_use_psk_identity_hint
9586 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9588 * Add RFC 3161 compliant time stamp request creation, response generation
9589 and response verification functionality.
9591 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9593 * Add initial support for TLS extensions, specifically for the server_name
9594 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9595 have new members for a hostname. The SSL data structure has an
9596 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9597 stored in that context to allow for session resumption, even after the
9598 SSL has been switched to a new SSL_CTX in reaction to a client's
9599 server_name extension.
9601 New functions (subject to change):
9603 SSL_get_servername()
9604 SSL_get_servername_type()
9607 New CTRL codes and macros (subject to change):
9609 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9610 - SSL_CTX_set_tlsext_servername_callback()
9611 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9612 - SSL_CTX_set_tlsext_servername_arg()
9613 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9615 openssl s_client has a new '-servername ...' option.
9617 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9618 '-key2 ...', '-servername_fatal' (subject to change). This allows
9619 testing the HostName extension for a specific single hostname ('-cert'
9620 and '-key' remain fallbacks for handshakes without HostName
9621 negotiation). If the unrecognized_name alert has to be sent, this by
9622 default is a warning; it becomes fatal with the '-servername_fatal'
9625 *Peter Sylvester, Remy Allais, Christophe Renou*
9627 * Whirlpool hash implementation is added.
9631 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9632 bn(64,32). Because of instruction set limitations it doesn't have
9633 any negative impact on performance. This was done mostly in order
9634 to make it possible to share assembler modules, such as bn_mul_mont
9635 implementations, between 32- and 64-bit builds without hassle.
9639 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9640 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9645 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9646 dedicated Montgomery multiplication procedure, is introduced.
9647 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9648 "64-bit" performance on certain 32-bit targets.
9652 * New option SSL_OP_NO_COMP to disable use of compression selectively
9653 in SSL structures. New SSL ctrl to set maximum send fragment size.
9654 Save memory by setting the I/O buffer sizes dynamically instead of
9655 using the maximum available value.
9659 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9660 in addition to the text details.
9664 * Very, very preliminary EXPERIMENTAL support for printing of general
9665 ASN1 structures. This currently produces rather ugly output and doesn't
9666 handle several customised structures at all.
9670 * Integrated support for PVK file format and some related formats such
9671 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9672 these in the 'rsa' and 'dsa' utilities.
9676 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9680 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9681 place for the (very old) "NETSCAPE" format certificates which are now
9682 handled using new ASN1 code equivalents.
9686 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9687 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9688 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9692 * Modify CRL distribution points extension code to print out previously
9693 unsupported fields. Enhance extension setting code to allow setting of
9698 * Add print and set support for Issuing Distribution Point CRL extension.
9702 * Change 'Configure' script to enable Camellia by default.
9709 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9711 * When rejecting SSL/TLS records due to an incorrect version number, never
9712 update s->server with a new major version number. As of
9713 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9714 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9715 the previous behavior could result in a read attempt at NULL when
9716 receiving specific incorrect SSL/TLS records once record payload
9717 protection is active. ([CVE-2010-0740])
9719 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9721 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9722 could be crashed if the relevant tables were not present (e.g. chrooted).
9724 *Tomas Hoger <thoger@redhat.com>*
9726 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9728 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9730 *Martin Olsson, Neel Mehta*
9732 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9733 accommodate for stack sorting, always a write lock!).
9737 * On some versions of WIN32 Heap32Next is very slow. This can cause
9738 excessive delays in the RAND_poll(): over a minute. As a workaround
9739 include a time check in the inner Heap32Next loop too.
9743 * The code that handled flushing of data in SSL/TLS originally used the
9744 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9745 the problem outlined in PR#1949. The fix suggested there however can
9746 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9747 of Apache). So instead simplify the code to flush unconditionally.
9748 This should be fine since flushing with no data to flush is a no op.
9752 * Handle TLS versions 2.0 and later properly and correctly use the
9753 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9754 off ancient servers have a habit of sticking around for a while...
9758 * Modify compression code so it frees up structures without using the
9759 ex_data callbacks. This works around a problem where some applications
9760 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9761 restarting) then use compression (e.g. SSL with compression) later.
9762 This results in significant per-connection memory leaks and
9763 has caused some security issues including CVE-2008-1678 and
9768 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9769 change when encrypting or decrypting.
9773 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9774 connect and renegotiate with servers which do not support RI.
9775 Until RI is more widely deployed this option is enabled by default.
9779 * Add "missing" ssl ctrls to clear options and mode.
9783 * If client attempts to renegotiate and doesn't support RI respond with
9784 a no_renegotiation alert as required by RFC5746. Some renegotiating
9785 TLS clients will continue a connection gracefully when they receive
9786 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9787 waiting for a server hello which it will never receive. Now we treat a
9788 received no_renegotiation alert as a fatal error. This is because
9789 applications requesting a renegotiation might well expect it to succeed
9790 and would have no code in place to handle the server denying it so the
9791 only safe thing to do is to terminate the connection.
9795 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9796 peer supports secure renegotiation and 0 otherwise. Print out peer
9797 renegotiation support in s_client/s_server.
9801 * Replace the highly broken and deprecated SPKAC certification method with
9802 the updated NID creation version. This should correctly handle UTF8.
9806 * Implement RFC5746. Re-enable renegotiation but require the extension
9807 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9808 turns out to be a bad idea. It has been replaced by
9809 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9810 SSL_CTX_set_options(). This is really not recommended unless you
9811 know what you are doing.
9813 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9815 * Fixes to stateless session resumption handling. Use initial_ctx when
9816 issuing and attempting to decrypt tickets in case it has changed during
9817 servername handling. Use a non-zero length session ID when attempting
9818 stateless session resumption: this makes it possible to determine if
9819 a resumption has occurred immediately after receiving server hello
9820 (several places in OpenSSL subtly assume this) instead of later in
9825 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9826 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9827 fixes for a few places where the return code is not checked
9830 *Julia Lawall <julia@diku.dk>*
9832 * Add --strict-warnings option to Configure script to include devteam
9833 warnings in other configurations.
9837 * Add support for --libdir option and LIBDIR variable in makefiles. This
9838 makes it possible to install openssl libraries in locations which
9839 have names other than "lib", for example "/usr/lib64" which some
9842 *Steve Henson, based on patch from Jeremy Utley*
9844 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9845 X690 8.9.12 and can produce some misleading textual output of OIDs.
9847 *Steve Henson, reported by Dan Kaminsky*
9849 * Delete MD2 from algorithm tables. This follows the recommendation in
9850 several standards that it is not used in new applications due to
9851 several cryptographic weaknesses. For binary compatibility reasons
9852 the MD2 API is still compiled in by default.
9856 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9861 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9862 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9865 *Guenter <lists@gknw.net>*
9867 * Fix the server certificate chain building code to use X509_verify_cert(),
9868 it used to have an ad-hoc builder which was unable to cope with anything
9869 other than a simple chain.
9871 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9873 * Don't check self signed certificate signatures in X509_verify_cert()
9874 by default (a flag can override this): it just wastes time without
9875 adding any security. As a useful side effect self signed root CAs
9876 with non-FIPS digests are now usable in FIPS mode.
9880 * In dtls1_process_out_of_seq_message() the check if the current message
9881 is already buffered was missing. For every new message was memory
9882 allocated, allowing an attacker to perform an denial of service attack
9883 with sending out of seq handshake messages until there is no memory
9884 left. Additionally every future message was buffered, even if the
9885 sequence number made no sense and would be part of another handshake.
9886 So only messages with sequence numbers less than 10 in advance will be
9887 buffered. ([CVE-2009-1378])
9889 *Robin Seggelmann, discovered by Daniel Mentz*
9891 * Records are buffered if they arrive with a future epoch to be
9892 processed after finishing the corresponding handshake. There is
9893 currently no limitation to this buffer allowing an attacker to perform
9894 a DOS attack with sending records with future epochs until there is no
9895 memory left. This patch adds the pqueue_size() function to determine
9896 the size of a buffer and limits the record buffer to 100 entries.
9899 *Robin Seggelmann, discovered by Daniel Mentz*
9901 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9902 parent structure is freed. ([CVE-2009-1379])
9906 * Handle non-blocking I/O properly in SSL_shutdown() call.
9908 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9910 * Add `2.5.4.*` OIDs
9912 *Ilya O. <vrghost@gmail.com>*
9914 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9916 * Disable renegotiation completely - this fixes a severe security
9917 problem ([CVE-2009-3555]) at the cost of breaking all
9918 renegotiation. Renegotiation can be re-enabled by setting
9919 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9920 run-time. This is really not recommended unless you know what
9925 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9927 * Don't set val to NULL when freeing up structures, it is freed up by
9928 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9929 zeroing past the valid field. ([CVE-2009-0789])
9931 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9933 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9934 checked correctly. This would allow some invalid signed attributes to
9935 appear to verify correctly. ([CVE-2009-0591])
9937 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9939 * Reject UniversalString and BMPString types with invalid lengths. This
9940 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9941 a legal length. ([CVE-2009-0590])
9945 * Set S/MIME signing as the default purpose rather than setting it
9946 unconditionally. This allows applications to override it at the store
9951 * Permit restricted recursion of ASN1 strings. This is needed in practice
9952 to handle some structures.
9956 * Improve efficiency of mem_gets: don't search whole buffer each time
9959 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9961 * New -hex option for openssl rand.
9965 * Print out UTF8String and NumericString when parsing ASN1.
9969 * Support NumericString type for name components.
9973 * Allow CC in the environment to override the automatically chosen
9974 compiler. Note that nothing is done to ensure flags work with the
9979 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9981 * Properly check EVP_VerifyFinal() and similar return values
9984 *Ben Laurie, Bodo Moeller, Google Security Team*
9986 * Enable TLS extensions by default.
9990 * Allow the CHIL engine to be loaded, whether the application is
9991 multithreaded or not. (This does not release the developer from the
9992 obligation to set up the dynamic locking callbacks.)
9994 *Sander Temme <sander@temme.net>*
9996 * Use correct exit code if there is an error in dgst command.
9998 *Steve Henson; problem pointed out by Roland Dirlewanger*
10000 * Tweak Configure so that you need to say "experimental-jpake" to enable
10001 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10005 * Add experimental JPAKE support, including demo authentication in
10006 s_client and s_server.
10010 * Set the comparison function in v3_addr_canonize().
10012 *Rob Austein <sra@hactrn.net>*
10014 * Add support for XMPP STARTTLS in s_client.
10016 *Philip Paeps <philip@freebsd.org>*
10018 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10019 to ensure that even with this option, only ciphersuites in the
10020 server's preference list will be accepted. (Note that the option
10021 applies only when resuming a session, so the earlier behavior was
10022 just about the algorithm choice for symmetric cryptography.)
10026 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10028 * Fix NULL pointer dereference if a DTLS server received
10029 ChangeCipherSpec as first record ([CVE-2009-1386]).
10033 * Fix a state transition in s3_srvr.c and d1_srvr.c
10034 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10036 *Nagendra Modadugu*
10038 * The fix in 0.9.8c that supposedly got rid of unsafe
10039 double-checked locking was incomplete for RSA blinding,
10040 addressing just one layer of what turns out to have been
10041 doubly unsafe triple-checked locking.
10043 So now fix this for real by retiring the MONT_HELPER macro
10044 in crypto/rsa/rsa_eay.c.
10046 *Bodo Moeller; problem pointed out by Marius Schilder*
10048 * Various precautionary measures:
10050 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10052 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10053 (NB: This would require knowledge of the secret session ticket key
10054 to exploit, in which case you'd be SOL either way.)
10056 - Change bn_nist.c so that it will properly handle input BIGNUMs
10057 outside the expected range.
10059 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10062 *Neel Mehta, Bodo Moeller*
10064 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10065 the load fails. Useful for distros.
10067 *Ben Laurie and the FreeBSD team*
10069 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10073 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10077 * Expand ENGINE to support engine supplied SSL client certificate functions.
10079 This work was sponsored by Logica.
10083 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10084 keystores. Support for SSL/TLS client authentication too.
10085 Not compiled unless enable-capieng specified to Configure.
10087 This work was sponsored by Logica.
10091 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10092 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10093 attribute creation routines such as certificate requests and PKCS#12
10098 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10100 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10101 handshake which could lead to a client crash as found using the
10102 Codenomicon TLS test suite ([CVE-2008-1672])
10104 *Steve Henson, Mark Cox*
10106 * Fix double free in TLS server name extensions which could lead to
10107 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10111 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10113 Clear the error queue to ensure that error entries left from
10114 older function calls do not interfere with the correct operation.
10116 *Lutz Jaenicke, Erik de Castro Lopo*
10118 * Remove root CA certificates of commercial CAs:
10120 The OpenSSL project does not recommend any specific CA and does not
10121 have any policy with respect to including or excluding any CA.
10122 Therefore, it does not make any sense to ship an arbitrary selection
10123 of root CA certificates with the OpenSSL software.
10127 * RSA OAEP patches to fix two separate invalid memory reads.
10128 The first one involves inputs when 'lzero' is greater than
10129 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10130 before the beginning of from). The second one involves inputs where
10131 the 'db' section contains nothing but zeroes (there is a one-byte
10132 invalid read after the end of 'db').
10134 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10136 * Partial backport from 0.9.9-dev:
10138 Introduce bn_mul_mont (dedicated Montgomery multiplication
10139 procedure) as a candidate for BIGNUM assembler implementation.
10140 While 0.9.9-dev uses assembler for various architectures, only
10141 x86_64 is available by default here in the 0.9.8 branch, and
10142 32-bit x86 is available through a compile-time setting.
10144 To try the 32-bit x86 assembler implementation, use Configure
10145 option "enable-montasm" (which exists only for this backport).
10147 As "enable-montasm" for 32-bit x86 disclaims code stability
10148 anyway, in this constellation we activate additional code
10149 backported from 0.9.9-dev for further performance improvements,
10150 namely BN_from_montgomery_word. (To enable this otherwise,
10151 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10153 *Andy Polyakov (backport partially by Bodo Moeller)*
10155 * Add TLS session ticket callback. This allows an application to set
10156 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10157 values. This is useful for key rollover for example where several key
10158 sets may exist with different names.
10162 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10163 This was broken until now in 0.9.8 releases, such that the only way
10164 a registered ENGINE could be used (assuming it initialises
10165 successfully on the host) was to explicitly set it as the default
10166 for the relevant algorithms. This is in contradiction with 0.9.7
10167 behaviour and the documentation. With this fix, when an ENGINE is
10168 registered into a given algorithm's table of implementations, the
10169 'uptodate' flag is reset so that auto-discovery will be used next
10170 time a new context for that algorithm attempts to select an
10173 *Ian Lister (tweaked by Geoff Thorpe)*
10175 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10176 implementation in the following ways:
10178 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10181 Lack of BER streaming support means one pass streaming processing is
10182 only supported if data is detached: setting the streaming flag is
10183 ignored for embedded content.
10185 CMS support is disabled by default and must be explicitly enabled
10186 with the enable-cms configuration option.
10190 * Update the GMP engine glue to do direct copies between BIGNUM and
10191 mpz_t when openssl and GMP use the same limb size. Otherwise the
10192 existing "conversion via a text string export" trick is still used.
10194 *Paul Sheer <paulsheer@gmail.com>*
10196 * Zlib compression BIO. This is a filter BIO which compressed and
10197 uncompresses any data passed through it.
10201 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10202 RFC3394 compatible AES key wrapping.
10206 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10207 sets string data without copying. X509_ALGOR_set0() and
10208 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10209 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10210 from an X509_ATTRIBUTE structure optionally checking it occurs only
10211 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10216 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10217 to get the expected BN_FLG_CONSTTIME behavior.
10219 *Bodo Moeller (Google)*
10223 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10224 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10225 - added some more tests to do_tests.pl
10226 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10227 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10228 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10229 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10230 - various changes to netware.pl to enable gcc-cross builds on Win32
10232 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10233 - various changes to fix missing prototype warnings
10234 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10235 - added AES, WHIRLPOOL and CPUID assembler code to build files
10236 - added missing AES assembler make rules to mk1mf.pl
10237 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10239 *Guenter Knauf <eflash@gmx.net>*
10241 * Implement certificate status request TLS extension defined in RFC3546.
10242 A client can set the appropriate parameters and receive the encoded
10243 OCSP response via a callback. A server can query the supplied parameters
10244 and set the encoded OCSP response in the callback. Add simplified examples
10245 to s_client and s_server.
10249 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10251 * Fix various bugs:
10252 + Binary incompatibility of ssl_ctx_st structure
10253 + DTLS interoperation with non-compliant servers
10254 + Don't call get_session_cb() without proposed session
10255 + Fix ia64 assembler code
10257 *Andy Polyakov, Steve Henson*
10259 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10261 * DTLS Handshake overhaul. There were longstanding issues with
10262 OpenSSL DTLS implementation, which were making it impossible for
10263 RFC 4347 compliant client to communicate with OpenSSL server.
10264 Unfortunately just fixing these incompatibilities would "cut off"
10265 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10266 server keeps tolerating non RFC compliant syntax. The opposite is
10267 not true, 0.9.8f client can not communicate with earlier server.
10268 This update even addresses CVE-2007-4995.
10272 * Changes to avoid need for function casts in OpenSSL: some compilers
10273 (gcc 4.2 and later) reject their use.
10274 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10277 * Add RFC4507 support to OpenSSL. This includes the corrections in
10278 RFC4507bis. The encrypted ticket format is an encrypted encoded
10279 SSL_SESSION structure, that way new session features are automatically
10282 If a client application caches session in an SSL_SESSION structure
10283 support is transparent because tickets are now stored in the encoded
10286 The SSL_CTX structure automatically generates keys for ticket
10287 protection in servers so again support should be possible
10288 with no application modification.
10290 If a client or server wishes to disable RFC4507 support then the option
10291 SSL_OP_NO_TICKET can be set.
10293 Add a TLS extension debugging callback to allow the contents of any client
10294 or server extensions to be examined.
10296 This work was sponsored by Google.
10300 * Add initial support for TLS extensions, specifically for the server_name
10301 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10302 have new members for a hostname. The SSL data structure has an
10303 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10304 stored in that context to allow for session resumption, even after the
10305 SSL has been switched to a new SSL_CTX in reaction to a client's
10306 server_name extension.
10308 New functions (subject to change):
10310 SSL_get_servername()
10311 SSL_get_servername_type()
10314 New CTRL codes and macros (subject to change):
10316 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10317 - SSL_CTX_set_tlsext_servername_callback()
10318 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10319 - SSL_CTX_set_tlsext_servername_arg()
10320 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10322 openssl s_client has a new '-servername ...' option.
10324 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10325 '-key2 ...', '-servername_fatal' (subject to change). This allows
10326 testing the HostName extension for a specific single hostname ('-cert'
10327 and '-key' remain fallbacks for handshakes without HostName
10328 negotiation). If the unrecognized_name alert has to be sent, this by
10329 default is a warning; it becomes fatal with the '-servername_fatal'
10332 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10334 * Add AES and SSE2 assembly language support to VC++ build.
10338 * Mitigate attack on final subtraction in Montgomery reduction.
10342 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10343 (which previously caused an internal error).
10347 * Squeeze another 10% out of IGE mode when in != out.
10351 * AES IGE mode speedup.
10353 *Dean Gaudet (Google)*
10355 * Add the Korean symmetric 128-bit cipher SEED (see
10356 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10357 add SEED ciphersuites from RFC 4162:
10359 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10360 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10361 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10362 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10364 To minimize changes between patchlevels in the OpenSSL 0.9.8
10365 series, SEED remains excluded from compilation unless OpenSSL
10366 is configured with 'enable-seed'.
10368 *KISA, Bodo Moeller*
10370 * Mitigate branch prediction attacks, which can be practical if a
10371 single processor is shared, allowing a spy process to extract
10372 information. For detailed background information, see
10373 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10374 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10375 and Necessary Software Countermeasures"). The core of the change
10376 are new versions BN_div_no_branch() and
10377 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10378 respectively, which are slower, but avoid the security-relevant
10379 conditional branches. These are automatically called by BN_div()
10380 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10381 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10382 remove a conditional branch.
10384 BN_FLG_CONSTTIME is the new name for the previous
10385 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10386 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10387 in the exponent causes BN_mod_exp_mont() to use the alternative
10388 implementation in BN_mod_exp_mont_consttime().) The old name
10389 remains as a deprecated alias.
10391 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10392 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10393 constant-time implementations for more than just exponentiation.
10394 Here too the old name is kept as a deprecated alias.
10396 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10397 the BN_BLINDING structure gets an independent copy of the
10398 modulus. This means that the previous `BIGNUM *m` argument to
10399 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10400 essentially becomes `const BIGNUM *m`, although we can't actually
10401 change this in the header file before 0.9.9. It allows
10402 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10403 enable BN_FLG_CONSTTIME.
10405 *Matthew D Wood (Intel Corp)*
10407 * In the SSL/TLS server implementation, be strict about session ID
10408 context matching (which matters if an application uses a single
10409 external cache for different purposes). Previously,
10410 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10411 set. This did ensure strict client verification, but meant that,
10412 with applications using a single external cache for quite
10413 different requirements, clients could circumvent ciphersuite
10414 restrictions for a given session ID context by starting a session
10415 in a different context.
10419 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10420 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10421 authentication-only ciphersuites.
10425 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10426 not complete and could lead to a possible single byte overflow
10427 ([CVE-2007-5135]) [Ben Laurie]
10429 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10431 * Since AES128 and AES256 (and similarly Camellia128 and
10432 Camellia256) share a single mask bit in the logic of
10433 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10434 kludge to work properly if AES128 is available and AES256 isn't
10435 (or if Camellia128 is available and Camellia256 isn't).
10439 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10440 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10441 When a point or a seed is encoded in a BIT STRING, we need to
10442 prevent the removal of trailing zero bits to get the proper DER
10443 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10444 of a NamedBitList, for which trailing 0 bits need to be removed.)
10448 * Have SSL/TLS server implementation tolerate "mismatched" record
10449 protocol version while receiving ClientHello even if the
10450 ClientHello is fragmented. (The server can't insist on the
10451 particular protocol version it has chosen before the ServerHello
10452 message has informed the client about his choice.)
10456 * Add RFC 3779 support.
10458 *Rob Austein for ARIN, Ben Laurie*
10460 * Load error codes if they are not already present instead of using a
10461 static variable. This allows them to be cleanly unloaded and reloaded.
10462 Improve header file function name parsing.
10466 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10467 or CAPABILITY handshake as required by RFCs.
10469 *Goetz Babin-Ebell*
10471 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10473 * Introduce limits to prevent malicious keys being able to
10474 cause a denial of service. ([CVE-2006-2940])
10476 *Steve Henson, Bodo Moeller*
10478 * Fix ASN.1 parsing of certain invalid structures that can result
10479 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10481 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10482 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10484 * Fix SSL client code which could crash if connecting to a
10485 malicious SSLv2 server. ([CVE-2006-4343])
10487 *Tavis Ormandy and Will Drewry, Google Security Team*
10489 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10490 match only those. Before that, "AES256-SHA" would be interpreted
10491 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10492 the same strength classification in 0.9.7h) as we currently only
10493 have a single AES bit in the ciphersuite description bitmap.
10494 That change, however, also applied to ciphersuite strings such as
10495 "RC4-MD5" that intentionally matched multiple ciphersuites --
10496 namely, SSL 2.0 ciphersuites in addition to the more common ones
10497 from SSL 3.0/TLS 1.0.
10499 So we change the selection algorithm again: Naming an explicit
10500 ciphersuite selects this one ciphersuite, and any other similar
10501 ciphersuite (same bitmap) from *other* protocol versions.
10502 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10503 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10505 Since SSL 2.0 does not have any ciphersuites for which the
10506 128/256 bit distinction would be relevant, this works for now.
10507 The proper fix will be to use different bits for AES128 and
10508 AES256, which would have avoided the problems from the beginning;
10509 however, bits are scarce, so we can only do this in a new release
10510 (not just a patchlevel) when we can change the SSL_CIPHER
10511 definition to split the single 'unsigned long mask' bitmap into
10512 multiple values to extend the available space.
10516 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10518 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10519 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10521 * Add AES IGE and biIGE modes.
10525 * Change the Unix randomness entropy gathering to use poll() when
10526 possible instead of select(), since the latter has some
10527 undesirable limitations.
10529 *Darryl Miles via Richard Levitte and Bodo Moeller*
10531 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10532 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10533 cannot be implicitly activated as part of, e.g., the "AES" alias.
10534 However, please upgrade to OpenSSL 0.9.9[-dev] for
10535 non-experimental use of the ECC ciphersuites to get TLS extension
10536 support, which is required for curve and point format negotiation
10537 to avoid potential handshake problems.
10541 * Disable rogue ciphersuites:
10543 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10544 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10545 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10547 The latter two were purportedly from
10548 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10551 Also deactivate the remaining ciphersuites from
10552 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10553 unofficial, and the ID has long expired.
10557 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10558 dual-core machines) and other potential thread-safety issues.
10562 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10563 versions), which is now available for royalty-free use
10564 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10565 Also, add Camellia TLS ciphersuites from RFC 4132.
10567 To minimize changes between patchlevels in the OpenSSL 0.9.8
10568 series, Camellia remains excluded from compilation unless OpenSSL
10569 is configured with 'enable-camellia'.
10573 * Disable the padding bug check when compression is in use. The padding
10574 bug check assumes the first packet is of even length, this is not
10575 necessarily true if compression is enabled and can result in false
10576 positives causing handshake failure. The actual bug test is ancient
10577 code so it is hoped that implementations will either have fixed it by
10578 now or any which still have the bug do not support compression.
10582 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10584 * When applying a cipher rule check to see if string match is an explicit
10585 cipher suite and only match that one cipher suite if it is.
10589 * Link in manifests for VC++ if needed.
10591 *Austin Ziegler <halostatue@gmail.com>*
10593 * Update support for ECC-based TLS ciphersuites according to
10594 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10595 TLS extensions, which are supported starting with the 0.9.9
10596 branch, not in the OpenSSL 0.9.8 branch).
10600 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10601 opaque EVP_CIPHER_CTX handling.
10605 * Fixes and enhancements to zlib compression code. We now only use
10606 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10607 to conform with the standards mentioned here:
10608 <http://www.zlib.net/DLL_FAQ.txt>
10609 Static zlib linking now works on Windows and the new --with-zlib-include
10610 --with-zlib-lib options to Configure can be used to supply the location
10611 of the headers and library. Gracefully handle case where zlib library
10616 * Several fixes and enhancements to the OID generation code. The old code
10617 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10618 handle numbers larger than ULONG_MAX, truncated printing and had a
10619 non standard OBJ_obj2txt() behaviour.
10623 * Add support for building of engines under engine/ as shared libraries
10624 under VC++ build system.
10628 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10629 Hopefully, we will not see any false combination of paths any more.
10633 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10635 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10636 (part of SSL_OP_ALL). This option used to disable the
10637 countermeasure against man-in-the-middle protocol-version
10638 rollback in the SSL 2.0 server implementation, which is a bad
10639 idea. ([CVE-2005-2969])
10641 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10642 for Information Security, National Institute of Advanced Industrial
10643 Science and Technology [AIST], Japan)*
10645 * Add two function to clear and return the verify parameter flags.
10649 * Keep cipherlists sorted in the source instead of sorting them at
10650 runtime, thus removing the need for a lock.
10654 * Avoid some small subgroup attacks in Diffie-Hellman.
10656 *Nick Mathewson and Ben Laurie*
10658 * Add functions for well-known primes.
10662 * Extended Windows CE support.
10664 *Satoshi Nakamura and Andy Polyakov*
10666 * Initialize SSL_METHOD structures at compile time instead of during
10667 runtime, thus removing the need for a lock.
10671 * Make PKCS7_decrypt() work even if no certificate is supplied by
10672 attempting to decrypt each encrypted key in turn. Add support to
10677 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10679 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10682 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10686 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10687 key into the same file any more.
10691 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10695 * Add -utf8 command line and config file option to 'ca'.
10697 *Stefan <stf@udoma.org*
10699 * Removed the macro des_crypt(), as it seems to conflict with some
10700 libraries. Use DES_crypt().
10704 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10705 involves renaming the source and generated shared-libs for
10706 both. The engines will accept the corrected or legacy ids
10707 ('ncipher' and '4758_cca' respectively) when binding. NB,
10708 this only applies when building 'shared'.
10710 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10712 * Add attribute functions to EVP_PKEY structure. Modify
10713 PKCS12_create() to recognize a CSP name attribute and
10714 use it. Make -CSP option work again in pkcs12 utility.
10718 * Add new functionality to the bn blinding code:
10719 - automatic re-creation of the BN_BLINDING parameters after
10720 a fixed number of uses (currently 32)
10721 - add new function for parameter creation
10722 - introduce flags to control the update behaviour of the
10723 BN_BLINDING parameters
10724 - hide BN_BLINDING structure
10725 Add a second BN_BLINDING slot to the RSA structure to improve
10726 performance when a single RSA object is shared among several
10731 * Add support for DTLS.
10733 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10735 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10736 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10740 * Remove buggy and incomplete DH cert support from
10741 ssl/ssl_rsa.c and ssl/s3_both.c
10745 * Use SHA-1 instead of MD5 as the default digest algorithm for
10746 the `apps/openssl` commands.
10750 * Compile clean with "-Wall -Wmissing-prototypes
10751 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10752 DEBUG_SAFESTACK must also be set.
10756 * Change ./Configure so that certain algorithms can be disabled by default.
10757 The new counterpiece to "no-xxx" is "enable-xxx".
10759 The patented RC5 and MDC2 algorithms will now be disabled unless
10760 "enable-rc5" and "enable-mdc2", respectively, are specified.
10762 (IDEA remains enabled despite being patented. This is because IDEA
10763 is frequently required for interoperability, and there is no license
10764 fee for non-commercial use. As before, "no-idea" can be used to
10765 avoid this algorithm.)
10769 * Add processing of proxy certificates (see RFC 3820). This work was
10770 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10771 EGEE (Enabling Grids for E-science in Europe).
10775 * RC4 performance overhaul on modern architectures/implementations, such
10776 as Intel P4, IA-64 and AMD64.
10780 * New utility extract-section.pl. This can be used specify an alternative
10781 section number in a pod file instead of having to treat each file as
10782 a separate case in Makefile. This can be done by adding two lines to the
10785 =for comment openssl_section:XXX
10787 The blank line is mandatory.
10791 * New arguments -certform, -keyform and -pass for s_client and s_server
10792 to allow alternative format key and certificate files and passphrase
10797 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10798 update associated structures and add various utility functions.
10800 Add new policy related verify parameters, include policy checking in
10801 standard verify code. Enhance 'smime' application with extra parameters
10802 to support policy checking and print out.
10806 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10807 Nehemiah processors. These extensions support AES encryption in hardware
10808 as well as RNG (though RNG support is currently disabled).
10810 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10812 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10816 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10818 *Andy Polyakov and a number of other people*
10820 * Improved PowerPC platform support. Most notably BIGNUM assembler
10821 implementation contributed by IBM.
10823 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10825 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10826 exponent rather than 'unsigned long'. There is a corresponding change to
10827 the new 'rsa_keygen' element of the RSA_METHOD structure.
10829 *Jelte Jansen, Geoff Thorpe*
10831 * Functionality for creating the initial serial number file is now
10832 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10834 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10835 number file to 1, which is bound to cause problems. To avoid
10836 the problems while respecting compatibility between different 0.9.7
10837 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10838 CA.pl for serial number initialization. With the new release 0.9.8,
10839 we can fix the problem directly in the 'ca' utility.)
10843 * Reduced header interdependencies by declaring more opaque objects in
10844 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10845 give fewer recursive includes, which could break lazy source code - so
10846 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10847 developers should define this symbol when building and using openssl to
10848 ensure they track the recommended behaviour, interfaces, [etc], but
10849 backwards-compatible behaviour prevails when this isn't defined.
10853 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10857 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10858 This will generate a random key of the appropriate length based on the
10859 cipher context. The EVP_CIPHER can provide its own random key generation
10860 routine to support keys of a specific form. This is used in the des and
10861 3des routines to generate a key of the correct parity. Update S/MIME
10862 code to use new functions and hence generate correct parity DES keys.
10863 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10864 valid (weak or incorrect parity).
10868 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10869 as looking them up. This is useful when the verified structure may contain
10870 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10871 present unless the new PKCS7_NO_CRL flag is asserted.
10875 * Extend ASN1 oid configuration module. It now additionally accepts the
10878 shortName = some long name, 1.2.3.4
10882 * Reimplemented the BN_CTX implementation. There is now no more static
10883 limitation on the number of variables it can handle nor the depth of the
10884 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10885 information can now expand as required, and rather than having a single
10886 static array of bignums, BN_CTX now uses a linked-list of such arrays
10887 allowing it to expand on demand whilst maintaining the usefulness of
10888 BN_CTX's "bundling".
10892 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10893 to allow all RSA operations to function using a single BN_CTX.
10897 * Preliminary support for certificate policy evaluation and checking. This
10898 is initially intended to pass the tests outlined in "Conformance Testing
10899 of Relying Party Client Certificate Path Processing Logic" v1.07.
10903 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10904 remained unused and not that useful. A variety of other little bignum
10905 tweaks and fixes have also been made continuing on from the audit (see
10910 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10911 associated ASN1, EVP and SSL functions and old ASN1 macros.
10915 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10916 and this should never fail. So the return value from the use of
10917 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10918 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10922 * BN_CTX_get() should return zero-valued bignums, providing the same
10923 initialised value as BN_new().
10925 *Geoff Thorpe, suggested by Ulf Möller*
10927 * Support for inhibitAnyPolicy certificate extension.
10931 * An audit of the BIGNUM code is underway, for which debugging code is
10932 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10933 is considered valid when processing BIGNUMs, and causes execution to
10934 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10935 further steps are taken to deliberately pollute unused data in BIGNUM
10936 structures to try and expose faulty code further on. For now, openssl will
10937 (in its default mode of operation) continue to tolerate the inconsistent
10938 forms that it has tolerated in the past, but authors and packagers should
10939 consider trying openssl and their own applications when compiled with
10940 these debugging symbols defined. It will help highlight potential bugs in
10941 their own code, and will improve the test coverage for OpenSSL itself. At
10942 some point, these tighter rules will become openssl's default to improve
10943 maintainability, though the assert()s and other overheads will remain only
10944 in debugging configurations. See bn.h for more details.
10946 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10948 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10949 that can only be obtained through BN_CTX_new() (which implicitly
10950 initialises it). The presence of this function only made it possible
10951 to overwrite an existing structure (and cause memory leaks).
10955 * Because of the callback-based approach for implementing LHASH as a
10956 template type, lh_insert() adds opaque objects to hash-tables and
10957 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10958 to clean up those corresponding objects before destroying the hash table
10959 (and losing the object pointers). So some over-zealous constifications in
10960 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10961 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10962 prototyped to have "const" restrictions on the object pointers they are
10963 given (and so aren't required to cast them away any more).
10967 * The tmdiff.h API was so ugly and minimal that our own timing utility
10968 (speed) prefers to use its own implementation. The two implementations
10969 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10970 its object type properly exposed (MS_TM) instead of casting to/from
10971 `char *`. This may still change yet if someone realises MS_TM and
10973 aren't necessarily the greatest nomenclatures - but this is what was used
10974 internally to the implementation so I've used that for now.
10978 * Ensure that deprecated functions do not get compiled when
10979 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10980 the self-tests were still using deprecated key-generation functions so
10981 these have been updated also.
10985 * Reorganise PKCS#7 code to separate the digest location functionality
10986 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10987 New function PKCS7_set_digest() to set the digest type for PKCS#7
10988 digestedData type. Add additional code to correctly generate the
10989 digestedData type and add support for this type in PKCS7 initialization
10994 * New function PKCS7_set0_type_other() this initializes a PKCS7
10995 structure of type "other".
10999 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11000 sure the loop does correctly stop and breaking ("division by zero")
11001 modulus operations are not performed. The (pre-generated) prime
11002 table crypto/bn/bn_prime.h was already correct, but it could not be
11003 re-generated on some platforms because of the "division by zero"
11004 situation in the script.
11006 *Ralf S. Engelschall*
11008 * Update support for ECC-based TLS ciphersuites according to
11009 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11010 SHA-1 now is only used for "small" curves (where the
11011 representation of a field element takes up to 24 bytes); for
11012 larger curves, the field element resulting from ECDH is directly
11013 used as premaster secret.
11015 *Douglas Stebila (Sun Microsystems Laboratories)*
11017 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11018 curve secp160r1 to the tests.
11020 *Douglas Stebila (Sun Microsystems Laboratories)*
11022 * Add the possibility to load symbols globally with DSO.
11024 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11026 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11027 control of the error stack.
11031 * Add support for STORE in ENGINE.
11035 * Add the STORE type. The intention is to provide a common interface
11036 to certificate and key stores, be they simple file-based stores, or
11037 HSM-type store, or LDAP stores, or...
11038 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11042 * Add a generic structure called OPENSSL_ITEM. This can be used to
11043 pass a list of arguments to any function as well as provide a way
11044 for a function to pass data back to the caller.
11048 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11049 works like BUF_strdup() but can be used to duplicate a portion of
11050 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11055 * Add the function sk_find_ex() which works like sk_find(), but will
11056 return an index to an element even if an exact match couldn't be
11057 found. The index is guaranteed to point at the element where the
11058 searched-for key would be inserted to preserve sorting order.
11062 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11063 takes an extra flags argument for optional functionality. Currently,
11064 the following flags are defined:
11066 OBJ_BSEARCH_VALUE_ON_NOMATCH
11067 This one gets OBJ_bsearch_ex() to return a pointer to the first
11068 element where the comparing function returns a negative or zero
11071 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11072 This one gets OBJ_bsearch_ex() to return a pointer to the first
11073 element where the comparing function returns zero. This is useful
11074 if there are more than one element where the comparing function
11079 * Make it possible to create self-signed certificates with 'openssl ca'
11080 in such a way that the self-signed certificate becomes part of the
11081 CA database and uses the same mechanisms for serial number generation
11082 as all other certificate signing. The new flag '-selfsign' enables
11083 this functionality. Adapt CA.sh and CA.pl.in.
11087 * Add functionality to check the public key of a certificate request
11088 against a given private. This is useful to check that a certificate
11089 request can be signed by that key (self-signing).
11093 * Make it possible to have multiple active certificates with the same
11094 subject in the CA index file. This is done only if the keyword
11095 'unique_subject' is set to 'no' in the main CA section (default
11096 if 'CA_default') of the configuration file. The value is saved
11097 with the database itself in a separate index attribute file,
11098 named like the index file with '.attr' appended to the name.
11102 * Generate multi-valued AVAs using '+' notation in config files for
11107 * Support for nameConstraints certificate extension.
11111 * Support for policyConstraints certificate extension.
11115 * Support for policyMappings certificate extension.
11119 * Make sure the default DSA_METHOD implementation only uses its
11120 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11121 and change its own handlers to be NULL so as to remove unnecessary
11122 indirection. This lets alternative implementations fallback to the
11123 default implementation more easily.
11127 * Support for directoryName in GeneralName related extensions
11132 * Make it possible to link applications using Makefile.shared.
11133 Make that possible even when linking against static libraries!
11137 * Support for single pass processing for S/MIME signing. This now
11138 means that S/MIME signing can be done from a pipe, in addition
11139 cleartext signing (multipart/signed type) is effectively streaming
11140 and the signed data does not need to be all held in memory.
11142 This is done with a new flag PKCS7_STREAM. When this flag is set
11143 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11144 is done after the data is output (and digests calculated) in
11145 SMIME_write_PKCS7().
11149 * Add full support for -rpath/-R, both in shared libraries and
11150 applications, at least on the platforms where it's known how
11155 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11156 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11157 will now compute a table of multiples of the generator that
11158 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11159 faster (notably in the case of a single point multiplication,
11160 scalar * generator).
11162 *Nils Larsch, Bodo Moeller*
11164 * IPv6 support for certificate extensions. The various extensions
11165 which use the IP:a.b.c.d can now take IPv6 addresses using the
11166 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11171 * Added an ENGINE that implements RSA by performing private key
11172 exponentiations with the GMP library. The conversions to and from
11173 GMP's mpz_t format aren't optimised nor are any montgomery forms
11174 cached, and on x86 it appears OpenSSL's own performance has caught up.
11175 However there are likely to be other architectures where GMP could
11176 provide a boost. This ENGINE is not built in by default, but it can be
11177 specified at Configure time and should be accompanied by the necessary
11178 linker additions, eg;
11179 ./config -DOPENSSL_USE_GMP -lgmp
11183 * "openssl engine" will not display ENGINE/DSO load failure errors when
11184 testing availability of engines with "-t" - the old behaviour is
11185 produced by increasing the feature's verbosity with "-tt".
11189 * ECDSA routines: under certain error conditions uninitialized BN objects
11190 could be freed. Solution: make sure initialization is performed early
11191 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11196 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11197 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11198 software implementations. For DSA and DH, parameter generation can
11199 also be overridden by providing the appropriate method callbacks.
11203 * Change the "progress" mechanism used in key-generation and
11204 primality testing to functions that take a new BN_GENCB pointer in
11205 place of callback/argument pairs. The new API functions have `_ex`
11206 postfixes and the older functions are reimplemented as wrappers for
11207 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11208 declarations of the old functions to help (graceful) attempts to
11209 migrate to the new functions. Also, the new key-generation API
11210 functions operate on a caller-supplied key-structure and return
11211 success/failure rather than returning a key or NULL - this is to
11212 help make "keygen" another member function of RSA_METHOD etc.
11214 Example for using the new callback interface:
11216 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11217 void *my_arg = ...;
11220 BN_GENCB_set(&my_cb, my_callback, my_arg);
11222 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11223 /* For the meaning of a, b in calls to my_callback(), see the
11224 * documentation of the function that calls the callback.
11225 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11226 * my_callback should return 1 if it wants BN_is_prime_ex()
11227 * to continue, or 0 to stop.
11232 * Change the ZLIB compression method to be stateful, and make it
11233 available to TLS with the number defined in
11234 draft-ietf-tls-compression-04.txt.
11238 * Add the ASN.1 structures and functions for CertificatePair, which
11239 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11241 CertificatePair ::= SEQUENCE {
11242 forward [0] Certificate OPTIONAL,
11243 reverse [1] Certificate OPTIONAL,
11244 -- at least one of the pair shall be present -- }
11246 Also implement the PEM functions to read and write certificate
11247 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11249 This needed to be defined, mostly for the sake of the LDAP
11250 attribute crossCertificatePair, but may prove useful elsewhere as
11255 * Make it possible to inhibit symlinking of shared libraries in
11256 Makefile.shared, for Cygwin's sake.
11260 * Extend the BIGNUM API by creating a function
11261 void BN_set_negative(BIGNUM *a, int neg);
11262 and a macro that behave like
11263 int BN_is_negative(const BIGNUM *a);
11265 to avoid the need to access 'a->neg' directly in applications.
11269 * Implement fast modular reduction for pseudo-Mersenne primes
11270 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11271 EC_GROUP_new_curve_GFp() will now automatically use this
11274 *Nils Larsch <nla@trustcenter.de>*
11276 * Add new lock type (CRYPTO_LOCK_BN).
11280 * Change the ENGINE framework to automatically load engines
11281 dynamically from specific directories unless they could be
11282 found to already be built in or loaded. Move all the
11283 current engines except for the cryptodev one to a new
11284 directory engines/.
11285 The engines in engines/ are built as shared libraries if
11286 the "shared" options was given to ./Configure or ./config.
11287 Otherwise, they are inserted in libcrypto.a.
11288 /usr/local/ssl/engines is the default directory for dynamic
11289 engines, but that can be overridden at configure time through
11290 the usual use of --prefix and/or --openssldir, and at run
11291 time with the environment variable OPENSSL_ENGINES.
11293 *Geoff Thorpe and Richard Levitte*
11295 * Add Makefile.shared, a helper makefile to build shared
11296 libraries. Adapt Makefile.org.
11300 * Add version info to Win32 DLLs.
11302 *Peter 'Luna' Runestig" <peter@runestig.com>*
11304 * Add new 'medium level' PKCS#12 API. Certificates and keys
11305 can be added using this API to created arbitrary PKCS#12
11306 files while avoiding the low-level API.
11308 New options to PKCS12_create(), key or cert can be NULL and
11309 will then be omitted from the output file. The encryption
11310 algorithm NIDs can be set to -1 for no encryption, the mac
11311 iteration count can be set to 0 to omit the mac.
11313 Enhance pkcs12 utility by making the -nokeys and -nocerts
11314 options work when creating a PKCS#12 file. New option -nomac
11315 to omit the mac, NONE can be set for an encryption algorithm.
11316 New code is modified to use the enhanced PKCS12_create()
11317 instead of the low-level API.
11321 * Extend ASN1 encoder to support indefinite length constructed
11322 encoding. This can output sequences tags and octet strings in
11323 this form. Modify pk7_asn1.c to support indefinite length
11324 encoding. This is experimental and needs additional code to
11325 be useful, such as an ASN1 bio and some enhanced streaming
11328 Extend template encode functionality so that tagging is passed
11329 down to the template encoder.
11333 * Let 'openssl req' fail if an argument to '-newkey' is not
11334 recognized instead of using RSA as a default.
11338 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11339 As these are not official, they are not included in "ALL";
11340 the "ECCdraft" ciphersuite group alias can be used to select them.
11342 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11344 * Add ECDH engine support.
11346 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11348 * Add ECDH in new directory crypto/ecdh/.
11350 *Douglas Stebila (Sun Microsystems Laboratories)*
11352 * Let BN_rand_range() abort with an error after 100 iterations
11353 without success (which indicates a broken PRNG).
11357 * Change BN_mod_sqrt() so that it verifies that the input value
11358 is really the square of the return value. (Previously,
11359 BN_mod_sqrt would show GIGO behaviour.)
11363 * Add named elliptic curves over binary fields from X9.62, SECG,
11364 and WAP/WTLS; add OIDs that were still missing.
11366 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11368 * Extend the EC library for elliptic curves over binary fields
11369 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11372 EC_GF2m_simple_method
11376 EC_GROUP_new_curve_GF2m
11377 EC_GROUP_set_curve_GF2m
11378 EC_GROUP_get_curve_GF2m
11379 EC_POINT_set_affine_coordinates_GF2m
11380 EC_POINT_get_affine_coordinates_GF2m
11381 EC_POINT_set_compressed_coordinates_GF2m
11383 Point compression for binary fields is disabled by default for
11384 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11387 As binary polynomials are represented as BIGNUMs, various members
11388 of the EC_GROUP and EC_POINT data structures can be shared
11389 between the implementations for prime fields and binary fields;
11390 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11391 are essentially identical to their `..._GFp` counterparts.
11392 (For simplicity, the `..._GFp` prefix has been dropped from
11393 various internal method names.)
11395 An internal 'field_div' method (similar to 'field_mul' and
11396 'field_sqr') has been added; this is used only for binary fields.
11398 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11400 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11401 through methods ('mul', 'precompute_mult').
11403 The generic implementations (now internally called 'ec_wNAF_mul'
11404 and 'ec_wNAF_precomputed_mult') remain the default if these
11405 methods are undefined.
11407 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11409 * New function EC_GROUP_get_degree, which is defined through
11410 EC_METHOD. For curves over prime fields, this returns the bit
11411 length of the modulus.
11413 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11415 * New functions EC_GROUP_dup, EC_POINT_dup.
11416 (These simply call ..._new and ..._copy).
11418 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11420 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11421 Polynomials are represented as BIGNUMs (where the sign bit is not
11422 used) in the following functions [macros]:
11425 BN_GF2m_sub [= BN_GF2m_add]
11426 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11427 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11428 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11430 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11431 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11432 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11433 BN_GF2m_cmp [= BN_ucmp]
11435 (Note that only the 'mod' functions are actually for fields GF(2^m).
11436 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11438 For some functions, an the irreducible polynomial defining a
11439 field can be given as an 'unsigned int[]' with strictly
11440 decreasing elements giving the indices of those bits that are set;
11441 i.e., p[] represents the polynomial
11442 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11444 p[0] > p[1] > ... > p[k] = 0.
11445 This applies to the following functions:
11448 BN_GF2m_mod_mul_arr
11449 BN_GF2m_mod_sqr_arr
11450 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11451 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11452 BN_GF2m_mod_exp_arr
11453 BN_GF2m_mod_sqrt_arr
11454 BN_GF2m_mod_solve_quad_arr
11458 Conversion can be performed by the following functions:
11463 bntest.c has additional tests for binary polynomial arithmetic.
11465 Two implementations for BN_GF2m_mod_div() are available.
11466 The default algorithm simply uses BN_GF2m_mod_inv() and
11467 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11468 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11469 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11471 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11473 * Add new error code 'ERR_R_DISABLED' that can be used when some
11474 functionality is disabled at compile-time.
11476 *Douglas Stebila <douglas.stebila@sun.com>*
11478 * Change default behaviour of 'openssl asn1parse' so that more
11479 information is visible when viewing, e.g., a certificate:
11481 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11482 mode the content of non-printable OCTET STRINGs is output in a
11483 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11484 avoid the appearance of a printable string.
11486 *Nils Larsch <nla@trustcenter.de>*
11488 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11490 EC_GROUP_set_asn1_flag()
11491 EC_GROUP_get_asn1_flag()
11492 EC_GROUP_set_point_conversion_form()
11493 EC_GROUP_get_point_conversion_form()
11494 These control ASN1 encoding details:
11495 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11496 has been set to OPENSSL_EC_NAMED_CURVE.
11497 - Points are encoded in uncompressed form by default; options for
11498 asn1_for are as for point2oct, namely
11499 POINT_CONVERSION_COMPRESSED
11500 POINT_CONVERSION_UNCOMPRESSED
11501 POINT_CONVERSION_HYBRID
11503 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11505 EC_GROUP_set_seed()
11506 EC_GROUP_get0_seed()
11507 EC_GROUP_get_seed_len()
11508 This is used only for ASN1 purposes (so far).
11510 *Nils Larsch <nla@trustcenter.de>*
11512 * Add 'field_type' member to EC_METHOD, which holds the NID
11513 of the appropriate field type OID. The new function
11514 EC_METHOD_get_field_type() returns this value.
11516 *Nils Larsch <nla@trustcenter.de>*
11519 EC_POINT_point2bn()
11520 EC_POINT_bn2point()
11521 EC_POINT_point2hex()
11522 EC_POINT_hex2point()
11523 providing useful interfaces to EC_POINT_point2oct() and
11524 EC_POINT_oct2point().
11526 *Nils Larsch <nla@trustcenter.de>*
11528 * Change internals of the EC library so that the functions
11529 EC_GROUP_set_generator()
11530 EC_GROUP_get_generator()
11531 EC_GROUP_get_order()
11532 EC_GROUP_get_cofactor()
11533 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11534 to methods, which would lead to unnecessary code duplication when
11535 adding different types of curves.
11537 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11539 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11540 arithmetic, and such that modified wNAFs are generated
11541 (which avoid length expansion in many cases).
11545 * Add a function EC_GROUP_check_discriminant() (defined via
11546 EC_METHOD) that verifies that the curve discriminant is non-zero.
11548 Add a function EC_GROUP_check() that makes some sanity tests
11549 on a EC_GROUP, its generator and order. This includes
11550 EC_GROUP_check_discriminant().
11552 *Nils Larsch <nla@trustcenter.de>*
11554 * Add ECDSA in new directory crypto/ecdsa/.
11556 Add applications 'openssl ecparam' and 'openssl ecdsa'
11557 (these are based on 'openssl dsaparam' and 'openssl dsa').
11559 ECDSA support is also included in various other files across the
11560 library. Most notably,
11561 - 'openssl req' now has a '-newkey ecdsa:file' option;
11562 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11563 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11564 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11565 them suitable for ECDSA where domain parameters must be
11566 extracted before the specific public key;
11567 - ECDSA engine support has been added.
11569 *Nils Larsch <nla@trustcenter.de>*
11571 * Include some named elliptic curves, and add OIDs from X9.62,
11572 SECG, and WAP/WTLS. Each curve can be obtained from the new
11574 EC_GROUP_new_by_curve_name(),
11575 and the list of available named curves can be obtained with
11576 EC_get_builtin_curves().
11577 Also add a 'curve_name' member to EC_GROUP objects, which can be
11579 EC_GROUP_set_curve_name()
11580 EC_GROUP_get_curve_name()
11582 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11584 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11585 was actually never needed) and in BN_mul(). The removal in BN_mul()
11586 required a small change in bn_mul_part_recursive() and the addition
11587 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11588 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11589 bn_sub_words() and bn_add_words() except they take arrays with
11594 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11596 * Cleanse PEM buffers before freeing them since they may contain
11599 *Benjamin Bennett <ben@psc.edu>*
11601 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11602 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11603 authentication-only ciphersuites.
11607 * Since AES128 and AES256 share a single mask bit in the logic of
11608 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11609 kludge to work properly if AES128 is available and AES256 isn't.
11613 * Expand security boundary to match 1.1.1 module.
11617 * Remove redundant features: hash file source, editing of test vectors
11618 modify fipsld to use external fips_premain.c signature.
11622 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11623 run algorithm test programs.
11627 * Make algorithm test programs more tolerant of whitespace.
11631 * Have SSL/TLS server implementation tolerate "mismatched" record
11632 protocol version while receiving ClientHello even if the
11633 ClientHello is fragmented. (The server can't insist on the
11634 particular protocol version it has chosen before the ServerHello
11635 message has informed the client about his choice.)
11639 * Load error codes if they are not already present instead of using a
11640 static variable. This allows them to be cleanly unloaded and reloaded.
11644 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11646 * Introduce limits to prevent malicious keys being able to
11647 cause a denial of service. ([CVE-2006-2940])
11649 *Steve Henson, Bodo Moeller*
11651 * Fix ASN.1 parsing of certain invalid structures that can result
11652 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11654 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11655 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11657 * Fix SSL client code which could crash if connecting to a
11658 malicious SSLv2 server. ([CVE-2006-4343])
11660 *Tavis Ormandy and Will Drewry, Google Security Team*
11662 * Change ciphersuite string processing so that an explicit
11663 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11664 will no longer include "AES128-SHA"), and any other similar
11665 ciphersuite (same bitmap) from *other* protocol versions (so that
11666 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11667 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11668 changes from 0.9.8b and 0.9.8d.
11672 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11674 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11675 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11677 * Change the Unix randomness entropy gathering to use poll() when
11678 possible instead of select(), since the latter has some
11679 undesirable limitations.
11681 *Darryl Miles via Richard Levitte and Bodo Moeller*
11683 * Disable rogue ciphersuites:
11685 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11686 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11687 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11689 The latter two were purportedly from
11690 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11693 Also deactivate the remaining ciphersuites from
11694 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11695 unofficial, and the ID has long expired.
11699 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11700 dual-core machines) and other potential thread-safety issues.
11704 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11706 * Adapt fipsld and the build system to link against the validated FIPS
11707 module in FIPS mode.
11711 * Fixes for VC++ 2005 build under Windows.
11715 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11716 from a Windows bash shell such as MSYS. It is autodetected from the
11717 "config" script when run from a VC++ environment. Modify standard VC++
11718 build to use fipscanister.o from the GNU make build.
11722 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11724 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11725 The value now differs depending on if you build for FIPS or not.
11726 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11727 safely run with a non-FIPSed libcrypto, as it may crash because of
11728 the difference induced by this change.
11732 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11734 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11735 (part of SSL_OP_ALL). This option used to disable the
11736 countermeasure against man-in-the-middle protocol-version
11737 rollback in the SSL 2.0 server implementation, which is a bad
11738 idea. ([CVE-2005-2969])
11740 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11741 for Information Security, National Institute of Advanced Industrial
11742 Science and Technology [AIST, Japan)]*
11744 * Minimal support for X9.31 signatures and PSS padding modes. This is
11745 mainly for FIPS compliance and not fully integrated at this stage.
11749 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11750 the exponentiation using a fixed-length exponent. (Otherwise,
11751 the information leaked through timing could expose the secret key
11752 after many signatures; cf. Bleichenbacher's attack on DSA with
11757 * Make a new fixed-window mod_exp implementation the default for
11758 RSA, DSA, and DH private-key operations so that the sequence of
11759 squares and multiplies and the memory access pattern are
11760 independent of the particular secret key. This will mitigate
11761 cache-timing and potential related attacks.
11763 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11764 and this is automatically used by BN_mod_exp_mont() if the new flag
11765 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11766 will use this BN flag for private exponents unless the flag
11767 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11768 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11770 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11772 * Change the client implementation for SSLv23_method() and
11773 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11774 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11775 (Previously, the SSL 2.0 backwards compatible Client Hello
11776 message format would be used even with SSL_OP_NO_SSLv2.)
11780 * Add support for smime-type MIME parameter in S/MIME messages which some
11785 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11786 a threadsafe manner. Modify rsa code to use new function and add calls
11787 to dsa and dh code (which had race conditions before).
11791 * Include the fixed error library code in the C error file definitions
11792 instead of fixing them up at runtime. This keeps the error code
11793 structures constant.
11797 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11799 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11802 * Fixes for newer kerberos headers. NB: the casts are needed because
11803 the 'length' field is signed on one version and unsigned on another
11804 with no (?) obvious way to tell the difference, without these VC++
11805 complains. Also the "definition" of FAR (blank) is no longer included
11806 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11807 some needed definitions.
11811 * Undo Cygwin change.
11815 * Added support for proxy certificates according to RFC 3820.
11816 Because they may be a security thread to unaware applications,
11817 they must be explicitly allowed in run-time. See
11818 docs/HOWTO/proxy_certificates.txt for further information.
11822 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11824 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11825 server and client random values. Previously
11826 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11827 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11829 This change has negligible security impact because:
11831 1. Server and client random values still have 24 bytes of pseudo random
11834 2. Server and client random values are sent in the clear in the initial
11837 3. The master secret is derived using the premaster secret (48 bytes in
11838 size for static RSA ciphersuites) as well as client server and random
11841 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11844 *Stephen Henson, reported by UK NISCC*
11846 * Use Windows randomness collection on Cygwin.
11850 * Fix hang in EGD/PRNGD query when communication socket is closed
11851 prematurely by EGD/PRNGD.
11853 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11855 * Prompt for pass phrases when appropriate for PKCS12 input format.
11859 * Back-port of selected performance improvements from development
11860 branch, as well as improved support for PowerPC platforms.
11864 * Add lots of checks for memory allocation failure, error codes to indicate
11865 failure and freeing up memory if a failure occurs.
11867 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11869 * Add new -passin argument to dgst.
11873 * Perform some character comparisons of different types in X509_NAME_cmp:
11874 this is needed for some certificates that re-encode DNs into UTF8Strings
11875 (in violation of RFC3280) and can't or won't issue name rollover
11880 * Make an explicit check during certificate validation to see that
11881 the CA setting in each certificate on the chain is correct. As a
11882 side effect always do the following basic checks on extensions,
11883 not just when there's an associated purpose to the check:
11885 - if there is an unhandled critical extension (unless the user
11886 has chosen to ignore this fault)
11887 - if the path length has been exceeded (if one is set at all)
11888 - that certain extensions fit the associated purpose (if one has
11893 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11895 * Avoid a race condition when CRLs are checked in a multi threaded
11896 environment. This would happen due to the reordering of the revoked
11897 entries during signature checking and serial number lookup. Now the
11898 encoding is cached and the serial number sort performed under a lock.
11899 Add new STACK function sk_is_sorted().
11903 * Add Delta CRL to the extension code.
11907 * Various fixes to s3_pkt.c so alerts are sent properly.
11909 *David Holmes <d.holmes@f5.com>*
11911 * Reduce the chances of duplicate issuer name and serial numbers (in
11912 violation of RFC3280) using the OpenSSL certificate creation utilities.
11913 This is done by creating a random 64 bit value for the initial serial
11914 number when a serial number file is created or when a self signed
11915 certificate is created using 'openssl req -x509'. The initial serial
11916 number file is created using 'openssl x509 -next_serial' in CA.pl
11917 rather than being initialized to 1.
11921 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11923 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11924 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11926 *Joe Orton, Steve Henson*
11928 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11931 *Joe Orton, Steve Henson*
11933 * Make it possible to have multiple active certificates with the same
11934 subject in the CA index file. This is done only if the keyword
11935 'unique_subject' is set to 'no' in the main CA section (default
11936 if 'CA_default') of the configuration file. The value is saved
11937 with the database itself in a separate index attribute file,
11938 named like the index file with '.attr' appended to the name.
11942 * X509 verify fixes. Disable broken certificate workarounds when
11943 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11944 keyUsage extension present. Don't accept CRLs with unhandled critical
11945 extensions: since verify currently doesn't process CRL extensions this
11946 rejects a CRL with *any* critical extensions. Add new verify error codes
11951 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11952 A clarification of RFC2560 will require the use of OCTET STRINGs and
11953 some implementations cannot handle the current raw format. Since OpenSSL
11954 copies and compares OCSP nonces as opaque blobs without any attempt at
11955 parsing them this should not create any compatibility issues.
11959 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11960 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11961 this HMAC (and other) operations are several times slower than OpenSSL
11966 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11968 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11970 * Use the correct content when signing type "other".
11974 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11976 * Fix various bugs revealed by running the NISCC test suite:
11978 Stop out of bounds reads in the ASN1 code when presented with
11979 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11981 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11983 If verify callback ignores invalid public key errors don't try to check
11984 certificate signature with the NULL public key.
11988 * New -ignore_err option in ocsp application to stop the server
11989 exiting on the first error in a request.
11993 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11994 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11999 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12000 extra data after the compression methods not only for TLS 1.0
12001 but also for SSL 3.0 (as required by the specification).
12003 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12005 * Change X509_certificate_type() to mark the key as exported/exportable
12006 when it's 512 *bits* long, not 512 bytes.
12010 * Change AES_cbc_encrypt() so it outputs exact multiple of
12011 blocks during encryption.
12015 * Various fixes to base64 BIO and non blocking I/O. On write
12016 flushes were not handled properly if the BIO retried. On read
12017 data was not being buffered properly and had various logic bugs.
12018 This also affects blocking I/O when the data being decoded is a
12023 * Various S/MIME bugfixes and compatibility changes:
12024 output correct application/pkcs7 MIME type if
12025 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12026 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12027 of files as .eml work). Correctly handle very long lines in MIME
12032 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12034 * Countermeasure against the Klima-Pokorny-Rosa extension of
12035 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12036 a protocol version number mismatch like a decryption error
12037 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12041 * Turn on RSA blinding by default in the default implementation
12042 to avoid a timing attack. Applications that don't want it can call
12043 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12044 They would be ill-advised to do so in most cases.
12046 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12048 * Change RSA blinding code so that it works when the PRNG is not
12049 seeded (in this case, the secret RSA exponent is abused as
12050 an unpredictable seed -- if it is not unpredictable, there
12051 is no point in blinding anyway). Make RSA blinding thread-safe
12052 by remembering the creator's thread ID in rsa->blinding and
12053 having all other threads use local one-time blinding factors
12054 (this requires more computation than sharing rsa->blinding, but
12055 avoids excessive locking; and if an RSA object is not shared
12056 between threads, blinding will still be very fast).
12060 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12061 ENGINE as defaults for all supported algorithms irrespective of
12062 the 'flags' parameter. 'flags' is now honoured, so applications
12063 should make sure they are passing it correctly.
12067 * Target "mingw" now allows native Windows code to be generated in
12068 the Cygwin environment as well as with the MinGW compiler.
12072 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12074 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12075 via timing by performing a MAC computation even if incorrect
12076 block cipher padding has been found. This is a countermeasure
12077 against active attacks where the attacker has to distinguish
12078 between bad padding and a MAC verification error. ([CVE-2003-0078])
12080 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12081 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12082 Martin Vuagnoux (EPFL, Ilion)*
12084 * Make the no-err option work as intended. The intention with no-err
12085 is not to have the whole error stack handling routines removed from
12086 libcrypto, it's only intended to remove all the function name and
12087 reason texts, thereby removing some of the footprint that may not
12088 be interesting if those errors aren't displayed anyway.
12090 NOTE: it's still possible for any application or module to have its
12091 own set of error texts inserted. The routines are there, just not
12092 used by default when no-err is given.
12096 * Add support for FreeBSD on IA64.
12098 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12100 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12101 Kerberos function mit_des_cbc_cksum(). Before this change,
12102 the value returned by DES_cbc_cksum() was like the one from
12103 mit_des_cbc_cksum(), except the bytes were swapped.
12105 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12107 * Allow an application to disable the automatic SSL chain building.
12108 Before this a rather primitive chain build was always performed in
12109 ssl3_output_cert_chain(): an application had no way to send the
12110 correct chain if the automatic operation produced an incorrect result.
12112 Now the chain builder is disabled if either:
12114 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12116 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12118 The reasoning behind this is that an application would not want the
12119 auto chain building to take place if extra chain certificates are
12120 present and it might also want a means of sending no additional
12121 certificates (for example the chain has two certificates and the
12126 * Add the possibility to build without the ENGINE framework.
12128 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12130 * Under Win32 gmtime() can return NULL: check return value in
12131 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12135 * DSA routines: under certain error conditions uninitialized BN objects
12136 could be freed. Solution: make sure initialization is performed early
12137 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12138 Nils Larsch <nla@trustcenter.de> via PR#459)
12142 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12143 checked on reconnect on the client side, therefore session resumption
12144 could still fail with a "ssl session id is different" error. This
12145 behaviour is masked when SSL_OP_ALL is used due to
12146 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12147 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12148 followup to PR #377.
12152 * IA-32 assembler support enhancements: unified ELF targets, support
12153 for SCO/Caldera platforms, fix for Cygwin shared build.
12157 * Add support for FreeBSD on sparc64. As a consequence, support for
12158 FreeBSD on non-x86 processors is separate from x86 processors on
12159 the config script, much like the NetBSD support.
12161 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12163 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12165 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12168 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12169 code (06) was taken as the first octet of the session ID and the last
12170 octet was ignored consequently. As a result SSLv2 client side session
12171 caching could not have worked due to the session ID mismatch between
12173 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12178 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12179 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12184 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12185 seems that in spite of existing for more than a year, many application
12186 author have done nothing to provide the necessary callbacks, which
12187 means that this particular engine will not work properly anywhere.
12188 This is a very unfortunate situation which forces us, in the name
12189 of usability, to give the hw_ncipher.c a static lock, which is part
12191 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12192 appear in 0.9.8 or later. We EXPECT application authors to have
12193 dealt properly with this when 0.9.8 is released (unless we actually
12194 make such changes in the libcrypto locking code that changes will
12195 have to be made anyway).
12199 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12200 octets have been read, EOF or an error occurs. Without this change
12201 some truncated ASN1 structures will not produce an error.
12205 * Disable Heimdal support, since it hasn't been fully implemented.
12206 Still give the possibility to force the use of Heimdal, but with
12207 warnings and a request that patches get sent to openssl-dev.
12211 * Add the VC-CE target, introduce the WINCE sysname, and add
12212 INSTALL.WCE and appropriate conditionals to make it build.
12214 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12216 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12217 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12218 edit numbers of the version.
12220 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12222 * Introduce safe string copy and catenation functions
12223 (BUF_strlcpy() and BUF_strlcat()).
12225 *Ben Laurie (CHATS) and Richard Levitte*
12227 * Avoid using fixed-size buffers for one-line DNs.
12229 *Ben Laurie (CHATS)*
12231 * Add BUF_MEM_grow_clean() to avoid information leakage when
12232 resizing buffers containing secrets, and use where appropriate.
12234 *Ben Laurie (CHATS)*
12236 * Avoid using fixed size buffers for configuration file location.
12238 *Ben Laurie (CHATS)*
12240 * Avoid filename truncation for various CA files.
12242 *Ben Laurie (CHATS)*
12244 * Use sizeof in preference to magic numbers.
12246 *Ben Laurie (CHATS)*
12248 * Avoid filename truncation in cert requests.
12250 *Ben Laurie (CHATS)*
12252 * Add assertions to check for (supposedly impossible) buffer
12255 *Ben Laurie (CHATS)*
12257 * Don't cache truncated DNS entries in the local cache (this could
12258 potentially lead to a spoofing attack).
12260 *Ben Laurie (CHATS)*
12262 * Fix various buffers to be large enough for hex/decimal
12263 representations in a platform independent manner.
12265 *Ben Laurie (CHATS)*
12267 * Add CRYPTO_realloc_clean() to avoid information leakage when
12268 resizing buffers containing secrets, and use where appropriate.
12270 *Ben Laurie (CHATS)*
12272 * Add BIO_indent() to avoid much slightly worrying code to do
12275 *Ben Laurie (CHATS)*
12277 * Convert sprintf()/BIO_puts() to BIO_printf().
12279 *Ben Laurie (CHATS)*
12281 * buffer_gets() could terminate with the buffer only half
12284 *Ben Laurie (CHATS)*
12286 * Add assertions to prevent user-supplied crypto functions from
12287 overflowing internal buffers by having large block sizes, etc.
12289 *Ben Laurie (CHATS)*
12291 * New OPENSSL_assert() macro (similar to assert(), but enabled
12294 *Ben Laurie (CHATS)*
12296 * Eliminate unused copy of key in RC4.
12298 *Ben Laurie (CHATS)*
12300 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12302 *Ben Laurie (CHATS)*
12304 * Fix off-by-one error in EGD path.
12306 *Ben Laurie (CHATS)*
12308 * If RANDFILE path is too long, ignore instead of truncating.
12310 *Ben Laurie (CHATS)*
12312 * Eliminate unused and incorrectly sized X.509 structure
12315 *Ben Laurie (CHATS)*
12317 * Eliminate unused and dangerous function knumber().
12319 *Ben Laurie (CHATS)*
12321 * Eliminate unused and dangerous structure, KSSL_ERR.
12323 *Ben Laurie (CHATS)*
12325 * Protect against overlong session ID context length in an encoded
12326 session object. Since these are local, this does not appear to be
12329 *Ben Laurie (CHATS)*
12331 * Change from security patch (see 0.9.6e below) that did not affect
12332 the 0.9.6 release series:
12334 Remote buffer overflow in SSL3 protocol - an attacker could
12335 supply an oversized master key in Kerberos-enabled versions.
12338 *Ben Laurie (CHATS)*
12340 * Change the SSL kerb5 codes to match RFC 2712.
12344 * Make -nameopt work fully for req and add -reqopt switch.
12346 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12348 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12350 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12352 * Make sure tests can be performed even if the corresponding algorithms
12353 have been removed entirely. This was also the last step to make
12354 OpenSSL compilable with DJGPP under all reasonable conditions.
12356 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12358 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12359 to allow version independent disabling of normally unselected ciphers,
12360 which may be activated as a side-effect of selecting a single cipher.
12362 (E.g., cipher list string "RSA" enables ciphersuites that are left
12363 out of "ALL" because they do not provide symmetric encryption.
12364 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12366 *Lutz Jaenicke, Bodo Moeller*
12368 * Add appropriate support for separate platform-dependent build
12369 directories. The recommended way to make a platform-dependent
12370 build directory is the following (tested on Linux), maybe with
12373 # Place yourself outside of the OpenSSL source tree. In
12374 # this example, the environment variable OPENSSL_SOURCE
12375 # is assumed to contain the absolute OpenSSL source directory.
12376 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12377 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12378 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12379 mkdir -p `dirname $F`
12380 ln -s $OPENSSL_SOURCE/$F $F
12383 To be absolutely sure not to disturb the source tree, a "make clean"
12384 is a good thing. If it isn't successful, don't worry about it,
12385 it probably means the source directory is very clean.
12389 * Make sure any ENGINE control commands make local copies of string
12390 pointers passed to them whenever necessary. Otherwise it is possible
12391 the caller may have overwritten (or deallocated) the original string
12392 data when a later ENGINE operation tries to use the stored values.
12394 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12396 * Improve diagnostics in file reading and command-line digests.
12398 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12400 * Add AES modes CFB and OFB to the object database. Correct an
12401 error in AES-CFB decryption.
12405 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12406 allows existing EVP_CIPHER_CTX structures to be reused after
12407 calling `EVP_*Final()`. This behaviour is used by encryption
12408 BIOs and some applications. This has the side effect that
12409 applications must explicitly clean up cipher contexts with
12410 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12414 * Check the values of dna and dnb in bn_mul_recursive before calling
12415 bn_mul_comba (a non zero value means the a or b arrays do not contain
12416 n2 elements) and fallback to bn_mul_normal if either is not zero.
12420 * Fix escaping of non-ASCII characters when using the -subj option
12421 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12425 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12426 form for "surname", serialNumber has no short form.
12427 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12428 therefore remove "mail" short name for "internet 7".
12429 The OID for unique identifiers in X509 certificates is
12430 x500UniqueIdentifier, not uniqueIdentifier.
12431 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12435 * Add an "init" command to the ENGINE config module and auto initialize
12436 ENGINEs. Without any "init" command the ENGINE will be initialized
12437 after all ctrl commands have been executed on it. If init=1 the
12438 ENGINE is initialized at that point (ctrls before that point are run
12439 on the uninitialized ENGINE and after on the initialized one). If
12440 init=0 then the ENGINE will not be initialized at all.
12444 * Fix the 'app_verify_callback' interface so that the user-defined
12445 argument is actually passed to the callback: In the
12446 SSL_CTX_set_cert_verify_callback() prototype, the callback
12447 declaration has been changed from
12450 int (*cb)(X509_STORE_CTX *,void *);
12451 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12452 i=s->ctx->app_verify_callback(&ctx)
12453 has been changed into
12454 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12456 To update applications using SSL_CTX_set_cert_verify_callback(),
12457 a dummy argument can be added to their callback functions.
12459 *D. K. Smetters <smetters@parc.xerox.com>*
12461 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12463 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12465 * Add and OPENSSL_LOAD_CONF define which will cause
12466 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12467 This allows older applications to transparently support certain
12468 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12469 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12470 load the config file and OPENSSL_add_all_algorithms_conf() which will
12471 always load it have also been added.
12475 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12476 Adjust NIDs and EVP layer.
12478 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12480 * Config modules support in openssl utility.
12482 Most commands now load modules from the config file,
12483 though in a few (such as version) this isn't done
12484 because it couldn't be used for anything.
12486 In the case of ca and req the config file used is
12487 the same as the utility itself: that is the -config
12488 command line option can be used to specify an
12493 * Move default behaviour from OPENSSL_config(). If appname is NULL
12494 use "openssl_conf" if filename is NULL use default openssl config file.
12498 * Add an argument to OPENSSL_config() to allow the use of an alternative
12499 config section name. Add a new flag to tolerate a missing config file
12500 and move code to CONF_modules_load_file().
12504 * Support for crypto accelerator cards from Accelerated Encryption
12505 Processing, www.aep.ie. (Use engine 'aep')
12506 The support was copied from 0.9.6c [engine] and adapted/corrected
12507 to work with the new engine framework.
12509 *AEP Inc. and Richard Levitte*
12511 * Support for SureWare crypto accelerator cards from Baltimore
12512 Technologies. (Use engine 'sureware')
12513 The support was copied from 0.9.6c [engine] and adapted
12514 to work with the new engine framework.
12518 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12519 make the newer ENGINE framework commands for the CHIL engine work.
12521 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12523 * Make it possible to produce shared libraries on ReliantUNIX.
12525 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12527 * Add the configuration target debug-linux-ppro.
12528 Make 'openssl rsa' use the general key loading routines
12529 implemented in `apps.c`, and make those routines able to
12530 handle the key format FORMAT_NETSCAPE and the variant
12533 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12535 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12537 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12539 * Add -keyform to rsautl, and document -engine.
12541 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12543 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12544 BIO_R_NO_SUCH_FILE error code rather than the generic
12545 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12549 * Add new functions
12550 ERR_peek_last_error
12551 ERR_peek_last_error_line
12552 ERR_peek_last_error_line_data.
12553 These are similar to
12555 ERR_peek_error_line
12556 ERR_peek_error_line_data,
12557 but report on the latest error recorded rather than the first one
12558 still in the error queue.
12560 *Ben Laurie, Bodo Moeller*
12562 * default_algorithms option in ENGINE config module. This allows things
12564 default_algorithms = ALL
12565 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12569 * Preliminary ENGINE config module.
12573 * New experimental application configuration code.
12577 * Change the AES code to follow the same name structure as all other
12578 symmetric ciphers, and behave the same way. Move everything to
12579 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12581 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12583 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12585 *Ben Laurie and Theo de Raadt*
12587 * Add option to output public keys in req command.
12589 *Massimiliano Pala madwolf@openca.org*
12591 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12592 (up to about 10% better than before for P-192 and P-224).
12596 * New functions/macros
12598 SSL_CTX_set_msg_callback(ctx, cb)
12599 SSL_CTX_set_msg_callback_arg(ctx, arg)
12600 SSL_set_msg_callback(ssl, cb)
12601 SSL_set_msg_callback_arg(ssl, arg)
12603 to request calling a callback function
12605 void cb(int write_p, int version, int content_type,
12606 const void *buf, size_t len, SSL *ssl, void *arg)
12608 whenever a protocol message has been completely received
12609 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12610 protocol version according to which the SSL library interprets
12611 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12612 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12613 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12614 specification (change_cipher_spec(20), alert(21), handshake(22)).
12615 'buf' and 'len' point to the actual message, 'ssl' to the
12616 SSL object, and 'arg' is the application-defined value set by
12617 SSL[_CTX]_set_msg_callback_arg().
12619 'openssl s_client' and 'openssl s_server' have new '-msg' options
12620 to enable a callback that displays all protocol messages.
12624 * Change the shared library support so shared libraries are built as
12625 soon as the corresponding static library is finished, and thereby get
12626 openssl and the test programs linked against the shared library.
12627 This still only happens when the keyword "shard" has been given to
12628 the configuration scripts.
12630 NOTE: shared library support is still an experimental thing, and
12631 backward binary compatibility is still not guaranteed.
12633 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12635 * Add support for Subject Information Access extension.
12637 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12639 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12640 additional bytes when new memory had to be allocated, not just
12641 when reusing an existing buffer.
12645 * New command line and configuration option 'utf8' for the req command.
12646 This allows field values to be specified as UTF8 strings.
12650 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12651 runs for the former and machine-readable output for the latter.
12655 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12656 of the e-mail address in the DN (i.e., it will go into a certificate
12657 extension only). The new configuration file option 'email_in_dn = no'
12658 has the same effect.
12660 *Massimiliano Pala madwolf@openca.org*
12662 * Change all functions with names starting with `des_` to be starting
12663 with `DES_` instead. Add wrappers that are compatible with libdes,
12664 but are named `_ossl_old_des_*`. Finally, add macros that map the
12665 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12666 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12667 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12670 Since we provide two compatibility mappings, the user needs to
12671 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12672 compatibility is desired. The default (i.e., when that macro
12673 isn't defined) is OpenSSL 0.9.6c compatibility.
12675 There are also macros that enable and disable the support of old
12676 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12677 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12678 are defined, the default will apply: to support the old des routines.
12680 In either case, one must include openssl/des.h to get the correct
12681 definitions. Do not try to just include openssl/des_old.h, that
12684 NOTE: This is a major break of an old API into a new one. Software
12685 authors are encouraged to switch to the `DES_` style functions. Some
12686 time in the future, des_old.h and the libdes compatibility functions
12687 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12688 default), and then completely removed.
12692 * Test for certificates which contain unsupported critical extensions.
12693 If such a certificate is found during a verify operation it is
12694 rejected by default: this behaviour can be overridden by either
12695 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12696 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12697 X509_supported_extension() has also been added which returns 1 if a
12698 particular extension is supported.
12702 * Modify the behaviour of EVP cipher functions in similar way to digests
12703 to retain compatibility with existing code.
12707 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12708 compatibility with existing code. In particular the 'ctx' parameter does
12709 not have to be to be initialized before the call to EVP_DigestInit() and
12710 it is tidied up after a call to EVP_DigestFinal(). New function
12711 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12712 EVP_MD_CTX_copy() changed to not require the destination to be
12713 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12714 requires the destination to be valid.
12716 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12717 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12721 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12722 so that complete 'Handshake' protocol structures are kept in memory
12723 instead of overwriting 'msg_type' and 'length' with 'body' data.
12727 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12729 *Massimo Santin via Richard Levitte*
12731 * Major restructuring to the underlying ENGINE code. This includes
12732 reduction of linker bloat, separation of pure "ENGINE" manipulation
12733 (initialisation, etc) from functionality dealing with implementations
12734 of specific crypto interfaces. This change also introduces integrated
12735 support for symmetric ciphers and digest implementations - so ENGINEs
12736 can now accelerate these by providing EVP_CIPHER and EVP_MD
12737 implementations of their own. This is detailed in
12738 [crypto/engine/README.md](crypto/engine/README.md)
12739 as it couldn't be adequately described here. However, there are a few
12740 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12741 were changed in the original introduction of ENGINE code have now
12742 reverted back - the hooking from this code to ENGINE is now a good
12743 deal more passive and at run-time, operations deal directly with
12744 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12745 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12746 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12747 they were not being used by the framework as there is no concept of a
12748 BIGNUM_METHOD and they could not be generalised to the new
12749 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12750 ENGINE_cpy() has been removed as it cannot be consistently defined in
12755 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12759 * Change mkdef.pl to sort symbols that get the same entry number,
12760 and make sure the automatically generated functions `ERR_load_*`
12761 become part of libeay.num as well.
12765 * New function SSL_renegotiate_pending(). This returns true once
12766 renegotiation has been requested (either SSL_renegotiate() call
12767 or HelloRequest/ClientHello received from the peer) and becomes
12768 false once a handshake has been completed.
12769 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12770 sends a HelloRequest, but does not ensure that a handshake takes
12771 place. SSL_renegotiate_pending() is useful for checking if the
12772 client has followed the request.)
12776 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12777 By default, clients may request session resumption even during
12778 renegotiation (if session ID contexts permit); with this option,
12779 session resumption is possible only in the first handshake.
12781 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12782 more bits available for options that should not be part of
12783 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12787 * Add some demos for certificate and certificate request creation.
12791 * Make maximum certificate chain size accepted from the peer application
12792 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12793 "Douglas E. Engert" <deengert@anl.gov>.
12797 * Add support for shared libraries for Unixware-7
12798 (Boyd Lynn Gerber <gerberb@zenez.com>).
12802 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12803 be done prior to destruction. Use this to unload error strings from
12804 ENGINEs that load their own error strings. NB: This adds two new API
12805 functions to "get" and "set" this destroy handler in an ENGINE.
12809 * Alter all existing ENGINE implementations (except "openssl" and
12810 "openbsd") to dynamically instantiate their own error strings. This
12811 makes them more flexible to be built both as statically-linked ENGINEs
12812 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12813 Also, add stub code to each that makes building them as self-contained
12814 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12818 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12819 implementations into applications that are completely implemented in
12820 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12821 commands that can be used to configure what shared-library to load and
12822 to control aspects of the way it is handled. Also, made an update to
12823 the [README-Engine.md](README-Engine.md) file
12824 that brings its information up-to-date and
12825 provides some information and instructions on the "dynamic" ENGINE
12826 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12830 * Make it possible to unload ranges of ERR strings with a new
12831 "ERR_unload_strings" function.
12835 * Add a copy() function to EVP_MD.
12839 * Make EVP_MD routines take a context pointer instead of just the
12840 md_data void pointer.
12844 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12845 that the digest can only process a single chunk of data
12846 (typically because it is provided by a piece of
12847 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12848 is only going to provide a single chunk of data, and hence the
12849 framework needn't accumulate the data for oneshot drivers.
12853 * As with "ERR", make it possible to replace the underlying "ex_data"
12854 functions. This change also alters the storage and management of global
12855 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12856 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12857 index counters. The API functions that use this state have been changed
12858 to take a "class_index" rather than pointers to the class's local STACK
12859 and counter, and there is now an API function to dynamically create new
12860 classes. This centralisation allows us to (a) plug a lot of the
12861 thread-safety problems that existed, and (b) makes it possible to clean
12862 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12863 such data would previously have always leaked in application code and
12864 workarounds were in place to make the memory debugging turn a blind eye
12865 to it. Application code that doesn't use this new function will still
12866 leak as before, but their memory debugging output will announce it now
12867 rather than letting it slide.
12869 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12870 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12871 has a return value to indicate success or failure.
12875 * Make it possible to replace the underlying "ERR" functions such that the
12876 global state (2 LHASH tables and 2 locks) is only used by the "default"
12877 implementation. This change also adds two functions to "get" and "set"
12878 the implementation prior to it being automatically set the first time
12879 any other ERR function takes place. Ie. an application can call "get",
12880 pass the return value to a module it has just loaded, and that module
12881 can call its own "set" function using that value. This means the
12882 module's "ERR" operations will use (and modify) the error state in the
12883 application and not in its own statically linked copy of OpenSSL code.
12887 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12888 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12889 the operation, and provides a more encapsulated way for external code
12890 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12891 to use these functions rather than manually incrementing the counts.
12893 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12897 * Add EVP test program.
12901 * Add symmetric cipher support to ENGINE. Expect the API to change!
12905 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12906 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12907 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12908 These allow a CRL to be built without having to access X509_CRL fields
12909 directly. Modify 'ca' application to use new functions.
12913 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12914 bug workarounds. Rollback attack detection is a security feature.
12915 The problem will only arise on OpenSSL servers when TLSv1 is not
12916 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12917 Software authors not wanting to support TLSv1 will have special reasons
12918 for their choice and can explicitly enable this option.
12920 *Bodo Moeller, Lutz Jaenicke*
12922 * Rationalise EVP so it can be extended: don't include a union of
12923 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12924 (similar to those existing for EVP_CIPHER_CTX).
12929 EVP_MD_CTX_init(&md); /* new function call */
12930 EVP_DigestInit(&md, EVP_sha1());
12931 EVP_DigestUpdate(&md, in, len);
12932 EVP_DigestFinal(&md, out, NULL);
12933 EVP_MD_CTX_cleanup(&md); /* new function call */
12937 * Make DES key schedule conform to the usual scheme, as well as
12938 correcting its structure. This means that calls to DES functions
12939 now have to pass a pointer to a des_key_schedule instead of a
12940 plain des_key_schedule (which was actually always a pointer
12943 des_key_schedule ks;
12945 des_set_key_checked(..., &ks);
12946 des_ncbc_encrypt(..., &ks, ...);
12948 (Note that a later change renames 'des_...' into 'DES_...'.)
12952 * Initial reduction of linker bloat: the use of some functions, such as
12953 PEM causes large amounts of unused functions to be linked in due to
12954 poor organisation. For example pem_all.c contains every PEM function
12955 which has a knock on effect of linking in large amounts of (unused)
12956 ASN1 code. Grouping together similar functions and splitting unrelated
12957 functions prevents this.
12961 * Cleanup of EVP macros.
12965 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12966 correct `_ecb suffix`.
12970 * Add initial OCSP responder support to ocsp application. The
12971 revocation information is handled using the text based index
12972 use by the ca application. The responder can either handle
12973 requests generated internally, supplied in files (for example
12974 via a CGI script) or using an internal minimal server.
12978 * Add configuration choices to get zlib compression for TLS.
12982 * Changes to Kerberos SSL for RFC 2712 compliance:
12983 1. Implemented real KerberosWrapper, instead of just using
12984 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12985 2. Implemented optional authenticator field of KerberosWrapper.
12987 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12988 and authenticator structs; see crypto/krb5/.
12990 Generalized Kerberos calls to support multiple Kerberos libraries.
12991 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12992 via Richard Levitte*
12994 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12995 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12996 values for each of the key sizes rather than having just
12997 parameters (and 'speed' generating keys each time).
13001 * Speed up EVP routines.
13004 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13005 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13006 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13007 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13009 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13010 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13011 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13014 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13016 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13020 * Added the OS2-EMX target.
13022 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13024 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13025 New functions to support `NCONF` routines in extension code.
13026 New function `CONF_set_nconf()`
13027 to allow functions which take an `NCONF` to also handle the old `LHASH`
13028 structure: this means that the old `CONF` compatible routines can be
13029 retained (in particular w.rt. extensions) without having to duplicate the
13030 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13034 * Enhance the general user interface with mechanisms for inner control
13035 and with possibilities to have yes/no kind of prompts.
13039 * Change all calls to low-level digest routines in the library and
13040 applications to use EVP. Add missing calls to HMAC_cleanup() and
13041 don't assume HMAC_CTX can be copied using memcpy().
13043 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13045 * Add the possibility to control engines through control names but with
13046 arbitrary arguments instead of just a string.
13047 Change the key loaders to take a UI_METHOD instead of a callback
13048 function pointer. NOTE: this breaks binary compatibility with earlier
13049 versions of OpenSSL [engine].
13050 Adapt the nCipher code for these new conditions and add a card insertion
13055 * Enhance the general user interface with mechanisms to better support
13056 dialog box interfaces, application-defined prompts, the possibility
13057 to use defaults (for example default passwords from somewhere else)
13058 and interrupts/cancellations.
13062 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13063 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13067 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13068 tidy up some unnecessarily weird code in 'sk_new()').
13070 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13072 * Change the key loading routines for ENGINEs to use the same kind
13073 callback (pem_password_cb) as all other routines that need this
13078 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13079 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13080 than this minimum value is recommended.
13084 * New random seeder for OpenVMS, using the system process statistics
13085 that are easily reachable.
13089 * Windows apparently can't transparently handle global
13090 variables defined in DLLs. Initialisations such as:
13092 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13094 won't compile. This is used by the any applications that need to
13095 declare their own ASN1 modules. This was fixed by adding the option
13096 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13097 needed for static libraries under Win32.
13101 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13102 setting of purpose and trust fields. New X509_STORE trust and
13103 purpose functions and tidy up setting in other SSL functions.
13107 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13108 structure. These are inherited by X509_STORE_CTX when it is
13109 initialised. This allows various defaults to be set in the
13110 X509_STORE structure (such as flags for CRL checking and custom
13111 purpose or trust settings) for functions which only use X509_STORE_CTX
13112 internally such as S/MIME.
13114 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13115 trust settings if they are not set in X509_STORE. This allows X509_STORE
13116 purposes and trust (in S/MIME for example) to override any set by default.
13118 Add command line options for CRL checking to smime, s_client and s_server
13123 * Initial CRL based revocation checking. If the CRL checking flag(s)
13124 are set then the CRL is looked up in the X509_STORE structure and
13125 its validity and signature checked, then if the certificate is found
13126 in the CRL the verify fails with a revoked error.
13128 Various new CRL related callbacks added to X509_STORE_CTX structure.
13130 Command line options added to 'verify' application to support this.
13132 This needs some additional work, such as being able to handle multiple
13133 CRLs with different times, extension based lookup (rather than just
13134 by subject name) and ultimately more complete V2 CRL extension
13139 * Add a general user interface API (crypto/ui/). This is designed
13140 to replace things like des_read_password and friends (backward
13141 compatibility functions using this new API are provided).
13142 The purpose is to remove prompting functions from the DES code
13143 section as well as provide for prompting through dialog boxes in
13144 a window system and the like.
13148 * Add "ex_data" support to ENGINE so implementations can add state at a
13149 per-structure level rather than having to store it globally.
13153 * Make it possible for ENGINE structures to be copied when retrieved by
13154 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13155 This causes the "original" ENGINE structure to act like a template,
13156 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13157 operational state can be localised to each ENGINE structure, despite the
13158 fact they all share the same "methods". New ENGINE structures returned in
13159 this case have no functional references and the return value is the single
13160 structural reference. This matches the single structural reference returned
13161 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13166 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13167 needs to match any other type at all we need to manually clear the
13172 * Changes to the "openssl engine" utility to include;
13173 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13174 about an ENGINE's available control commands.
13175 - executing control commands from command line arguments using the
13176 '-pre' and '-post' switches. '-post' is only used if '-t' is
13177 specified and the ENGINE is successfully initialised. The syntax for
13178 the individual commands are colon-separated, for example;
13179 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13183 * New dynamic control command support for ENGINEs. ENGINEs can now
13184 declare their own commands (numbers), names (strings), descriptions,
13185 and input types for run-time discovery by calling applications. A
13186 subset of these commands are implicitly classed as "executable"
13187 depending on their input type, and only these can be invoked through
13188 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13189 can be based on user input, config files, etc). The distinction is
13190 that "executable" commands cannot return anything other than a boolean
13191 result and can only support numeric or string input, whereas some
13192 discoverable commands may only be for direct use through
13193 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13194 pointers, or other custom uses. The "executable" commands are to
13195 support parameterisations of ENGINE behaviour that can be
13196 unambiguously defined by ENGINEs and used consistently across any
13197 OpenSSL-based application. Commands have been added to all the
13198 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13199 control over shared-library paths without source code alterations.
13203 * Changed all ENGINE implementations to dynamically allocate their
13204 ENGINEs rather than declaring them statically. Apart from this being
13205 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13206 this also allows the implementations to compile without using the
13207 internal engine_int.h header.
13211 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13212 'const' value. Any code that should be able to modify a RAND_METHOD
13213 should already have non-const pointers to it (ie. they should only
13214 modify their own ones).
13218 * Made a variety of little tweaks to the ENGINE code.
13219 - "atalla" and "ubsec" string definitions were moved from header files
13220 to C code. "nuron" string definitions were placed in variables
13221 rather than hard-coded - allowing parameterisation of these values
13222 later on via ctrl() commands.
13223 - Removed unused "#if 0"'d code.
13224 - Fixed engine list iteration code so it uses ENGINE_free() to release
13225 structural references.
13226 - Constified the RAND_METHOD element of ENGINE structures.
13227 - Constified various get/set functions as appropriate and added
13228 missing functions (including a catch-all ENGINE_cpy that duplicates
13229 all ENGINE values onto a new ENGINE except reference counts/state).
13230 - Removed NULL parameter checks in get/set functions. Setting a method
13231 or function to NULL is a way of cancelling out a previously set
13232 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13233 and doesn't justify the extra error symbols and code.
13234 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13235 flags from engine_int.h to engine.h.
13236 - Changed prototypes for ENGINE handler functions (init(), finish(),
13237 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13241 * Implement binary inversion algorithm for BN_mod_inverse in addition
13242 to the algorithm using long division. The binary algorithm can be
13243 used only if the modulus is odd. On 32-bit systems, it is faster
13244 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13245 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13246 up to 450 bits. In 64-bit environments, the binary algorithm
13247 appears to be advantageous for much longer moduli; here we use it
13248 for moduli up to 2048 bits.
13252 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13253 could not support the combine flag in choice fields.
13257 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13258 extensions from a certificate request to the certificate.
13262 * Allow multiple 'certopt' and 'nameopt' options to be separated
13263 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13264 file: this allows the display of the certificate about to be
13265 signed to be customised, to allow certain fields to be included
13266 or excluded and extension details. The old system didn't display
13267 multicharacter strings properly, omitted fields not in the policy
13268 and couldn't display additional details such as extensions.
13272 * Function EC_POINTs_mul for multiple scalar multiplication
13273 of an arbitrary number of elliptic curve points
13274 \sum scalars[i]*points[i],
13275 optionally including the generator defined for the EC_GROUP:
13276 scalar*generator + \sum scalars[i]*points[i].
13278 EC_POINT_mul is a simple wrapper function for the typical case
13279 that the point list has just one item (besides the optional
13284 * First EC_METHODs for curves over GF(p):
13286 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13287 operations and provides various method functions that can also
13288 operate with faster implementations of modular arithmetic.
13290 EC_GFp_mont_method() reuses most functions that are part of
13291 EC_GFp_simple_method, but uses Montgomery arithmetic.
13293 *Bodo Moeller; point addition and point doubling
13294 implementation directly derived from source code provided by
13295 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13297 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13298 crypto/ec/ec_lib.c):
13300 Curves are EC_GROUP objects (with an optional group generator)
13301 based on EC_METHODs that are built into the library.
13303 Points are EC_POINT objects based on EC_GROUP objects.
13305 Most of the framework would be able to handle curves over arbitrary
13306 finite fields, but as there are no obvious types for fields other
13307 than GF(p), some functions are limited to that for now.
13311 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13312 that the file contains a complete HTTP response.
13316 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13317 change the def and num file printf format specifier from "%-40sXXX"
13318 to "%-39s XXX". The latter will always guarantee a space after the
13319 field while the former will cause them to run together if the field
13320 is 40 of more characters long.
13324 * Constify the cipher and digest 'method' functions and structures
13325 and modify related functions to take constant EVP_MD and EVP_CIPHER
13330 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13331 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13335 * Modify `EVP_Digest*()` routines so they now return values. Although the
13336 internal software routines can never fail additional hardware versions
13341 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13343 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13344 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13347 ERR_R_NESTED_ASN1_ERROR
13349 ERR_R_MISSING_ASN1_EOS
13350 were 4 .. 9, conflicting with
13351 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13353 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13354 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13356 Add new error code 'ERR_R_INTERNAL_ERROR'.
13360 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13365 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13366 sets the subject name for a new request or supersedes the
13367 subject name in a given request. Formats that can be parsed are
13368 'CN=Some Name, OU=myOU, C=IT'
13370 'CN=Some Name/OU=myOU/C=IT'.
13372 Add options '-batch' and '-verbose' to 'openssl req'.
13374 *Massimiliano Pala <madwolf@hackmasters.net>*
13376 * Introduce the possibility to access global variables through
13377 functions on platform were that's the best way to handle exporting
13378 global variables in shared libraries. To enable this functionality,
13379 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13380 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13381 is normally done by Configure or something similar).
13383 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13384 in the source file (foo.c) like this:
13386 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13387 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13389 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13390 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13392 OPENSSL_DECLARE_GLOBAL(int,foo);
13393 #define foo OPENSSL_GLOBAL_REF(foo)
13394 OPENSSL_DECLARE_GLOBAL(double,bar);
13395 #define bar OPENSSL_GLOBAL_REF(bar)
13397 The #defines are very important, and therefore so is including the
13398 header file everywhere where the defined globals are used.
13400 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13401 of ASN.1 items, but that structure is a bit different.
13403 The largest change is in util/mkdef.pl which has been enhanced with
13404 better and easier to understand logic to choose which symbols should
13405 go into the Windows .def files as well as a number of fixes and code
13406 cleanup (among others, algorithm keywords are now sorted
13407 lexicographically to avoid constant rewrites).
13411 * In BN_div() keep a copy of the sign of 'num' before writing the
13412 result to 'rm' because if rm==num the value will be overwritten
13413 and produce the wrong result if 'num' is negative: this caused
13414 problems with BN_mod() and BN_nnmod().
13418 * Function OCSP_request_verify(). This checks the signature on an
13419 OCSP request and verifies the signer certificate. The signer
13420 certificate is just checked for a generic purpose and OCSP request
13425 * Add OCSP_check_validity() function to check the validity of OCSP
13426 responses. OCSP responses are prepared in real time and may only
13427 be a few seconds old. Simply checking that the current time lies
13428 between thisUpdate and nextUpdate max reject otherwise valid responses
13429 caused by either OCSP responder or client clock inaccuracy. Instead
13430 we allow thisUpdate and nextUpdate to fall within a certain period of
13431 the current time. The age of the response can also optionally be
13432 checked. Two new options -validity_period and -status_age added to
13437 * If signature or public key algorithm is unrecognized print out its
13438 OID rather that just UNKNOWN.
13442 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13443 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13444 ID to be generated from the issuer certificate alone which can then be
13445 passed to OCSP_id_issuer_cmp().
13449 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13450 ASN1 modules to export functions returning ASN1_ITEM pointers
13451 instead of the ASN1_ITEM structures themselves. This adds several
13452 new macros which allow the underlying ASN1 function/structure to
13453 be accessed transparently. As a result code should not use ASN1_ITEM
13454 references directly (such as &X509_it) but instead use the relevant
13455 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13456 use of the new ASN1 code on platforms where exporting structures
13457 is problematical (for example in shared libraries) but exporting
13458 functions returning pointers to structures is not.
13462 * Add support for overriding the generation of SSL/TLS session IDs.
13463 These callbacks can be registered either in an SSL_CTX or per SSL.
13464 The purpose of this is to allow applications to control, if they wish,
13465 the arbitrary values chosen for use as session IDs, particularly as it
13466 can be useful for session caching in multiple-server environments. A
13467 command-line switch for testing this (and any client code that wishes
13468 to use such a feature) has been added to "s_server".
13470 *Geoff Thorpe, Lutz Jaenicke*
13472 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13473 of the form `#if defined(...) || defined(...) || ...` and
13474 `#if !defined(...) && !defined(...) && ...`. This also avoids
13475 the growing number of special cases it was previously handling.
13479 * Make all configuration macros available for application by making
13480 sure they are available in opensslconf.h, by giving them names starting
13481 with `OPENSSL_` to avoid conflicts with other packages and by making
13482 sure e_os2.h will cover all platform-specific cases together with
13484 Additionally, it is now possible to define configuration/platform-
13485 specific names (called "system identities"). In the C code, these
13486 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13487 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13488 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13493 * New option -set_serial to 'req' and 'x509' this allows the serial
13494 number to use to be specified on the command line. Previously self
13495 signed certificates were hard coded with serial number 0 and the
13496 CA options of 'x509' had to use a serial number in a file which was
13501 * New options to 'ca' utility to support V2 CRL entry extensions.
13502 Currently CRL reason, invalidity date and hold instruction are
13503 supported. Add new CRL extensions to V3 code and some new objects.
13507 * New function EVP_CIPHER_CTX_set_padding() this is used to
13508 disable standard block padding (aka PKCS#5 padding) in the EVP
13509 API, which was previously mandatory. This means that the data is
13510 not padded in any way and so the total length much be a multiple
13511 of the block size, otherwise an error occurs.
13515 * Initial (incomplete) OCSP SSL support.
13519 * New function OCSP_parse_url(). This splits up a URL into its host,
13520 port and path components: primarily to parse OCSP URLs. New -url
13521 option to ocsp utility.
13525 * New nonce behavior. The return value of OCSP_check_nonce() now
13526 reflects the various checks performed. Applications can decide
13527 whether to tolerate certain situations such as an absent nonce
13528 in a response when one was present in a request: the ocsp application
13529 just prints out a warning. New function OCSP_add1_basic_nonce()
13530 this is to allow responders to include a nonce in a response even if
13531 the request is nonce-less.
13535 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13536 skipped when using openssl x509 multiple times on a single input file,
13537 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13541 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13542 set string type: to handle setting ASN1_TIME structures. Fix ca
13543 utility to correctly initialize revocation date of CRLs.
13547 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13548 the clients preferred ciphersuites and rather use its own preferences.
13549 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13550 Internet Explorer by ensuring unchanged hash method during stepup.
13551 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13555 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13556 to aes and add a new 'exist' option to print out symbols that don't
13561 * Additional options to ocsp utility to allow flags to be set and
13562 additional certificates supplied.
13566 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13567 OCSP client a number of certificate to only verify the response
13572 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13573 handle the new API. Currently only ECB, CBC modes supported. Add new
13576 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13577 Encryption Standard (AES) Ciphersuites for Transport Layer
13578 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13579 not enabled by default and were not part of the "ALL" ciphersuite
13580 alias because they were not yet official; they could be
13581 explicitly requested by specifying the "AESdraft" ciphersuite
13582 group alias. In the final release of OpenSSL 0.9.7, the group
13583 alias is called "AES" and is part of "ALL".)
13585 *Ben Laurie, Steve Henson, Bodo Moeller*
13587 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13588 request to response.
13592 * Functions for OCSP responders. OCSP_request_onereq_count(),
13593 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13594 extract information from a certificate request. OCSP_response_create()
13595 creates a response and optionally adds a basic response structure.
13596 OCSP_basic_add1_status() adds a complete single response to a basic
13597 response and returns the OCSP_SINGLERESP structure just added (to allow
13598 extensions to be included for example). OCSP_basic_add1_cert() adds a
13599 certificate to a basic response and OCSP_basic_sign() signs a basic
13600 response with various flags. New helper functions ASN1_TIME_check()
13601 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13602 (converts ASN1_TIME to GeneralizedTime).
13606 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13607 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13608 structure from a certificate. X509_pubkey_digest() digests the public_key
13609 contents: this is used in various key identifiers.
13613 * Make sk_sort() tolerate a NULL argument.
13615 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13617 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13618 passed by the function are trusted implicitly. If any of them signed the
13619 response then it is assumed to be valid and is not verified.
13623 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13624 to data. This was previously part of the PKCS7 ASN1 code. This
13625 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13626 *Steve Henson, reported by Kenneth R. Robinette
13627 <support@securenetterm.com>*
13629 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13630 routines: without these tracing memory leaks is very painful.
13631 Fix leaks in PKCS12 and PKCS7 routines.
13635 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13636 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13637 effectively meant GeneralizedTime would never be used. Now it
13638 is initialised to -1 but X509_time_adj() now has to check the value
13639 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13640 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13641 *Steve Henson, reported by Kenneth R. Robinette
13642 <support@securenetterm.com>*
13644 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13645 result in a zero length in the ASN1_INTEGER structure which was
13646 not consistent with the structure when d2i_ASN1_INTEGER() was used
13647 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13648 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13649 where it did not print out a minus for negative ASN1_INTEGER.
13653 * Add summary printout to ocsp utility. The various functions which
13654 convert status values to strings have been renamed to:
13655 OCSP_response_status_str(), OCSP_cert_status_str() and
13656 OCSP_crl_reason_str() and are no longer static. New options
13657 to verify nonce values and to disable verification. OCSP response
13658 printout format cleaned up.
13662 * Add additional OCSP certificate checks. These are those specified
13663 in RFC2560. This consists of two separate checks: the CA of the
13664 certificate being checked must either be the OCSP signer certificate
13665 or the issuer of the OCSP signer certificate. In the latter case the
13666 OCSP signer certificate must contain the OCSP signing extended key
13667 usage. This check is performed by attempting to match the OCSP
13668 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13669 in the OCSP_CERTID structures of the response.
13673 * Initial OCSP certificate verification added to OCSP_basic_verify()
13674 and related routines. This uses the standard OpenSSL certificate
13675 verify routines to perform initial checks (just CA validity) and
13676 to obtain the certificate chain. Then additional checks will be
13677 performed on the chain. Currently the root CA is checked to see
13678 if it is explicitly trusted for OCSP signing. This is used to set
13679 a root CA as a global signing root: that is any certificate that
13680 chains to that CA is an acceptable OCSP signing certificate.
13684 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13685 extensions from a separate configuration file.
13686 As when reading extensions from the main configuration file,
13687 the '-extensions ...' option may be used for specifying the
13690 *Massimiliano Pala <madwolf@comune.modena.it>*
13692 * New OCSP utility. Allows OCSP requests to be generated or
13693 read. The request can be sent to a responder and the output
13694 parsed, outputted or printed in text form. Not complete yet:
13695 still needs to check the OCSP response validity.
13699 * New subcommands for 'openssl ca':
13700 `openssl ca -status <serial>` prints the status of the cert with
13701 the given serial number (according to the index file).
13702 `openssl ca -updatedb` updates the expiry status of certificates
13705 *Massimiliano Pala <madwolf@comune.modena.it>*
13707 * New '-newreq-nodes' command option to CA.pl. This is like
13708 '-newreq', but calls 'openssl req' with the '-nodes' option
13709 so that the resulting key is not encrypted.
13711 *Damien Miller <djm@mindrot.org>*
13713 * New configuration for the GNU Hurd.
13715 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13717 * Initial code to implement OCSP basic response verify. This
13718 is currently incomplete. Currently just finds the signer's
13719 certificate and verifies the signature on the response.
13723 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13724 value of OPENSSLDIR. This is available via the new '-d' option
13725 to 'openssl version', and is also included in 'openssl version -a'.
13729 * Allowing defining memory allocation callbacks that will be given
13730 file name and line number information in additional arguments
13731 (a `const char*` and an int). The basic functionality remains, as
13732 well as the original possibility to just replace malloc(),
13733 realloc() and free() by functions that do not know about these
13734 additional arguments. To register and find out the current
13735 settings for extended allocation functions, the following
13736 functions are provided:
13738 CRYPTO_set_mem_ex_functions
13739 CRYPTO_set_locked_mem_ex_functions
13740 CRYPTO_get_mem_ex_functions
13741 CRYPTO_get_locked_mem_ex_functions
13743 These work the same way as CRYPTO_set_mem_functions and friends.
13744 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13745 extended allocation function is enabled.
13746 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13747 a conventional allocation function is enabled.
13749 *Richard Levitte, Bodo Moeller*
13751 * Finish off removing the remaining LHASH function pointer casts.
13752 There should no longer be any prototype-casting required when using
13753 the LHASH abstraction, and any casts that remain are "bugs". See
13754 the callback types and macros at the head of lhash.h for details
13755 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13759 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13760 If /dev/[u]random devices are not available or do not return enough
13761 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13763 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13764 /etc/entropy will be queried once each in this sequence, querying stops
13765 when enough entropy was collected without querying more sockets.
13769 * Change the Unix RAND_poll() variant to be able to poll several
13770 random devices, as specified by DEVRANDOM, until a sufficient amount
13771 of data has been collected. We spend at most 10 ms on each file
13772 (select timeout) and read in non-blocking mode. DEVRANDOM now
13773 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13774 (previously it was just the string "/dev/urandom"), so on typical
13775 platforms the 10 ms delay will never occur.
13776 Also separate out the Unix variant to its own file, rand_unix.c.
13777 For VMS, there's a currently-empty rand_vms.c.
13781 * Move OCSP client related routines to ocsp_cl.c. These
13782 provide utility functions which an application needing
13783 to issue a request to an OCSP responder and analyse the
13784 response will typically need: as opposed to those which an
13785 OCSP responder itself would need which will be added later.
13787 OCSP_request_sign() signs an OCSP request with an API similar
13788 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13789 response. OCSP_response_get1_basic() extracts basic response
13790 from response. OCSP_resp_find_status(): finds and extracts status
13791 information from an OCSP_CERTID structure (which will be created
13792 when the request structure is built). These are built from lower
13793 level functions which work on OCSP_SINGLERESP structures but
13794 won't normally be used unless the application wishes to examine
13795 extensions in the OCSP response for example.
13797 Replace nonce routines with a pair of functions.
13798 OCSP_request_add1_nonce() adds a nonce value and optionally
13799 generates a random value. OCSP_check_nonce() checks the
13800 validity of the nonce in an OCSP response.
13804 * Change function OCSP_request_add() to OCSP_request_add0_id().
13805 This doesn't copy the supplied OCSP_CERTID and avoids the
13806 need to free up the newly created id. Change return type
13807 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13808 This can then be used to add extensions to the request.
13809 Deleted OCSP_request_new(), since most of its functionality
13810 is now in OCSP_REQUEST_new() (and the case insensitive name
13811 clash) apart from the ability to set the request name which
13812 will be added elsewhere.
13816 * Update OCSP API. Remove obsolete extensions argument from
13817 various functions. Extensions are now handled using the new
13818 OCSP extension code. New simple OCSP HTTP function which
13819 can be used to send requests and parse the response.
13823 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13824 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13825 uses the special reorder version of SET OF to sort the attributes
13826 and reorder them to match the encoded order. This resolves a long
13827 standing problem: a verify on a PKCS7 structure just after signing
13828 it used to fail because the attribute order did not match the
13829 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13830 it uses the received order. This is necessary to tolerate some broken
13831 software that does not order SET OF. This is handled by encoding
13832 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13833 to produce the required SET OF.
13837 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13838 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13839 files to get correct declarations of the ASN.1 item variables.
13843 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13844 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13845 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13846 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13847 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13848 ASN1_ITEM and no wrapper functions.
13852 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13853 replace the old function pointer based I/O routines. Change most of
13854 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13858 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13859 lines, recognize more "algorithms" that can be deselected, and make
13860 it complain about algorithm deselection that isn't recognised.
13864 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13865 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13866 to use new functions. Add NO_ASN1_OLD which can be set to remove
13867 some old style ASN1 functions: this can be used to determine if old
13868 code will still work when these eventually go away.
13872 * New extension functions for OCSP structures, these follow the
13873 same conventions as certificates and CRLs.
13877 * New function X509V3_add1_i2d(). This automatically encodes and
13878 adds an extension. Its behaviour can be customised with various
13879 flags to append, replace or delete. Various wrappers added for
13880 certificates and CRLs.
13884 * Fix to avoid calling the underlying ASN1 print routine when
13885 an extension cannot be parsed. Correct a typo in the
13886 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13890 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13891 entries for variables.
13895 * Add functionality to `apps/openssl.c` for detecting locking
13896 problems: As the program is single-threaded, all we have
13897 to do is register a locking callback using an array for
13898 storing which locks are currently held by the program.
13902 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13903 SSL_get_ex_data_X509_STORE_idx(), which is used in
13904 ssl_verify_cert_chain() and thus can be called at any time
13905 during TLS/SSL handshakes so that thread-safety is essential.
13906 Unfortunately, the ex_data design is not at all suited
13907 for multi-threaded use, so it probably should be abolished.
13911 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13913 *Broadcom, tweaked and integrated by Geoff Thorpe*
13915 * Move common extension printing code to new function
13916 X509V3_print_extensions(). Reorganise OCSP print routines and
13917 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13921 * New function X509_signature_print() to remove duplication in some
13926 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13927 set (this was treated exactly the same as SET OF previously). This
13928 is used to reorder the STACK representing the structure to match the
13929 encoding. This will be used to get round a problem where a PKCS7
13930 structure which was signed could not be verified because the STACK
13931 order did not reflect the encoded order.
13935 * Reimplement the OCSP ASN1 module using the new code.
13939 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13940 for its ASN1 operations. The old style function pointers still exist
13941 for now but they will eventually go away.
13945 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13946 completely replaces the old ASN1 functionality with a table driven
13947 encoder and decoder which interprets an ASN1_ITEM structure describing
13948 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13949 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13950 has also been converted to the new form.
13954 * Change BN_mod_exp_recp so that negative moduli are tolerated
13955 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13956 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13957 for negative moduli.
13961 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13962 of not touching the result's sign bit.
13966 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13971 * Changed the LHASH code to use prototypes for callbacks, and created
13972 macros to declare and implement thin (optionally static) functions
13973 that provide type-safety and avoid function pointer casting for the
13974 type-specific callbacks.
13978 * Added Kerberos Cipher Suites to be used with TLS, as written in
13980 *Veers Staats <staatsvr@asc.hpc.mil>,
13981 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13983 * Reformat the FAQ so the different questions and answers can be divided
13984 in sections depending on the subject.
13988 * Have the zlib compression code load ZLIB.DLL dynamically under
13993 * New function BN_mod_sqrt for computing square roots modulo a prime
13994 (using the probabilistic Tonelli-Shanks algorithm unless
13995 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13996 be handled deterministically).
13998 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14000 * Make BN_mod_inverse faster by explicitly handling small quotients
14001 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14002 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14006 * New function BN_kronecker.
14010 * Fix BN_gcd so that it works on negative inputs; the result is
14011 positive unless both parameters are zero.
14012 Previously something reasonably close to an infinite loop was
14013 possible because numbers could be growing instead of shrinking
14014 in the implementation of Euclid's algorithm.
14018 * Fix BN_is_word() and BN_is_one() macros to take into account the
14019 sign of the number in question.
14021 Fix BN_is_word(a,w) to work correctly for w == 0.
14023 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14024 because its test if the absolute value of 'a' equals 'w'.
14025 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14026 it exists mostly for use in the implementations of BN_is_zero(),
14027 BN_is_one(), and BN_is_word().
14031 * New function BN_swap.
14035 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14036 the exponentiation functions are more likely to produce reasonable
14037 results on negative inputs.
14041 * Change BN_mod_mul so that the result is always non-negative.
14042 Previously, it could be negative if one of the factors was negative;
14043 I don't think anyone really wanted that behaviour.
14047 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14048 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14049 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14050 and add new functions:
14059 BN_mod_lshift1_quick
14061 BN_mod_lshift_quick
14063 These functions always generate non-negative results.
14065 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14066 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14068 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14069 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14070 be reduced modulo `m`.
14072 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14075 The following entry accidentally appeared in the CHANGES file
14076 distributed with OpenSSL 0.9.7. The modifications described in
14077 it do *not* apply to OpenSSL 0.9.7.
14079 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14080 was actually never needed) and in BN_mul(). The removal in BN_mul()
14081 required a small change in bn_mul_part_recursive() and the addition
14082 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14083 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14084 bn_sub_words() and bn_add_words() except they take arrays with
14090 * In 'openssl passwd', verify passwords read from the terminal
14091 unless the '-salt' option is used (which usually means that
14092 verification would just waste user's time since the resulting
14093 hash is going to be compared with some given password hash)
14094 or the new '-noverify' option is used.
14096 This is an incompatible change, but it does not affect
14097 non-interactive use of 'openssl passwd' (passwords on the command
14098 line, '-stdin' option, '-in ...' option) and thus should not
14099 cause any problems.
14103 * Remove all references to RSAref, since there's no more need for it.
14107 * Make DSO load along a path given through an environment variable
14108 (SHLIB_PATH) with shl_load().
14112 * Constify the ENGINE code as a result of BIGNUM constification.
14113 Also constify the RSA code and most things related to it. In a
14114 few places, most notable in the depth of the ASN.1 code, ugly
14115 casts back to non-const were required (to be solved at a later
14120 * Make it so the openssl application has all engines loaded by default.
14124 * Constify the BIGNUM routines a little more.
14128 * Add the following functions:
14130 ENGINE_load_cswift()
14132 ENGINE_load_atalla()
14133 ENGINE_load_nuron()
14134 ENGINE_load_builtin_engines()
14136 That way, an application can itself choose if external engines that
14137 are built-in in OpenSSL shall ever be used or not. The benefit is
14138 that applications won't have to be linked with libdl or other dso
14139 libraries unless it's really needed.
14141 Changed 'openssl engine' to load all engines on demand.
14142 Changed the engine header files to avoid the duplication of some
14143 declarations (they differed!).
14147 * 'openssl engine' can now list capabilities.
14151 * Better error reporting in 'openssl engine'.
14155 * Never call load_dh_param(NULL) in s_server.
14159 * Add engine application. It can currently list engines by name and
14160 identity, and test if they are actually available.
14164 * Improve RPM specification file by forcing symbolic linking and making
14165 sure the installed documentation is also owned by root.root.
14167 *Damien Miller <djm@mindrot.org>*
14169 * Give the OpenSSL applications more possibilities to make use of
14170 keys (public as well as private) handled by engines.
14174 * Add OCSP code that comes from CertCo.
14178 * Add VMS support for the Rijndael code.
14182 * Added untested support for Nuron crypto accelerator.
14186 * Add support for external cryptographic devices. This code was
14187 previously distributed separately as the "engine" branch.
14189 *Geoff Thorpe, Richard Levitte*
14191 * Rework the filename-translation in the DSO code. It is now possible to
14192 have far greater control over how a "name" is turned into a filename
14193 depending on the operating environment and any oddities about the
14194 different shared library filenames on each system.
14198 * Support threads on FreeBSD-elf in Configure.
14202 * Fix for SHA1 assembly problem with MASM: it produces
14203 warnings about corrupt line number information when assembling
14204 with debugging information. This is caused by the overlapping
14207 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14210 NCONF_get_number() has no error checking at all. As a replacement,
14211 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14212 promoted strongly. The old NCONF_get_number is kept around for
14213 binary backward compatibility.
14214 Make it possible for methods to load from something other than a BIO,
14215 by providing a function pointer that is given a name instead of a BIO.
14216 For example, this could be used to load configuration data from an
14221 * Fix for non blocking accept BIOs. Added new I/O special reason
14222 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14223 with non blocking I/O was not possible because no retry code was
14224 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14229 * Added the beginnings of Rijndael support.
14233 * Fix for bug in DirectoryString mask setting. Add support for
14234 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14235 to allow certificate printing to more controllable, additional
14236 'certopt' option to 'x509' to allow new printing options to be
14241 * Clean old EAY MD5 hack from e_os.h.
14245 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14247 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14248 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14250 *Joe Orton, Steve Henson*
14252 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14254 * Fix additional bug revealed by the NISCC test suite:
14256 Stop bug triggering large recursion when presented with
14257 certain ASN.1 tags ([CVE-2003-0851])
14261 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14263 * Fix various bugs revealed by running the NISCC test suite:
14265 Stop out of bounds reads in the ASN1 code when presented with
14266 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14268 If verify callback ignores invalid public key errors don't try to check
14269 certificate signature with the NULL public key.
14273 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14274 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14279 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14280 extra data after the compression methods not only for TLS 1.0
14281 but also for SSL 3.0 (as required by the specification).
14283 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14285 * Change X509_certificate_type() to mark the key as exported/exportable
14286 when it's 512 *bits* long, not 512 bytes.
14290 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14292 * Countermeasure against the Klima-Pokorny-Rosa extension of
14293 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14294 a protocol version number mismatch like a decryption error
14295 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14299 * Turn on RSA blinding by default in the default implementation
14300 to avoid a timing attack. Applications that don't want it can call
14301 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14302 They would be ill-advised to do so in most cases.
14304 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14306 * Change RSA blinding code so that it works when the PRNG is not
14307 seeded (in this case, the secret RSA exponent is abused as
14308 an unpredictable seed -- if it is not unpredictable, there
14309 is no point in blinding anyway). Make RSA blinding thread-safe
14310 by remembering the creator's thread ID in rsa->blinding and
14311 having all other threads use local one-time blinding factors
14312 (this requires more computation than sharing rsa->blinding, but
14313 avoids excessive locking; and if an RSA object is not shared
14314 between threads, blinding will still be very fast).
14318 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14320 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14321 via timing by performing a MAC computation even if incorrect
14322 block cipher padding has been found. This is a countermeasure
14323 against active attacks where the attacker has to distinguish
14324 between bad padding and a MAC verification error. ([CVE-2003-0078])
14326 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14327 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14328 Martin Vuagnoux (EPFL, Ilion)*
14330 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14332 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14333 memory from its contents. This is done with a counter that will
14334 place alternating values in each byte. This can be used to solve
14335 two issues: 1) the removal of calls to memset() by highly optimizing
14336 compilers, and 2) cleansing with other values than 0, since those can
14337 be read through on certain media, for example a swap space on disk.
14341 * Bugfix: client side session caching did not work with external caching,
14342 because the session->cipher setting was not restored when reloading
14343 from the external cache. This problem was masked, when
14344 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14345 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14349 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14350 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14352 *Zeev Lieber <zeev-l@yahoo.com>*
14354 * Undo an undocumented change introduced in 0.9.6e which caused
14355 repeated calls to OpenSSL_add_all_ciphers() and
14356 OpenSSL_add_all_digests() to be ignored, even after calling
14361 * Change the default configuration reader to deal with last line not
14362 being properly terminated.
14366 * Change X509_NAME_cmp() so it applies the special rules on handling
14367 DN values that are of type PrintableString, as well as RDNs of type
14368 emailAddress where the value has the type ia5String.
14370 *stefank@valicert.com via Richard Levitte*
14372 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14373 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14374 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14375 the bitwise-OR of the two for use by the majority of applications
14376 wanting this behaviour, and update the docs. The documented
14377 behaviour and actual behaviour were inconsistent and had been
14378 changing anyway, so this is more a bug-fix than a behavioural
14381 *Geoff Thorpe, diagnosed by Nadav Har'El*
14383 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14384 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14388 * Fix initialization code race conditions in
14389 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14390 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14391 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14392 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14393 ssl2_get_cipher_by_char(),
14394 ssl3_get_cipher_by_char().
14396 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14398 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14399 the cached sessions are flushed, as the remove_cb() might use ex_data
14400 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14401 (see [openssl.org #212]).
14403 *Geoff Thorpe, Lutz Jaenicke*
14405 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14406 length, instead of the encoding length to d2i_ASN1_OBJECT.
14410 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14412 * [In 0.9.6g-engine release:]
14413 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14415 *Lynn Gazis <lgazis@rainbow.com>*
14417 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14419 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14420 and get fix the header length calculation.
14421 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14422 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14424 * Use proper error handling instead of 'assertions' in buffer
14425 overflow checks added in 0.9.6e. This prevents DoS (the
14426 assertions could call abort()).
14428 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14430 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14432 * Add various sanity checks to asn1_get_length() to reject
14433 the ASN1 length bytes if they exceed sizeof(long), will appear
14434 negative or the content length exceeds the length of the
14437 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14439 * Fix cipher selection routines: ciphers without encryption had no flags
14440 for the cipher strength set and where therefore not handled correctly
14441 by the selection routines (PR #130).
14445 * Fix EVP_dsa_sha macro.
14450 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14451 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14452 that was added in OpenSSL 0.9.6d.
14454 As the countermeasure turned out to be incompatible with some
14455 broken SSL implementations, the new option is part of SSL_OP_ALL.
14456 SSL_OP_ALL is usually employed when compatibility with weird SSL
14457 implementations is desired (e.g. '-bugs' option to 's_client' and
14458 's_server'), so the new option is automatically set in many
14463 * Changes in security patch:
14465 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14466 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14467 Air Force Materiel Command, USAF, under agreement number
14470 * Add various sanity checks to asn1_get_length() to reject
14471 the ASN1 length bytes if they exceed sizeof(long), will appear
14472 negative or the content length exceeds the length of the
14473 supplied buffer. ([CVE-2002-0659])
14475 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14477 * Assertions for various potential buffer overflows, not known to
14478 happen in practice.
14480 *Ben Laurie (CHATS)*
14482 * Various temporary buffers to hold ASCII versions of integers were
14483 too small for 64 bit platforms. ([CVE-2002-0655])
14484 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14486 * Remote buffer overflow in SSL3 protocol - an attacker could
14487 supply an oversized session ID to a client. ([CVE-2002-0656])
14489 *Ben Laurie (CHATS)*
14491 * Remote buffer overflow in SSL2 protocol - an attacker could
14492 supply an oversized client master key. ([CVE-2002-0656])
14494 *Ben Laurie (CHATS)*
14496 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14498 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14499 encoded as NULL) with id-dsa-with-sha1.
14501 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14503 * Check various `X509_...()` return values in `apps/req.c`.
14505 *Nils Larsch <nla@trustcenter.de>*
14507 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14508 an end-of-file condition would erroneously be flagged, when the CRLF
14509 was just at the end of a processed block. The bug was discovered when
14510 processing data through a buffering memory BIO handing the data to a
14511 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14512 <ptsekov@syntrex.com> and Nedelcho Stanev.
14516 * Implement a countermeasure against a vulnerability recently found
14517 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14518 before application data chunks to avoid the use of known IVs
14519 with data potentially chosen by the attacker.
14523 * Fix length checks in ssl3_get_client_hello().
14527 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14528 to prevent ssl3_read_internal() from incorrectly assuming that
14529 ssl3_read_bytes() found application data while handshake
14530 processing was enabled when in fact s->s3->in_read_app_data was
14531 merely automatically cleared during the initial handshake.
14533 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14535 * Fix object definitions for Private and Enterprise: they were not
14536 recognized in their shortname (=lowercase) representation. Extend
14537 obj_dat.pl to issue an error when using undefined keywords instead
14538 of silently ignoring the problem (Svenning Sorensen
14539 <sss@sss.dnsalias.net>).
14543 * Fix DH_generate_parameters() so that it works for 'non-standard'
14544 generators, i.e. generators other than 2 and 5. (Previously, the
14545 code did not properly initialise the 'add' and 'rem' values to
14546 BN_generate_prime().)
14548 In the new general case, we do not insist that 'generator' is
14549 actually a primitive root: This requirement is rather pointless;
14550 a generator of the order-q subgroup is just as good, if not
14555 * Map new X509 verification errors to alerts. Discovered and submitted by
14556 Tom Wu <tom@arcot.com>.
14560 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14561 returning non-zero before the data has been completely received
14562 when using non-blocking I/O.
14564 *Bodo Moeller; problem pointed out by John Hughes*
14566 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14568 *Ben Laurie, Lutz Jaenicke*
14570 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14571 Yoram Zahavi <YoramZ@gilian.com>).
14575 * Add information about CygWin 1.3 and on, and preserve proper
14576 configuration for the versions before that.
14578 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14580 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14581 check whether we deal with a copy of a session and do not delete from
14582 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14583 <izhar@checkpoint.com>.
14587 * Do not store session data into the internal session cache, if it
14588 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14589 flag is set). Proposed by Aslam <aslam@funk.com>.
14593 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14598 * [In 0.9.6d-engine release:]
14599 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14601 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14603 * Add the configuration target linux-s390x.
14605 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14607 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14608 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14609 variable as an indication that a ClientHello message has been
14610 received. As the flag value will be lost between multiple
14611 invocations of ssl3_accept when using non-blocking I/O, the
14612 function may not be aware that a handshake has actually taken
14613 place, thus preventing a new session from being added to the
14616 To avoid this problem, we now set s->new_session to 2 instead of
14617 using a local variable.
14619 *Lutz Jaenicke, Bodo Moeller*
14621 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14622 if the SSL_R_LENGTH_MISMATCH error is detected.
14624 *Geoff Thorpe, Bodo Moeller*
14626 * New 'shared_ldflag' column in Configure platform table.
14630 * Fix EVP_CIPHER_mode macro.
14632 *"Dan S. Camper" <dan@bti.net>*
14634 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14635 type, we must throw them away by setting rr->length to 0.
14637 *D P Chang <dpc@qualys.com>*
14639 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14641 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14642 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14643 worked incorrectly for those cases where range = `10..._2` and
14644 `3*range` is two bits longer than range.)
14648 * Only add signing time to PKCS7 structures if it is not already
14653 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14654 OBJ_ld_ce should be OBJ_id_ce.
14655 Also some ip-pda OIDs in crypto/objects/objects.txt were
14656 incorrect (cf. RFC 3039).
14658 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14660 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14661 returns early because it has nothing to do.
14663 *Andy Schneider <andy.schneider@bjss.co.uk>*
14665 * [In 0.9.6c-engine release:]
14666 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14668 *Andy Schneider <andy.schneider@bjss.co.uk>*
14670 * [In 0.9.6c-engine release:]
14671 Add support for Cryptographic Appliance's keyserver technology.
14672 (Use engine 'keyclient')
14674 *Cryptographic Appliances and Geoff Thorpe*
14676 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14677 is called via tools/c89.sh because arguments have to be
14678 rearranged (all '-L' options must appear before the first object
14681 *Richard Shapiro <rshapiro@abinitio.com>*
14683 * [In 0.9.6c-engine release:]
14684 Add support for Broadcom crypto accelerator cards, backported
14687 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14689 * [In 0.9.6c-engine release:]
14690 Add support for SureWare crypto accelerator cards from
14691 Baltimore Technologies. (Use engine 'sureware')
14693 *Baltimore Technologies and Mark Cox*
14695 * [In 0.9.6c-engine release:]
14696 Add support for crypto accelerator cards from Accelerated
14697 Encryption Processing, www.aep.ie. (Use engine 'aep')
14699 *AEP Inc. and Mark Cox*
14701 * Add a configuration entry for gcc on UnixWare.
14703 *Gary Benson <gbenson@redhat.com>*
14705 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14706 messages are stored in a single piece (fixed-length part and
14707 variable-length part combined) and fix various bugs found on the way.
14711 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14712 instead. BIO_gethostbyname() does not know what timeouts are
14713 appropriate, so entries would stay in cache even when they have
14715 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14717 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14718 faced with a pathologically small ClientHello fragment that does
14719 not contain client_version: Instead of aborting with an error,
14720 simply choose the highest available protocol version (i.e.,
14721 TLS 1.0 unless it is disabled). In practice, ClientHello
14722 messages are never sent like this, but this change gives us
14723 strictly correct behaviour at least for TLS.
14727 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14728 never resets s->method to s->ctx->method when called from within
14729 one of the SSL handshake functions.
14731 *Bodo Moeller; problem pointed out by Niko Baric*
14733 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14734 (sent using the client's version number) if client_version is
14735 smaller than the protocol version in use. Also change
14736 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14737 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14738 the client will at least see that alert.
14742 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14747 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14748 client receives HelloRequest while in a handshake.
14750 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14752 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14753 should end in 'break', not 'goto end' which circumvents various
14754 cleanups done in state SSL_ST_OK. But session related stuff
14755 must be disabled for SSL_ST_OK in the case that we just sent a
14758 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14759 before just sending a HelloRequest.
14761 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14763 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14764 reveal whether illegal block cipher padding was found or a MAC
14765 verification error occurred. (Neither SSLerr() codes nor alerts
14766 are directly visible to potential attackers, but the information
14767 may leak via logfiles.)
14769 Similar changes are not required for the SSL 2.0 implementation
14770 because the number of padding bytes is sent in clear for SSL 2.0,
14771 and the extra bytes are just ignored. However ssl/s2_pkt.c
14772 failed to verify that the purported number of padding bytes is in
14777 * Add OpenUNIX-8 support including shared libraries
14778 (Boyd Lynn Gerber <gerberb@zenez.com>).
14782 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14783 'wristwatch attack' using huge encoding parameters (cf.
14784 James H. Manger's CRYPTO 2001 paper). Note that the
14785 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14786 encoding parameters and hence was not vulnerable.
14790 * BN_sqr() bug fix.
14792 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14794 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14795 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14796 followed by modular reduction.
14798 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14800 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14801 equivalent based on BN_pseudo_rand() instead of BN_rand().
14805 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14806 This function was broken, as the check for a new client hello message
14807 to handle SGC did not allow these large messages.
14808 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14812 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14816 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14817 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14821 * Rework the configuration and shared library support for Tru64 Unix.
14822 The configuration part makes use of modern compiler features and
14823 still retains old compiler behavior for those that run older versions
14824 of the OS. The shared library support part includes a variant that
14825 uses the RPATH feature, and is available through the special
14826 configuration target "alpha-cc-rpath", which will never be selected
14829 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14831 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14832 with the same message size as in ssl3_get_certificate_request().
14833 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14834 messages might inadvertently be reject as too long.
14836 *Petr Lampa <lampa@fee.vutbr.cz>*
14838 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14842 * Modified SSL library such that the verify_callback that has been set
14843 specifically for an SSL object with SSL_set_verify() is actually being
14844 used. Before the change, a verify_callback set with this function was
14845 ignored and the verify_callback() set in the SSL_CTX at the time of
14846 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14847 to allow the necessary settings.
14851 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14852 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14853 done automatically (in contradiction to the requirements of the C
14854 standard). This made problems when used from OpenSSH.
14858 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14859 dh->length and always used
14861 BN_rand_range(priv_key, dh->p).
14863 BN_rand_range() is not necessary for Diffie-Hellman, and this
14864 specific range makes Diffie-Hellman unnecessarily inefficient if
14865 dh->length (recommended exponent length) is much smaller than the
14866 length of dh->p. We could use BN_rand_range() if the order of
14867 the subgroup was stored in the DH structure, but we only have
14872 BN_rand(priv_key, l, ...)
14874 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14881 RSA_eay_public_encrypt
14882 RSA_eay_private_decrypt
14883 RSA_eay_private_encrypt (signing)
14884 RSA_eay_public_decrypt (signature verification)
14886 (default implementations for RSA_public_encrypt,
14887 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14888 always reject numbers >= n.
14892 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14893 to synchronize access to 'locking_thread'. This is necessary on
14894 systems where access to 'locking_thread' (an 'unsigned long'
14895 variable) is not atomic.
14899 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14900 *before* setting the 'crypto_lock_rand' flag. The previous code had
14901 a race condition if 0 is a valid thread ID.
14903 *Travis Vitek <vitek@roguewave.com>*
14905 * Add support for shared libraries under Irix.
14907 *Albert Chin-A-Young <china@thewrittenword.com>*
14909 * Add configuration option to build on Linux on both big-endian and
14910 little-endian MIPS.
14912 *Ralf Baechle <ralf@uni-koblenz.de>*
14914 * Add the possibility to create shared libraries on HP-UX.
14918 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14920 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14921 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14922 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14923 PRNG state recovery was possible based on the output of
14924 one PRNG request appropriately sized to gain knowledge on
14925 'md' followed by enough consecutive 1-byte PRNG requests
14926 to traverse all of 'state'.
14928 1. When updating 'md_local' (the current thread's copy of 'md')
14929 during PRNG output generation, hash all of the previous
14930 'md_local' value, not just the half used for PRNG output.
14932 2. Make the number of bytes from 'state' included into the hash
14933 independent from the number of PRNG bytes requested.
14935 The first measure alone would be sufficient to avoid
14936 Markku-Juhani's attack. (Actually it had never occurred
14937 to me that the half of 'md_local' used for chaining was the
14938 half from which PRNG output bytes were taken -- I had always
14939 assumed that the secret half would be used.) The second
14940 measure makes sure that additional data from 'state' is never
14941 mixed into 'md_local' in small portions; this heuristically
14942 further strengthens the PRNG.
14946 * Fix crypto/bn/asm/mips3.s.
14950 * When only the key is given to "enc", the IV is undefined. Print out
14951 an error message in this case.
14955 * Handle special case when X509_NAME is empty in X509 printing routines.
14959 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14960 positive and less than q.
14964 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14965 used: it isn't thread safe and the add_lock_callback should handle
14968 *Paul Rose <Paul.Rose@bridge.com>*
14970 * Verify that incoming data obeys the block size in
14971 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14977 *Ulf Möller, Bodo Möller*
14979 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14980 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14981 when fixing the server behaviour for backwards-compatible 'client
14982 hello' messages. (Note that the attack is impractical against
14983 SSL 3.0 and TLS 1.0 anyway because length and version checking
14984 means that the probability of guessing a valid ciphertext is
14985 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14988 Before 0.9.5, the countermeasure (hide the error by generating a
14989 random 'decryption result') did not work properly because
14990 ERR_clear_error() was missing, meaning that SSL_get_error() would
14991 detect the supposedly ignored error.
14993 Both problems are now fixed.
14997 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14998 (previously it was 1024).
15002 * Fix for compatibility mode trust settings: ignore trust settings
15003 unless some valid trust or reject settings are present.
15007 * Fix for blowfish EVP: its a variable length cipher.
15011 * Fix various bugs related to DSA S/MIME verification. Handle missing
15012 parameters in DSA public key structures and return an error in the
15013 DSA routines if parameters are absent.
15017 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15018 in the current directory if neither $RANDFILE nor $HOME was set.
15019 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15020 caused some confusion to Windows users who haven't defined $HOME.
15021 Thus RAND_file_name() is changed again: e_os.h can define a
15022 DEFAULT_HOME, which will be used if $HOME is not set.
15023 For Windows, we use "C:"; on other platforms, we still require
15024 environment variables.
15026 * Move 'if (!initialized) RAND_poll()' into regions protected by
15027 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15028 having multiple threads call RAND_poll() concurrently.
15032 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15033 combination of a flag and a thread ID variable.
15034 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15035 flag), *other* threads can enter ssleay_add_bytes without obeying
15036 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15037 that they do not hold after the first thread unsets add_do_not_lock).
15041 * Change bctest again: '-x' expressions are not available in all
15042 versions of 'test'.
15046 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15048 * Fix a couple of memory leaks in PKCS7_dataDecode()
15050 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15052 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15053 the default extension for executables, if any. Also, make the perl
15054 scripts that use symlink() to test if it really exists and use "cp"
15055 if it doesn't. All this made OpenSSL compilable and installable in
15060 * Fix for asn1_GetSequence() for indefinite length constructed data.
15061 If SEQUENCE is length is indefinite just set c->slen to the total
15062 amount of data available.
15064 *Steve Henson, reported by shige@FreeBSD.org*
15066 *This change does not apply to 0.9.7.*
15068 * Change bctest to avoid here-documents inside command substitution
15069 (workaround for FreeBSD /bin/sh bug).
15070 For compatibility with Ultrix, avoid shell functions (introduced
15071 in the bctest version that searches along $PATH).
15075 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15076 with des_encrypt() defined on some operating systems, like Solaris
15081 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15082 On the Importance of Eliminating Errors in Cryptographic
15083 Computations, J. Cryptology 14 (2001) 2, 101-119,
15084 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15088 * MIPS assembler BIGNUM division bug fix.
15092 * Disabled incorrect Alpha assembler code.
15096 * Fix PKCS#7 decode routines so they correctly update the length
15097 after reading an EOC for the EXPLICIT tag.
15101 *This change does not apply to 0.9.7.*
15103 * Fix bug in PKCS#12 key generation routines. This was triggered
15104 if a 3DES key was generated with a 0 initial byte. Include
15105 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15106 (but broken) behaviour.
15110 * Enhance bctest to search for a working bc along $PATH and print
15113 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15115 * Fix memory leaks in err.c: free err_data string if necessary;
15116 don't write to the wrong index in ERR_set_error_data.
15120 * Implement ssl23_peek (analogous to ssl23_read), which previously
15125 * Replace rdtsc with `_emit` statements for VC++ version 5.
15127 *Jeremy Cooper <jeremy@baymoo.org>*
15129 * Make it possible to reuse SSLv2 sessions.
15133 * In copy_email() check for >= 0 as a return value for
15134 X509_NAME_get_index_by_NID() since 0 is a valid index.
15136 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15138 * Avoid coredump with unsupported or invalid public keys by checking if
15139 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15140 PKCS7_verify() fails with non detached data.
15144 * Don't use getenv in library functions when run as setuid/setgid.
15145 New function OPENSSL_issetugid().
15149 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15150 due to incorrect handling of multi-threading:
15152 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15154 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15156 3. Count how many times MemCheck_off() has been called so that
15157 nested use can be treated correctly. This also avoids
15158 inband-signalling in the previous code (which relied on the
15159 assumption that thread ID 0 is impossible).
15163 * Add "-rand" option also to s_client and s_server.
15167 * Fix CPU detection on Irix 6.x.
15168 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15169 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15171 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15176 *This change does not apply to 0.9.7.*
15178 * Use the cached encoding of an X509_NAME structure rather than
15179 copying it. This is apparently the reason for the libsafe "errors"
15180 but the code is actually correct.
15184 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15185 Bleichenbacher's DSA attack.
15186 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15187 to be set and top=0 forces the highest bit to be set; top=-1 is new
15188 and leaves the highest bit random.
15190 *Ulf Moeller, Bodo Moeller*
15192 * In the `NCONF_...`-based implementations for `CONF_...` queries
15193 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15194 a temporary CONF structure with the data component set to NULL
15195 (which gives segmentation faults in lh_retrieve).
15196 Instead, use NULL for the CONF pointer in CONF_get_string and
15197 CONF_get_number (which may use environment variables) and directly
15198 return NULL from CONF_get_section.
15202 * Fix potential buffer overrun for EBCDIC.
15206 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15207 keyUsage if basicConstraints absent for a CA.
15211 * Make SMIME_write_PKCS7() write mail header values with a format that
15212 is more generally accepted (no spaces before the semicolon), since
15213 some programs can't parse those values properly otherwise. Also make
15214 sure BIO's that break lines after each write do not create invalid
15219 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15220 macros previously used would not encode an empty SEQUENCE OF
15221 and break the signature.
15225 *This change does not apply to 0.9.7.*
15227 * Zero the premaster secret after deriving the master secret in
15232 * Add some EVP_add_digest_alias registrations (as found in
15233 OpenSSL_add_all_digests()) to SSL_library_init()
15234 aka OpenSSL_add_ssl_algorithms(). This provides improved
15235 compatibility with peers using X.509 certificates
15236 with unconventional AlgorithmIdentifier OIDs.
15240 * Fix for Irix with NO_ASM.
15242 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15244 * ./config script fixes.
15246 *Ulf Moeller, Richard Levitte*
15248 * Fix 'openssl passwd -1'.
15252 * Change PKCS12_key_gen_asc() so it can cope with non null
15253 terminated strings whose length is passed in the passlen
15254 parameter, for example from PEM callbacks. This was done
15255 by adding an extra length parameter to asc2uni().
15257 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15259 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15260 call failed, free the DSA structure.
15264 * Fix to uni2asc() to cope with zero length Unicode strings.
15265 These are present in some PKCS#12 files.
15269 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15270 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15271 when writing a 32767 byte record.
15273 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15275 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15276 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15278 (RSA objects have a reference count access to which is protected
15279 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15280 so they are meant to be shared between threads.)
15281 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15282 "Reddie, Steven" <Steven.Reddie@ca.com>*
15284 * Fix a deadlock in CRYPTO_mem_leaks().
15288 * Use better test patterns in bntest.
15292 * rand_win.c fix for Borland C.
15296 * BN_rshift bugfix for n == 0.
15300 * Add a 'bctest' script that checks for some known 'bc' bugs
15301 so that 'make test' does not abort just because 'bc' is broken.
15305 * Store verify_result within SSL_SESSION also for client side to
15306 avoid potential security hole. (Re-used sessions on the client side
15307 always resulted in verify_result==X509_V_OK, not using the original
15308 result of the server certificate verification.)
15312 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15313 SSL3_RT_APPLICATION_DATA, return 0.
15314 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15319 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15320 releases, have been re-implemented by renaming the previous
15321 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15322 and ssl3_read_internal, respectively, and adding 'peek' parameters
15323 to them. The new ssl[23]_{read,peek} functions are calls to
15324 ssl[23]_read_internal with the 'peek' flag set appropriately.
15325 A 'peek' parameter has also been added to ssl3_read_bytes, which
15326 does the actual work for ssl3_read_internal.
15330 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15331 the method-specific "init()" handler. Also clean up ex_data after
15332 calling the method-specific "finish()" handler. Previously, this was
15333 happening the other way round.
15337 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15338 The previous value, 12, was not always sufficient for BN_mod_exp().
15342 * Make sure that shared libraries get the internal name engine with
15343 the full version number and not just 0. This should mark the
15344 shared libraries as not backward compatible. Of course, this should
15345 be changed again when we can guarantee backward binary compatibility.
15349 * Fix typo in get_cert_by_subject() in by_dir.c
15351 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15353 * Rework the system to generate shared libraries:
15355 - Make note of the expected extension for the shared libraries and
15356 if there is a need for symbolic links from for example libcrypto.so.0
15357 to libcrypto.so.0.9.7. There is extended info in Configure for
15360 - Make as few rebuilds of the shared libraries as possible.
15362 - Still avoid linking the OpenSSL programs with the shared libraries.
15364 - When installing, install the shared libraries separately from the
15369 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15371 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15372 and not in SSL_clear because the latter is also used by the
15373 accept/connect functions; previously, the settings made by
15374 SSL_set_read_ahead would be lost during the handshake.
15376 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15378 * Correct util/mkdef.pl to be selective about disabled algorithms.
15379 Previously, it would create entries for disabled algorithms no
15384 * Added several new manual pages for SSL_* function.
15388 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15390 * In ssl23_get_client_hello, generate an error message when faced
15391 with an initial SSL 3.0/TLS record that is too small to contain the
15392 first two bytes of the ClientHello message, i.e. client_version.
15393 (Note that this is a pathologic case that probably has never happened
15394 in real life.) The previous approach was to use the version number
15395 from the record header as a substitute; but our protocol choice
15396 should not depend on that one because it is not authenticated
15397 by the Finished messages.
15401 * More robust randomness gathering functions for Windows.
15403 *Jeffrey Altman <jaltman@columbia.edu>*
15405 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15406 not set then we don't setup the error code for issuer check errors
15407 to avoid possibly overwriting other errors which the callback does
15408 handle. If an application does set the flag then we assume it knows
15409 what it is doing and can handle the new informational codes
15414 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15415 a general "ANY" type, as such it should be able to decode anything
15416 including tagged types. However it didn't check the class so it would
15417 wrongly interpret tagged types in the same way as their universal
15418 counterpart and unknown types were just rejected. Changed so that the
15419 tagged and unknown types are handled in the same way as a SEQUENCE:
15420 that is the encoding is stored intact. There is also a new type
15421 "V_ASN1_OTHER" which is used when the class is not universal, in this
15422 case we have no idea what the actual type is so we just lump them all
15427 * On VMS, stdout may very well lead to a file that is written to
15428 in a record-oriented fashion. That means that every write() will
15429 write a separate record, which will be read separately by the
15430 programs trying to read from it. This can be very confusing.
15432 The solution is to put a BIO filter in the way that will buffer
15433 text until a linefeed is reached, and then write everything a
15434 line at a time, so every record written will be an actual line,
15435 not chunks of lines and not (usually doesn't happen, but I've
15436 seen it once) several lines in one record. BIO_f_linebuffer() is
15439 Currently, it's a VMS-only method, because that's where it has
15440 been tested well enough.
15444 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15445 it can return incorrect results.
15446 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15447 but it was in 0.9.6-beta[12].)
15451 * Disable the check for content being present when verifying detached
15452 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15453 include zero length content when signing messages.
15457 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15458 BIO_ctrl (for BIO pairs).
15462 * Add DSO method for VMS.
15466 * Bug fix: Montgomery multiplication could produce results with the
15471 * Add RPM specification openssl.spec and modify it to build three
15472 packages. The default package contains applications, application
15473 documentation and run-time libraries. The devel package contains
15474 include files, static libraries and function documentation. The
15475 doc package contains the contents of the doc directory. The original
15476 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15480 * Add a large number of documentation files for many SSL routines.
15482 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15484 * Add a configuration entry for Sony News 4.
15486 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15488 * Don't set the two most significant bits to one when generating a
15489 random number < q in the DSA library.
15493 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15494 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15495 the underlying transport is blocking) if a handshake took place.
15496 (The default behaviour is needed by applications such as s_client
15497 and s_server that use select() to determine when to use SSL_read;
15498 but for applications that know in advance when to expect data, it
15499 just makes things more complicated.)
15503 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15508 * Add a few more EBCDIC conditionals that make `req` and `x509`
15509 work better on such systems.
15511 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15513 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15514 Update PKCS12_parse() so it copies the friendlyName and the
15515 keyid to the certificates aux info.
15519 * Fix bug in PKCS7_verify() which caused an infinite loop
15520 if there was more than one signature.
15522 *Sven Uszpelkat <su@celocom.de>*
15524 * Major change in util/mkdef.pl to include extra information
15525 about each symbol, as well as presenting variables as well
15526 as functions. This change means that there's n more need
15527 to rebuild the .num files when some algorithms are excluded.
15531 * Allow the verify time to be set by an application,
15532 rather than always using the current time.
15536 * Phase 2 verify code reorganisation. The certificate
15537 verify code now looks up an issuer certificate by a
15538 number of criteria: subject name, authority key id
15539 and key usage. It also verifies self signed certificates
15540 by the same criteria. The main comparison function is
15541 X509_check_issued() which performs these checks.
15543 Lot of changes were necessary in order to support this
15544 without completely rewriting the lookup code.
15546 Authority and subject key identifier are now cached.
15548 The LHASH 'certs' is X509_STORE has now been replaced
15549 by a STACK_OF(X509_OBJECT). This is mainly because an
15550 LHASH can't store or retrieve multiple objects with
15551 the same hash value.
15553 As a result various functions (which were all internal
15554 use only) have changed to handle the new X509_STORE
15555 structure. This will break anything that messed round
15556 with X509_STORE internally.
15558 The functions X509_STORE_add_cert() now checks for an
15559 exact match, rather than just subject name.
15561 The X509_STORE API doesn't directly support the retrieval
15562 of multiple certificates matching a given criteria, however
15563 this can be worked round by performing a lookup first
15564 (which will fill the cache with candidate certificates)
15565 and then examining the cache for matches. This is probably
15566 the best we can do without throwing out X509_LOOKUP
15567 entirely (maybe later...).
15569 The X509_VERIFY_CTX structure has been enhanced considerably.
15571 All certificate lookup operations now go via a get_issuer()
15572 callback. Although this currently uses an X509_STORE it
15573 can be replaced by custom lookups. This is a simple way
15574 to bypass the X509_STORE hackery necessary to make this
15575 work and makes it possible to use more efficient techniques
15576 in future. A very simple version which uses a simple
15577 STACK for its trusted certificate store is also provided
15578 using X509_STORE_CTX_trusted_stack().
15580 The verify_cb() and verify() callbacks now have equivalents
15581 in the X509_STORE_CTX structure.
15583 X509_STORE_CTX also has a 'flags' field which can be used
15584 to customise the verify behaviour.
15588 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15589 excludes S/MIME capabilities.
15593 * When a certificate request is read in keep a copy of the
15594 original encoding of the signed data and use it when outputting
15595 again. Signatures then use the original encoding rather than
15596 a decoded, encoded version which may cause problems if the
15597 request is improperly encoded.
15601 * For consistency with other BIO_puts implementations, call
15602 buffer_write(b, ...) directly in buffer_puts instead of calling
15605 In BIO_puts, increment b->num_write as in BIO_write.
15607 *Peter.Sylvester@EdelWeb.fr*
15609 * Fix BN_mul_word for the case where the word is 0. (We have to use
15610 BN_zero, we may not return a BIGNUM with an array consisting of
15611 words set to zero.)
15615 * Avoid calling abort() from within the library when problems are
15616 detected, except if preprocessor symbols have been defined
15617 (such as REF_CHECK, BN_DEBUG etc.).
15621 * New openssl application 'rsautl'. This utility can be
15622 used for low-level RSA operations. DER public key
15623 BIO/fp routines also added.
15627 * New Configure entry and patches for compiling on QNX 4.
15629 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15631 * A demo state-machine implementation was sponsored by
15632 Nuron (<http://www.nuron.com/>) and is now available in
15633 demos/state_machine.
15637 * New options added to the 'dgst' utility for signature
15638 generation and verification.
15642 * Unrecognized PKCS#7 content types are now handled via a
15643 catch all ASN1_TYPE structure. This allows unsupported
15644 types to be stored as a "blob" and an application can
15645 encode and decode it manually.
15649 * Fix various signed/unsigned issues to make a_strex.c
15650 compile under VC++.
15652 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15654 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15655 length if passed a buffer. ASN1_INTEGER_to_BN failed
15656 if passed a NULL BN and its argument was negative.
15658 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15660 * Modification to PKCS#7 encoding routines to output definite
15661 length encoding. Since currently the whole structures are in
15662 memory there's not real point in using indefinite length
15663 constructed encoding. However if OpenSSL is compiled with
15664 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15668 * Added BIO_vprintf() and BIO_vsnprintf().
15672 * Added more prefixes to parse for in the strings written
15673 through a logging bio, to cover all the levels that are available
15674 through syslog. The prefixes are now:
15676 PANIC, EMERG, EMR => LOG_EMERG
15677 ALERT, ALR => LOG_ALERT
15678 CRIT, CRI => LOG_CRIT
15679 ERROR, ERR => LOG_ERR
15680 WARNING, WARN, WAR => LOG_WARNING
15681 NOTICE, NOTE, NOT => LOG_NOTICE
15682 INFO, INF => LOG_INFO
15683 DEBUG, DBG => LOG_DEBUG
15685 and as before, if none of those prefixes are present at the
15686 beginning of the string, LOG_ERR is chosen.
15688 On Win32, the `LOG_*` levels are mapped according to this:
15690 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15691 LOG_WARNING => EVENTLOG_WARNING_TYPE
15692 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15696 * Made it possible to reconfigure with just the configuration
15697 argument "reconf" or "reconfigure". The command line arguments
15698 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15699 and are retrieved from there when reconfiguring.
15705 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15707 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15711 * The obj_dat.pl script was messing up the sorting of object
15712 names. The reason was that it compared the quoted version
15713 of strings as a result "OCSP" > "OCSP Signing" because
15714 " > SPACE. Changed script to store unquoted versions of
15715 names and add quotes on output. It was also omitting some
15716 names from the lookup table if they were given a default
15717 value (that is if SN is missing it is given the same
15718 value as LN and vice versa), these are now added on the
15719 grounds that if an object has a name we should be able to
15720 look it up. Finally added warning output when duplicate
15721 short or long names are found.
15725 * Changes needed for Tandem NSK.
15727 *Scott Uroff <scott@xypro.com>*
15729 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15730 RSA_padding_check_SSLv23(), special padding was never detected
15731 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15732 version rollback attacks was not effective.
15734 In s23_clnt.c, don't use special rollback-attack detection padding
15735 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15736 client; similarly, in s23_srvr.c, don't do the rollback check if
15737 SSL 2.0 is the only protocol enabled in the server.
15741 * Make it possible to get hexdumps of unprintable data with 'openssl
15742 asn1parse'. By implication, the functions ASN1_parse_dump() and
15743 BIO_dump_indent() are added.
15747 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15748 these print out strings and name structures based on various
15749 flags including RFC2253 support and proper handling of
15750 multibyte characters. Added options to the 'x509' utility
15751 to allow the various flags to be set.
15755 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15756 Also change the functions X509_cmp_current_time() and
15757 X509_gmtime_adj() work with an ASN1_TIME structure,
15758 this will enable certificates using GeneralizedTime in validity
15759 dates to be checked.
15763 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15764 negative public key encodings) on by default,
15765 NO_NEG_PUBKEY_BUG can be set to disable it.
15769 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15770 content octets. An i2c_ASN1_OBJECT is unnecessary because
15771 the encoding can be trivially obtained from the structure.
15775 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15776 not read locks (`CRYPTO_r_[un]lock`).
15780 * A first attempt at creating official support for shared
15781 libraries through configuration. I've kept it so the
15782 default is static libraries only, and the OpenSSL programs
15783 are always statically linked for now, but there are
15784 preparations for dynamic linking in place.
15785 This has been tested on Linux and Tru64.
15789 * Randomness polling function for Win9x, as described in:
15790 Peter Gutmann, Software Generation of Practically Strong
15795 * Fix so PRNG is seeded in req if using an already existing
15800 * New options to smime application. -inform and -outform
15801 allow alternative formats for the S/MIME message including
15802 PEM and DER. The -content option allows the content to be
15803 specified separately. This should allow things like Netscape
15804 form signing output easier to verify.
15808 * Fix the ASN1 encoding of tags using the 'long form'.
15812 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15813 STRING types. These convert content octets to and from the
15814 underlying type. The actual tag and length octets are
15815 already assumed to have been read in and checked. These
15816 are needed because all other string types have virtually
15817 identical handling apart from the tag. By having versions
15818 of the ASN1 functions that just operate on content octets
15819 IMPLICIT tagging can be handled properly. It also allows
15820 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15821 and ASN1_INTEGER are identical apart from the tag.
15825 * Change the handling of OID objects as follows:
15827 - New object identifiers are inserted in objects.txt, following
15828 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15829 - objects.pl is used to process obj_mac.num and create a new
15831 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15834 This is currently kind of a hack, and the perl code in objects.pl
15835 isn't very elegant, but it works as I intended. The simplest way
15836 to check that it worked correctly is to look in obj_dat.h and
15837 check the array nid_objs and make sure the objects haven't moved
15838 around (this is important!). Additions are OK, as well as
15839 consistent name changes.
15843 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15847 * Addition of the command line parameter '-rand file' to 'openssl req'.
15848 The given file adds to whatever has already been seeded into the
15849 random pool through the RANDFILE configuration file option or
15850 environment variable, or the default random state file.
15854 * mkstack.pl now sorts each macro group into lexical order.
15855 Previously the output order depended on the order the files
15856 appeared in the directory, resulting in needless rewriting
15861 * Patches to make OpenSSL compile under Win32 again. Mostly
15862 work arounds for the VC++ problem that it treats func() as
15863 func(void). Also stripped out the parts of mkdef.pl that
15864 added extra typesafe functions: these no longer exist.
15868 * Reorganisation of the stack code. The macros are now all
15869 collected in safestack.h . Each macro is defined in terms of
15870 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15871 DEBUG_SAFESTACK is now handled in terms of function casts,
15872 this has the advantage of retaining type safety without the
15873 use of additional functions. If DEBUG_SAFESTACK is not defined
15874 then the non typesafe macros are used instead. Also modified the
15875 mkstack.pl script to handle the new form. Needs testing to see
15876 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15877 the default if no major problems. Similar behaviour for ASN1_SET_OF
15878 and PKCS12_STACK_OF.
15882 * When some versions of IIS use the 'NET' form of private key the
15883 key derivation algorithm is different. Normally MD5(password) is
15884 used as a 128 bit RC4 key. In the modified case
15885 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15886 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15887 as the old Netscape_RSA functions except they have an additional
15888 'sgckey' parameter which uses the modified algorithm. Also added
15889 an -sgckey command line option to the rsa utility. Thanks to
15890 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15891 algorithm to openssl-dev.
15895 * The evp_local.h macros were using 'c.##kname' which resulted in
15896 invalid expansion on some systems (SCO 5.0.5 for example).
15897 Corrected to 'c.kname'.
15899 *Phillip Porch <root@theporch.com>*
15901 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15902 a STACK of email addresses from a certificate or request, these look
15903 in the subject name and the subject alternative name extensions and
15904 omit any duplicate addresses.
15908 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15909 This makes DSA verification about 2 % faster.
15913 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15914 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15915 plus overhead for 1024 bit moduli).
15916 This makes exponentiations about 0.5 % faster for 1024 bit
15917 exponents (as measured by "openssl speed rsa2048").
15921 * Rename memory handling macros to avoid conflicts with other
15923 Malloc => OPENSSL_malloc
15924 Malloc_locked => OPENSSL_malloc_locked
15925 Realloc => OPENSSL_realloc
15926 Free => OPENSSL_free
15930 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15931 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15935 * CygWin32 support.
15937 *John Jarvie <jjarvie@newsguy.com>*
15939 * The type-safe stack code has been rejigged. It is now only compiled
15940 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15941 by default all type-specific stack functions are "#define"d back to
15942 standard stack functions. This results in more streamlined output
15943 but retains the type-safety checking possibilities of the original
15948 * The STACK code has been cleaned up, and certain type declarations
15949 that didn't make a lot of sense have been brought in line. This has
15950 also involved a cleanup of sorts in safestack.h to more correctly
15951 map type-safe stack functions onto their plain stack counterparts.
15952 This work has also resulted in a variety of "const"ifications of
15953 lots of the code, especially `_cmp` operations which should normally
15954 be prototyped with "const" parameters anyway.
15958 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15959 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15960 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15961 where all of 'md' is used each time the PRNG is used, but 'state'
15962 is used only indexed by a cyclic counter. As entropy may not be
15963 well distributed from the beginning, 'md' is important as a
15964 chaining variable. However, the output function chains only half
15965 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15966 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15967 in all of 'state' being rewritten, with the new values depending
15968 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15972 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15973 the handshake is continued after ssl_verify_cert_chain();
15974 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15975 can lead to 'unexplainable' connection aborts later.
15977 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15979 * Major EVP API cipher revision.
15980 Add hooks for extra EVP features. This allows various cipher
15981 parameters to be set in the EVP interface. Support added for variable
15982 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15983 setting of RC2 and RC5 parameters.
15985 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15988 Remove lots of duplicated code from the EVP library. For example *every*
15989 cipher init() function handles the 'iv' in the same way according to the
15990 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15991 for CFB and OFB modes they zero ctx->num.
15993 New functionality allows removal of S/MIME code RC2 hack.
15995 Most of the routines have the same form and so can be declared in terms
15998 By shifting this to the top level EVP_CipherInit() it can be removed from
15999 all individual ciphers. If the cipher wants to handle IVs or keys
16000 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16003 Change lots of functions like EVP_EncryptUpdate() to now return a
16004 value: although software versions of the algorithms cannot fail
16005 any installed hardware versions can.
16009 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16010 this option is set, tolerate broken clients that send the negotiated
16011 protocol version number instead of the requested protocol version
16016 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16017 i.e. non-zero for export ciphersuites, zero otherwise.
16018 Previous versions had this flag inverted, inconsistent with
16019 rsa_tmp_cb (..._TMP_RSA_CB).
16021 *Bodo Moeller; problem reported by Amit Chopra*
16023 * Add missing DSA library text string. Work around for some IIS
16024 key files with invalid SEQUENCE encoding.
16028 * Add a document (doc/standards.txt) that list all kinds of standards
16029 and so on that are implemented in OpenSSL.
16033 * Enhance c_rehash script. Old version would mishandle certificates
16034 with the same subject name hash and wouldn't handle CRLs at all.
16035 Added -fingerprint option to crl utility, to support new c_rehash
16040 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16044 * Fix for SSL server purpose checking. Server checking was
16045 rejecting certificates which had extended key usage present
16046 but no ssl client purpose.
16048 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16050 * Make PKCS#12 code work with no password. The PKCS#12 spec
16051 is a little unclear about how a blank password is handled.
16052 Since the password in encoded as a BMPString with terminating
16053 double NULL a zero length password would end up as just the
16054 double NULL. However no password at all is different and is
16055 handled differently in the PKCS#12 key generation code. NS
16056 treats a blank password as zero length. MSIE treats it as no
16057 password on export: but it will try both on import. We now do
16058 the same: PKCS12_parse() tries zero length and no password if
16059 the password is set to "" or NULL (NULL is now a valid password:
16060 it wasn't before) as does the pkcs12 application.
16064 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16065 perror when PEM_read_bio_X509_REQ fails, the error message must
16066 be obtained from the error queue.
16070 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16071 it in ERR_remove_state if appropriate, and change ERR_get_state
16072 accordingly to avoid race conditions (this is necessary because
16073 thread_hash is no longer constant once set).
16077 * Bugfix for linux-elf makefile.one.
16081 * RSA_get_default_method() will now cause a default
16082 RSA_METHOD to be chosen if one doesn't exist already.
16083 Previously this was only set during a call to RSA_new()
16084 or RSA_new_method(NULL) meaning it was possible for
16085 RSA_get_default_method() to return NULL.
16089 * Added native name translation to the existing DSO code
16090 that will convert (if the flag to do so is set) filenames
16091 that are sufficiently small and have no path information
16092 into a canonical native form. Eg. "blah" converted to
16093 "libblah.so" or "blah.dll" etc.
16097 * New function ERR_error_string_n(e, buf, len) which is like
16098 ERR_error_string(e, buf), but writes at most 'len' bytes
16099 including the 0 terminator. For ERR_error_string_n, 'buf'
16102 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16104 * CONF library reworked to become more general. A new CONF
16105 configuration file reader "class" is implemented as well as a
16106 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16107 old `CONF_*` functions are still there, but are reimplemented to
16108 work in terms of the new functions. Also, a set of functions
16109 to handle the internal storage of the configuration data is
16110 provided to make it easier to write new configuration file
16111 reader "classes" (I can definitely see something reading a
16112 configuration file in XML format, for example), called `_CONF_*`,
16113 or "the configuration storage API"...
16115 The new configuration file reading functions are:
16117 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16118 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16120 NCONF_default, NCONF_WIN32
16122 NCONF_dump_fp, NCONF_dump_bio
16124 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16125 NCONF_new creates a new CONF object. This works in the same way
16126 as other interfaces in OpenSSL, like the BIO interface.
16127 `NCONF_dump_*` dump the internal storage of the configuration file,
16128 which is useful for debugging. All other functions take the same
16129 arguments as the old `CONF_*` functions with the exception of the
16130 first that must be a `CONF *` instead of a `LHASH *`.
16132 To make it easier to use the new classes with the old `CONF_*` functions,
16133 the function CONF_set_default_method is provided.
16137 * Add '-tls1' option to 'openssl ciphers', which was already
16138 mentioned in the documentation but had not been implemented.
16139 (This option is not yet really useful because even the additional
16140 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16144 * Initial DSO code added into libcrypto for letting OpenSSL (and
16145 OpenSSL-based applications) load shared libraries and bind to
16146 them in a portable way.
16148 *Geoff Thorpe, with contributions from Richard Levitte*
16150 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16152 * Make sure _lrotl and _lrotr are only used with MSVC.
16154 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16155 (the default implementation of RAND_status).
16157 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16158 to '-clrext' (= clear extensions), as intended and documented.
16159 *Bodo Moeller; inconsistency pointed out by Michael Attili
16160 <attili@amaxo.com>*
16162 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16163 was larger than the MD block size.
16165 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16167 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16168 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16169 using the passed key: if the passed key was a private key the result
16170 of X509_print(), for example, would be to print out all the private key
16175 * des_quad_cksum() byte order bug fix.
16176 *Ulf Möller, using the problem description in krb4-0.9.7, where
16177 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16179 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16182 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16184 * For easily testing in shell scripts whether some command
16185 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16186 returns with exit code 0 iff no command of the given name is available.
16187 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16188 the output goes to stdout and nothing is printed to stderr.
16189 Additional arguments are always ignored.
16191 Since for each cipher there is a command of the same name,
16192 the 'no-cipher' compilation switches can be tested this way.
16194 ('openssl no-XXX' is not able to detect pseudo-commands such
16195 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16199 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16203 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16204 is set; it will be thrown away anyway because each handshake creates
16206 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16207 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16208 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16209 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16213 * New s_client option -ign_eof: EOF at stdin is ignored, and
16214 'Q' and 'R' lose their special meanings (quit/renegotiate).
16215 This is part of what -quiet does; unlike -quiet, -ign_eof
16216 does not suppress any output.
16220 * Add compatibility options to the purpose and trust code. The
16221 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16222 accepts a certificate or CA, this was the previous behaviour,
16223 with all the associated security issues.
16225 X509_TRUST_COMPAT is the old trust behaviour: only and
16226 automatically trust self signed roots in certificate store. A
16227 new trust setting X509_TRUST_DEFAULT is used to specify that
16228 a purpose has no associated trust setting and it should instead
16229 use the value in the default purpose.
16233 * Fix the PKCS#8 DSA private key code so it decodes keys again
16234 and fix a memory leak.
16238 * In util/mkerr.pl (which implements 'make errors'), preserve
16239 reason strings from the previous version of the .c file, as
16240 the default to have only downcase letters (and digits) in
16241 automatically generated reasons codes is not always appropriate.
16245 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16246 using strerror. Previously, ERR_reason_error_string() returned
16247 library names as reason strings for SYSerr; but SYSerr is a special
16248 case where small numbers are errno values, not library numbers.
16252 * Add '-dsaparam' option to 'openssl dhparam' application. This
16253 converts DSA parameters into DH parameters. (When creating parameters,
16254 DSA_generate_parameters is used.)
16258 * Include 'length' (recommended exponent length) in C code generated
16259 by 'openssl dhparam -C'.
16263 * The second argument to set_label in perlasm was already being used
16264 so couldn't be used as a "file scope" flag. Moved to third argument
16269 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16270 instead of RAND_bytes for encryption IVs and salts.
16274 * Include RAND_status() into RAND_METHOD instead of implementing
16275 it only for md_rand.c Otherwise replacing the PRNG by calling
16276 RAND_set_rand_method would be impossible.
16280 * Don't let DSA_generate_key() enter an infinite loop if the random
16281 number generation fails.
16285 * New 'rand' application for creating pseudo-random output.
16289 * Added configuration support for Linux/IA64
16291 *Rolf Haberrecker <rolf@suse.de>*
16293 * Assembler module support for Mingw32.
16297 * Shared library support for HPUX (in shlib/).
16299 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16301 * Shared library support for Solaris gcc.
16303 *Lutz Behnke <behnke@trustcenter.de>*
16305 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16307 * PKCS7_encrypt() was adding text MIME headers twice because they
16308 were added manually and by SMIME_crlf_copy().
16312 * In bntest.c don't call BN_rand with zero bits argument.
16314 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16316 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16317 case was implemented. This caused BN_div_recp() to fail occasionally.
16321 * Add an optional second argument to the set_label() in the perl
16322 assembly language builder. If this argument exists and is set
16323 to 1 it signals that the assembler should use a symbol whose
16324 scope is the entire file, not just the current function. This
16325 is needed with MASM which uses the format label:: for this scope.
16327 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16329 * Change the ASN1 types so they are typedefs by default. Before
16330 almost all types were #define'd to ASN1_STRING which was causing
16331 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16336 * Change names of new functions to the new get1/get0 naming
16337 convention: After 'get1', the caller owns a reference count
16338 and has to call `..._free`; 'get0' returns a pointer to some
16339 data structure without incrementing reference counters.
16340 (Some of the existing 'get' functions increment a reference
16341 counter, some don't.)
16342 Similarly, 'set1' and 'add1' functions increase reference
16343 counters or duplicate objects.
16347 * Allow for the possibility of temp RSA key generation failure:
16348 the code used to assume it always worked and crashed on failure.
16352 * Fix potential buffer overrun problem in BIO_printf().
16353 *Ulf Möller, using public domain code by Patrick Powell; problem
16354 pointed out by David Sacerdote <das33@cornell.edu>*
16356 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16357 RAND_egd() and RAND_status(). In the command line application,
16358 the EGD socket can be specified like a seed file using RANDFILE
16363 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16364 Some CAs (e.g. Verisign) distribute certificates in this form.
16368 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16369 list to exclude them. This means that no special compilation option
16370 is needed to use anonymous DH: it just needs to be included in the
16375 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16376 EVP_MD_type. The old functionality is available in a new macro called
16377 EVP_MD_md(). Change code that uses it and update docs.
16381 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16382 where the `void *` argument is replaced by a function pointer argument.
16383 Previously `void *` was abused to point to functions, which works on
16384 many platforms, but is not correct. As these functions are usually
16385 called by macros defined in OpenSSL header files, most source code
16386 should work without changes.
16390 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16391 sections with information on -D... compiler switches used for
16392 compiling the library so that applications can see them. To enable
16393 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16394 must be defined. E.g.,
16395 #define OPENSSL_ALGORITHM_DEFINES
16396 #include <openssl/opensslconf.h>
16397 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16399 *Richard Levitte, Ulf and Bodo Möller*
16401 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16406 * Change the 'other' type in certificate aux info to a STACK_OF
16407 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16408 the required ASN1 format: arbitrary types determined by an OID.
16412 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16413 argument to 'req'. This is not because the function is newer or
16414 better than others it just uses the work 'NEW' in the certificate
16415 request header lines. Some software needs this.
16419 * Reorganise password command line arguments: now passwords can be
16420 obtained from various sources. Delete the PEM_cb function and make
16421 it the default behaviour: i.e. if the callback is NULL and the
16422 usrdata argument is not NULL interpret it as a null terminated pass
16423 phrase. If usrdata and the callback are NULL then the pass phrase
16424 is prompted for as usual.
16428 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16429 the support is automatically enabled. The resulting binaries will
16430 autodetect the card and use it if present.
16432 *Ben Laurie and Compaq Inc.*
16434 * Work around for Netscape hang bug. This sends certificate request
16435 and server done in one record. Since this is perfectly legal in the
16436 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16437 the bugs/SSLv3 entry for more info.
16441 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16445 * Add -rand argument to smime and pkcs12 applications and read/write
16450 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16454 * Add command line password options to the remaining applications.
16458 * Bug fix for BN_div_recp() for numerators with an even number of
16463 * More tests in bntest.c, and changed test_bn output.
16467 * ./config recognizes MacOS X now.
16471 * Bug fix for BN_div() when the first words of num and divisor are
16472 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16476 * Add support for various broken PKCS#8 formats, and command line
16477 options to produce them.
16481 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16482 get temporary BIGNUMs from a BN_CTX.
16486 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16491 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16492 include a #define from the old name to the new. The original intent
16493 was that statically linked binaries could for example just call
16494 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16495 link with digests. This never worked because SSLeay_add_all_digests()
16496 and SSLeay_add_all_ciphers() were in the same source file so calling
16497 one would link with the other. They are now in separate source files.
16501 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16505 * Use a less unusual form of the Miller-Rabin primality test (it used
16506 a binary algorithm for exponentiation integrated into the Miller-Rabin
16507 loop, our standard modexp algorithms are faster).
16511 * Support for the EBCDIC character set completed.
16513 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16515 * Source code cleanups: use const where appropriate, eliminate casts,
16516 use `void *` instead of `char *` in lhash.
16520 * Bugfix: ssl3_send_server_key_exchange was not restartable
16521 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16522 this the server could overwrite ephemeral keys that the client
16527 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16528 using 50 iterations of the Rabin-Miller test.
16530 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16531 iterations of the Rabin-Miller test as required by the appendix
16532 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16533 As BN_is_prime_fasttest includes trial division, DSA parameter
16534 generation becomes much faster.
16536 This implies a change for the callback functions in DSA_is_prime
16537 and DSA_generate_parameters: The callback function is called once
16538 for each positive witness in the Rabin-Miller test, not just
16539 occasionally in the inner loop; and the parameters to the
16540 callback function now provide an iteration count for the outer
16541 loop rather than for the current invocation of the inner loop.
16542 DSA_generate_parameters additionally can call the callback
16543 function with an 'iteration count' of -1, meaning that a
16544 candidate has passed the trial division test (when q is generated
16545 from an application-provided seed, trial division is skipped).
16549 * New function BN_is_prime_fasttest that optionally does trial
16550 division before starting the Rabin-Miller test and has
16551 an additional BN_CTX * argument (whereas BN_is_prime always
16552 has to allocate at least one BN_CTX).
16553 'callback(1, -1, cb_arg)' is called when a number has passed the
16554 trial division stage.
16558 * Fix for bug in CRL encoding. The validity dates weren't being handled
16563 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16567 * New function BN_pseudo_rand().
16571 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16572 bignum version of BN_from_montgomery() with the working code from
16573 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16578 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16579 made it impossible to use the same SSL_SESSION data structure in
16580 SSL2 clients in multiple threads.
16584 * The return value of RAND_load_file() no longer counts bytes obtained
16585 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16586 to seed the PRNG (previously an explicit byte count was required).
16588 *Ulf Möller, Bodo Möller*
16590 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16591 used `char *` instead of `void *` and had casts all over the place.
16595 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16599 * Retain source code compatibility for BN_prime_checks macro:
16600 BN_is_prime(..., BN_prime_checks, ...) now uses
16601 BN_prime_checks_for_size to determine the appropriate number of
16602 Rabin-Miller iterations.
16606 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16607 DH_CHECK_P_NOT_SAFE_PRIME.
16608 (Check if this is true? OpenPGP calls them "strong".)
16612 * Merge the functionality of "dh" and "gendh" programs into a new program
16613 "dhparam". The old programs are retained for now but will handle DH keys
16614 (instead of parameters) in future.
16618 * Make the ciphers, s_server and s_client programs check the return values
16619 when a new cipher list is set.
16623 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16624 ciphers. Before when the 56bit ciphers were enabled the sorting was
16627 The syntax for the cipher sorting has been extended to support sorting by
16628 cipher-strength (using the strength_bits hard coded in the tables).
16629 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16631 Fix a bug in the cipher-command parser: when supplying a cipher command
16632 string with an "undefined" symbol (neither command nor alphanumeric
16633 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16634 an error is flagged.
16636 Due to the strength-sorting extension, the code of the
16637 ssl_create_cipher_list() function was completely rearranged. I hope that
16638 the readability was also increased :-)
16640 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16642 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16643 for the first serial number and places 2 in the serial number file. This
16644 avoids problems when the root CA is created with serial number zero and
16645 the first user certificate has the same issuer name and serial number
16650 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16651 the new code. Add documentation for this stuff.
16655 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16656 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16657 structures and behave in an analogous way to the X509v3 functions:
16658 they shouldn't be called directly but wrapper functions should be used
16661 So we also now have some wrapper functions that call the X509at functions
16662 when passed certificate requests. (TO DO: similar things can be done with
16663 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16664 things. Some of these need some d2i or i2d and print functionality
16665 because they handle more complex structures.)
16669 * Add missing #ifndefs that caused missing symbols when building libssl
16670 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16671 NO_RSA in `ssl/s2*.c`.
16673 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16675 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16676 has a return value which indicates the quality of the random data
16677 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16678 error queue. New function RAND_pseudo_bytes() generates output that is
16679 guaranteed to be unique but not unpredictable. RAND_add is like
16680 RAND_seed, but takes an extra argument for an entropy estimate
16681 (RAND_seed always assumes full entropy).
16685 * Do more iterations of Rabin-Miller probable prime test (specifically,
16686 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16687 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16688 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16689 false-positive rate of at most 2^-80 for random input.
16693 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16697 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16698 in the 0.9.5 release), this returns the chain
16699 from an X509_CTX structure with a dup of the stack and all
16700 the X509 reference counts upped: so the stack will exist
16701 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16704 Also make SSL_SESSION_print() print out the verify return
16709 * Add manpage for the pkcs12 command. Also change the default
16710 behaviour so MAC iteration counts are used unless the new
16711 -nomaciter option is used. This improves file security and
16712 only older versions of MSIE (4.0 for example) need it.
16716 * Honor the no-xxx Configure options when creating .DEF files.
16720 * Add PKCS#10 attributes to field table: challengePassword,
16721 unstructuredName and unstructuredAddress. These are taken from
16722 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16723 international characters are used.
16725 More changes to X509_ATTRIBUTE code: allow the setting of types
16726 based on strings. Remove the 'loc' parameter when adding
16727 attributes because these will be a SET OF encoding which is sorted
16732 * Initial changes to the 'req' utility to allow request generation
16733 automation. This will allow an application to just generate a template
16734 file containing all the field values and have req construct the
16737 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16738 used all over the place including certificate requests and PKCS#7
16739 structures. They are currently handled manually where necessary with
16740 some primitive wrappers for PKCS#7. The new functions behave in a
16741 manner analogous to the X509 extension functions: they allow
16742 attributes to be looked up by NID and added.
16744 Later something similar to the X509V3 code would be desirable to
16745 automatically handle the encoding, decoding and printing of the
16746 more complex types. The string types like challengePassword can
16747 be handled by the string table functions.
16749 Also modified the multi byte string table handling. Now there is
16750 a 'global mask' which masks out certain types. The table itself
16751 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16752 is useful when for example there is only one permissible type
16753 (as in countryName) and using the mask might result in no valid
16758 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16759 SSL_get_peer_finished to allow applications to obtain the latest
16760 Finished messages sent to the peer or expected from the peer,
16761 respectively. (SSL_get_peer_finished is usually the Finished message
16762 actually received from the peer, otherwise the protocol will be aborted.)
16764 As the Finished message are message digests of the complete handshake
16765 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16766 be used for external authentication procedures when the authentication
16767 provided by SSL/TLS is not desired or is not enough.
16771 * Enhanced support for Alpha Linux is added. Now ./config checks if
16772 the host supports BWX extension and if Compaq C is present on the
16773 $PATH. Just exploiting of the BWX extension results in 20-30%
16774 performance kick for some algorithms, e.g. DES and RC4 to mention
16775 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16780 * Add support for MS "fast SGC". This is arguably a violation of the
16781 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16782 weak crypto and after checking the certificate is SGC a second one
16783 with strong crypto. MS SGC stops the first handshake after receiving
16784 the server certificate message and sends a second client hello. Since
16785 a server will typically do all the time consuming operations before
16786 expecting any further messages from the client (server key exchange
16787 is the most expensive) there is little difference between the two.
16789 To get OpenSSL to support MS SGC we have to permit a second client
16790 hello message after we have sent server done. In addition we have to
16791 reset the MAC if we do get this second client hello.
16795 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16796 if a DER encoded private key is RSA or DSA traditional format. Changed
16797 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16798 format DER encoded private key. Newer code should use PKCS#8 format which
16799 has the key type encoded in the ASN1 structure. Added DER private key
16800 support to pkcs8 application.
16804 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16805 ciphersuites has been selected (as required by the SSL 3/TLS 1
16806 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16807 is set, we interpret this as a request to violate the specification
16808 (the worst that can happen is a handshake failure, and 'correct'
16809 behaviour would result in a handshake failure anyway).
16813 * In SSL_CTX_add_session, take into account that there might be multiple
16814 SSL_SESSION structures with the same session ID (e.g. when two threads
16815 concurrently obtain them from an external cache).
16816 The internal cache can handle only one SSL_SESSION with a given ID,
16817 so if there's a conflict, we now throw out the old one to achieve
16822 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16823 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16824 some routines that use cipher OIDs: some ciphers do not have OIDs
16825 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16830 * Simplify the trust setting structure and code. Now we just have
16831 two sequences of OIDs for trusted and rejected settings. These will
16832 typically have values the same as the extended key usage extension
16833 and any application specific purposes.
16835 The trust checking code now has a default behaviour: it will just
16836 check for an object with the same NID as the passed id. Functions can
16837 be provided to override either the default behaviour or the behaviour
16838 for a given id. SSL client, server and email already have functions
16839 in place for compatibility: they check the NID and also return "trusted"
16840 if the certificate is self signed.
16844 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16845 traditional format into an EVP_PKEY structure.
16849 * Add a password callback function PEM_cb() which either prompts for
16850 a password if usr_data is NULL or otherwise assumes it is a null
16851 terminated password. Allow passwords to be passed on command line
16852 environment or config files in a few more utilities.
16856 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16857 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16858 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16859 Update documentation.
16863 * Support for ASN1 "NULL" type. This could be handled before by using
16864 ASN1_TYPE but there wasn't any function that would try to read a NULL
16865 and produce an error if it couldn't. For compatibility we also have
16866 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16867 don't allocate anything because they don't need to.
16871 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16874 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16876 * Rebuild of the memory allocation routines used by OpenSSL code and
16877 possibly others as well. The purpose is to make an interface that
16878 provide hooks so anyone can build a separate set of allocation and
16879 deallocation routines to be used by OpenSSL, for example memory
16880 pool implementations, or something else, which was previously hard
16881 since Malloc(), Realloc() and Free() were defined as macros having
16882 the values malloc, realloc and free, respectively (except for Win32
16883 compilations). The same is provided for memory debugging code.
16884 OpenSSL already comes with functionality to find memory leaks, but
16885 this gives people a chance to debug other memory problems.
16887 With these changes, a new set of functions and macros have appeared:
16889 CRYPTO_set_mem_debug_functions() [F]
16890 CRYPTO_get_mem_debug_functions() [F]
16891 CRYPTO_dbg_set_options() [F]
16892 CRYPTO_dbg_get_options() [F]
16893 CRYPTO_malloc_debug_init() [M]
16895 The memory debug functions are NULL by default, unless the library
16896 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16897 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16898 gives the standard debugging functions that come with OpenSSL) or
16899 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16900 provided by the library user) must be used. When the standard
16901 debugging functions are used, CRYPTO_dbg_set_options can be used to
16902 request additional information:
16903 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16904 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16906 Also, things like CRYPTO_set_mem_functions will always give the
16907 expected result (the new set of functions is used for allocation
16908 and deallocation) at all times, regardless of platform and compiler
16911 To finish it up, some functions that were never use in any other
16912 way than through macros have a new API and new semantic:
16914 CRYPTO_dbg_malloc()
16915 CRYPTO_dbg_realloc()
16918 All macros of value have retained their old syntax.
16920 *Richard Levitte and Bodo Moeller*
16922 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16923 ordering of SMIMECapabilities wasn't in "strength order" and there
16924 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16929 * Some ASN1 types with illegal zero length encoding (INTEGER,
16930 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16932 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16934 * Merge in my S/MIME library for OpenSSL. This provides a simple
16935 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16936 functionality to handle multipart/signed properly) and a utility
16937 called 'smime' to call all this stuff. This is based on code I
16938 originally wrote for Celo who have kindly allowed it to be
16939 included in OpenSSL.
16943 * Add variants des_set_key_checked and des_set_key_unchecked of
16944 des_set_key (aka des_key_sched). Global variable des_check_key
16945 decides which of these is called by des_set_key; this way
16946 des_check_key behaves as it always did, but applications and
16947 the library itself, which was buggy for des_check_key == 1,
16948 have a cleaner way to pick the version they need.
16952 * New function PKCS12_newpass() which changes the password of a
16957 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16958 dynamic mix. In both cases the ids can be used as an index into the
16959 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16960 functions so they accept a list of the field values and the
16961 application doesn't need to directly manipulate the X509_TRUST
16966 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16971 * Modify the way the V3 extension code looks up extensions. This now
16972 works in a similar way to the object code: we have some "standard"
16973 extensions in a static table which is searched with OBJ_bsearch()
16974 and the application can add dynamic ones if needed. The file
16975 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16976 updated whenever a new extension is added to the core code and kept
16977 in ext_nid order. There is a simple program 'tabtest.c' which checks
16978 this. New extensions are not added too often so this file can readily
16979 be maintained manually.
16981 There are two big advantages in doing things this way. The extensions
16982 can be looked up immediately and no longer need to be "added" using
16983 X509V3_add_standard_extensions(): this function now does nothing.
16984 Side note: I get *lots* of email saying the extension code doesn't
16985 work because people forget to call this function.
16986 Also no dynamic allocation is done unless new extensions are added:
16987 so if we don't add custom extensions there is no need to call
16988 X509V3_EXT_cleanup().
16992 * Modify enc utility's salting as follows: make salting the default. Add a
16993 magic header, so unsalted files fail gracefully instead of just decrypting
16994 to garbage. This is because not salting is a big security hole, so people
16995 should be discouraged from doing it.
16999 * Fixes and enhancements to the 'x509' utility. It allowed a message
17000 digest to be passed on the command line but it only used this
17001 parameter when signing a certificate. Modified so all relevant
17002 operations are affected by the digest parameter including the
17003 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17004 DSA key was used because it didn't fix the digest.
17008 * Initial certificate chain verify code. Currently tests the untrusted
17009 certificates for consistency with the verify purpose (which is set
17010 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17012 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17013 this is because it will reject chains with invalid extensions whereas
17014 every previous version of OpenSSL and SSLeay made no checks at all.
17016 Trust code: checks the root CA for the relevant trust settings. Trust
17017 settings have an initial value consistent with the verify purpose: e.g.
17018 if the verify purpose is for SSL client use it expects the CA to be
17019 trusted for SSL client use. However the default value can be changed to
17020 permit custom trust settings: one example of this would be to only trust
17021 certificates from a specific "secure" set of CAs.
17023 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17024 which should be used for version portability: especially since the
17025 verify structure is likely to change more often now.
17027 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17028 to set them. If not set then assume SSL clients will verify SSL servers
17031 Two new options to the verify program: -untrusted allows a set of
17032 untrusted certificates to be passed in and -purpose which sets the
17033 intended purpose of the certificate. If a purpose is set then the
17034 new chain verify code is used to check extension consistency.
17038 * Support for the authority information access extension.
17042 * Modify RSA and DSA PEM read routines to transparently handle
17043 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17044 public keys in a format compatible with certificate
17045 SubjectPublicKeyInfo structures. Unfortunately there were already
17046 functions called *_PublicKey_* which used various odd formats so
17047 these are retained for compatibility: however the DSA variants were
17048 never in a public release so they have been deleted. Changed dsa/rsa
17049 utilities to handle the new format: note no releases ever handled public
17050 keys so we should be OK.
17052 The primary motivation for this change is to avoid the same fiasco
17053 that dogs private keys: there are several incompatible private key
17054 formats some of which are standard and some OpenSSL specific and
17055 require various evil hacks to allow partial transparent handling and
17056 even then it doesn't work with DER formats. Given the option anything
17057 other than PKCS#8 should be dumped: but the other formats have to
17058 stay in the name of compatibility.
17060 With public keys and the benefit of hindsight one standard format
17061 is used which works with EVP_PKEY, RSA or DSA structures: though
17062 it clearly returns an error if you try to read the wrong kind of key.
17064 Added a -pubkey option to the 'x509' utility to output the public key.
17065 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17066 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17067 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17068 that do the same as the `EVP_PKEY_assign_*()` except they up the
17069 reference count of the added key (they don't "swallow" the
17074 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17075 CRLs would fail if the file contained no certificates or no CRLs:
17076 added a new function to read in both types and return the number
17077 read: this means that if none are read it will be an error. The
17078 DER versions of the certificate and CRL reader would always fail
17079 because it isn't possible to mix certificates and CRLs in DER format
17080 without choking one or the other routine. Changed this to just read
17081 a certificate: this is the best we can do. Also modified the code
17082 in `apps/verify.c` to take notice of return codes: it was previously
17083 attempting to read in certificates from NULL pointers and ignoring
17084 any errors: this is one reason why the cert and CRL reader seemed
17085 to work. It doesn't check return codes from the default certificate
17086 routines: these may well fail if the certificates aren't installed.
17090 * Code to support otherName option in GeneralName.
17094 * First update to verify code. Change the verify utility
17095 so it warns if it is passed a self signed certificate:
17096 for consistency with the normal behaviour. X509_verify
17097 has been modified to it will now verify a self signed
17098 certificate if *exactly* the same certificate appears
17099 in the store: it was previously impossible to trust a
17100 single self signed certificate. This means that:
17101 openssl verify ss.pem
17102 now gives a warning about a self signed certificate but
17103 openssl verify -CAfile ss.pem ss.pem
17108 * For servers, store verify_result in SSL_SESSION data structure
17109 (and add it to external session representation).
17110 This is needed when client certificate verifications fails,
17111 but an application-provided verification callback (set by
17112 SSL_CTX_set_cert_verify_callback) allows accepting the session
17113 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17114 but returns 1): When the session is reused, we have to set
17115 ssl->verify_result to the appropriate error code to avoid
17118 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17120 * Fix a bug in the new PKCS#7 code: it didn't consider the
17121 case in PKCS7_dataInit() where the signed PKCS7 structure
17122 didn't contain any existing data because it was being created.
17124 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17126 * Add a salt to the key derivation routines in enc.c. This
17127 forms the first 8 bytes of the encrypted file. Also add a
17128 -S option to allow a salt to be input on the command line.
17132 * New function X509_cmp(). Oddly enough there wasn't a function
17133 to compare two certificates. We do this by working out the SHA1
17134 hash and comparing that. X509_cmp() will be needed by the trust
17139 * SSL_get1_session() is like SSL_get_session(), but increments
17140 the reference count in the SSL_SESSION returned.
17142 *Geoff Thorpe <geoff@eu.c2.net>*
17144 * Fix for 'req': it was adding a null to request attributes.
17145 Also change the X509_LOOKUP and X509_INFO code to handle
17146 certificate auxiliary information.
17150 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17155 * Add the possibility to add extra information to the memory leak
17156 detecting output, to form tracebacks, showing from where each
17157 allocation was originated: CRYPTO_push_info("constant string") adds
17158 the string plus current file name and line number to a per-thread
17159 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17160 is like calling CYRPTO_pop_info() until the stack is empty.
17161 Also updated memory leak detection code to be multi-thread-safe.
17165 * Add options -text and -noout to pkcs7 utility and delete the
17166 encryption options which never did anything. Update docs.
17170 * Add options to some of the utilities to allow the pass phrase
17171 to be included on either the command line (not recommended on
17172 OSes like Unix) or read from the environment. Update the
17173 manpages and fix a few bugs.
17177 * Add a few manpages for some of the openssl commands.
17181 * Fix the -revoke option in ca. It was freeing up memory twice,
17182 leaking and not finding already revoked certificates.
17186 * Extensive changes to support certificate auxiliary information.
17187 This involves the use of X509_CERT_AUX structure and X509_AUX
17188 functions. An X509_AUX function such as PEM_read_X509_AUX()
17189 can still read in a certificate file in the usual way but it
17190 will also read in any additional "auxiliary information". By
17191 doing things this way a fair degree of compatibility can be
17192 retained: existing certificates can have this information added
17193 using the new 'x509' options.
17195 Current auxiliary information includes an "alias" and some trust
17196 settings. The trust settings will ultimately be used in enhanced
17197 certificate chain verification routines: currently a certificate
17198 can only be trusted if it is self signed and then it is trusted
17203 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17204 The problem was that one of the replacement routines had not been working
17205 since SSLeay releases. For now the offending routine has been replaced
17206 with non-optimised assembler. Even so, this now gives around 95%
17207 performance improvement for 1024 bit RSA signs.
17211 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17212 handling. Most clients have the effective key size in bits equal to
17213 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17214 A few however don't do this and instead use the size of the decrypted key
17215 to determine the RC2 key length and the AlgorithmIdentifier to determine
17216 the effective key length. In this case the effective key length can still
17217 be 40 bits but the key length can be 168 bits for example. This is fixed
17218 by manually forcing an RC2 key into the EVP_PKEY structure because the
17219 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17220 the key length and effective key length are equal.
17224 * Add a bunch of functions that should simplify the creation of
17225 X509_NAME structures. Now you should be able to do:
17226 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17227 and have it automatically work out the correct field type and fill in
17228 the structures. The more adventurous can try:
17229 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17230 and it will (hopefully) work out the correct multibyte encoding.
17234 * Change the 'req' utility to use the new field handling and multibyte
17235 copy routines. Before the DN field creation was handled in an ad hoc
17236 way in req, ca, and x509 which was rather broken and didn't support
17237 BMPStrings or UTF8Strings. Since some software doesn't implement
17238 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17239 using the dirstring_type option. See the new comment in the default
17240 openssl.cnf for more info.
17244 * Make crypto/rand/md_rand.c more robust:
17245 - Assure unique random numbers after fork().
17246 - Make sure that concurrent threads access the global counter and
17247 md serializably so that we never lose entropy in them
17248 or use exactly the same state in multiple threads.
17249 Access to the large state is not always serializable because
17250 the additional locking could be a performance killer, and
17251 md should be large enough anyway.
17255 * New file `apps/app_rand.c` with commonly needed functionality
17256 for handling the random seed file.
17258 Use the random seed file in some applications that previously did not:
17260 dsaparam -genkey (which also ignored its '-rand' option),
17263 x509 (when signing).
17264 Except on systems with /dev/urandom, it is crucial to have a random
17265 seed file at least for key creation, DSA signing, and for DH exchanges;
17266 for RSA signatures we could do without one.
17268 gendh and gendsa (unlike genrsa) used to read only the first byte
17269 of each file listed in the '-rand' option. The function as previously
17270 found in genrsa is now in app_rand.c and is used by all programs
17271 that support '-rand'.
17275 * In RAND_write_file, use mode 0600 for creating files;
17276 don't just chmod when it may be too late.
17280 * Report an error from X509_STORE_load_locations
17281 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17285 * New function ASN1_mbstring_copy() this copies a string in either
17286 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17287 into an ASN1_STRING type. A mask of permissible types is passed
17288 and it chooses the "minimal" type to use or an error if not type
17293 * Add function equivalents to the various macros in asn1.h. The old
17294 macros are retained with an `M_` prefix. Code inside the library can
17295 use the `M_` macros. External code (including the openssl utility)
17296 should *NOT* in order to be "shared library friendly".
17300 * Add various functions that can check a certificate's extensions
17301 to see if it usable for various purposes such as SSL client,
17302 server or S/MIME and CAs of these types. This is currently
17303 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17304 verification. Also added a -purpose flag to x509 utility to
17305 print out all the purposes.
17309 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17314 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17315 for, obtain and decode and extension and obtain its critical flag.
17316 This allows all the necessary extension code to be handled in a
17317 single function call.
17321 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17322 platforms. See crypto/rc4/rc4_enc.c for further details.
17326 * New -noout option to asn1parse. This causes no output to be produced
17327 its main use is when combined with -strparse and -out to extract data
17328 from a file (which may not be in ASN.1 format).
17332 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17333 when producing the local key id.
17335 *Richard Levitte <levitte@stacken.kth.se>*
17337 * New option -dhparam in s_server. This allows a DH parameter file to be
17338 stated explicitly. If it is not stated then it tries the first server
17339 certificate file. The previous behaviour hard coded the filename
17344 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17345 a public key to be input or output. For example:
17346 openssl rsa -in key.pem -pubout -out pubkey.pem
17347 Also added necessary DSA public key functions to handle this.
17351 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17352 in the message. This was handled by allowing
17353 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17355 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17357 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17358 to the end of the strings whereas this didn't. This would cause problems
17359 if strings read with d2i_ASN1_bytes() were later modified.
17361 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17363 * Fix for base64 decode bug. When a base64 bio reads only one line of
17364 data and it contains EOF it will end up returning an error. This is
17365 caused by input 46 bytes long. The cause is due to the way base64
17366 BIOs find the start of base64 encoded data. They do this by trying a
17367 trial decode on each line until they find one that works. When they
17368 do a flag is set and it starts again knowing it can pass all the
17369 data directly through the decoder. Unfortunately it doesn't reset
17370 the context it uses. This means that if EOF is reached an attempt
17371 is made to pass two EOFs through the context and this causes the
17372 resulting error. This can also cause other problems as well. As is
17373 usual with these problems it takes *ages* to find and the fix is
17374 trivial: move one line.
17376 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17378 * Ugly workaround to get s_client and s_server working under Windows. The
17379 old code wouldn't work because it needed to select() on sockets and the
17380 tty (for keypresses and to see if data could be written). Win32 only
17381 supports select() on sockets so we select() with a 1s timeout on the
17382 sockets and then see if any characters are waiting to be read, if none
17383 are present then we retry, we also assume we can always write data to
17384 the tty. This isn't nice because the code then blocks until we've
17385 received a complete line of data and it is effectively polling the
17386 keyboard at 1s intervals: however it's quite a bit better than not
17387 working at all :-) A dedicated Windows application might handle this
17388 with an event loop for example.
17392 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17393 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17394 will be called when RSA_sign() and RSA_verify() are used. This is useful
17395 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17396 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17397 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17398 This necessitated the support of an extra signature type NID_md5_sha1
17399 for SSL signatures and modifications to the SSL library to use it instead
17400 of calling RSA_public_decrypt() and RSA_private_encrypt().
17404 * Add new -verify -CAfile and -CApath options to the crl program, these
17405 will lookup a CRL issuers certificate and verify the signature in a
17406 similar way to the verify program. Tidy up the crl program so it
17407 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17408 less strict. It will now permit CRL extensions even if it is not
17409 a V2 CRL: this will allow it to tolerate some broken CRLs.
17413 * Initialize all non-automatic variables each time one of the openssl
17414 sub-programs is started (this is necessary as they may be started
17415 multiple times from the "OpenSSL>" prompt).
17417 *Lennart Bang, Bodo Moeller*
17419 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17420 removing all other RSA functionality (this is what NO_RSA does). This
17421 is so (for example) those in the US can disable those operations covered
17422 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17427 * Non-copying interface to BIO pairs.
17428 (still largely untested)
17432 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17433 ASCII string. This was handled independently in various places before.
17437 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17438 UTF8 strings a character at a time.
17442 * Use client_version from client hello to select the protocol
17443 (s23_srvr.c) and for RSA client key exchange verification
17444 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17448 * Add various utility functions to handle SPKACs, these were previously
17449 handled by poking round in the structure internals. Added new function
17450 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17451 print, verify and generate SPKACs. Based on an original idea from
17452 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17456 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17460 * Allow the config file extension section to be overwritten on the
17461 command line. Based on an original idea from Massimiliano Pala
17462 <madwolf@comune.modena.it>. The new option is called -extensions
17463 and can be applied to ca, req and x509. Also -reqexts to override
17464 the request extensions in req and -crlexts to override the crl extensions
17469 * Add new feature to the SPKAC handling in ca. Now you can include
17470 the same field multiple times by preceding it by "XXXX." for example:
17473 this is the same syntax as used in the req config file.
17477 * Allow certificate extensions to be added to certificate requests. These
17478 are specified in a 'req_extensions' option of the req section of the
17479 config file. They can be printed out with the -text option to req but
17480 are otherwise ignored at present.
17484 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17485 data read consists of only the final block it would not decrypted because
17486 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17487 A misplaced 'break' also meant the decrypted final block might not be
17488 copied until the next read.
17492 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17493 a few extra parameters to the DH structure: these will be useful if
17494 for example we want the value of 'q' or implement X9.42 DH.
17498 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17499 provides hooks that allow the default DSA functions or functions on a
17500 "per key" basis to be replaced. This allows hardware acceleration and
17501 hardware key storage to be handled without major modification to the
17502 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17503 associated functions.
17507 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17508 as "read only": it can't be written to and the buffer it points to will
17509 not be freed. Reading from a read only BIO is much more efficient than
17510 a normal memory BIO. This was added because there are several times when
17511 an area of memory needs to be read from a BIO. The previous method was
17512 to create a memory BIO and write the data to it, this results in two
17513 copies of the data and an O(n^2) reading algorithm. There is a new
17514 function BIO_new_mem_buf() which creates a read only memory BIO from
17515 an area of memory. Also modified the PKCS#7 routines to use read only
17520 * Bugfix: ssl23_get_client_hello did not work properly when called in
17521 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17522 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17523 but a retry condition occurred while trying to read the rest.
17527 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17528 NID_pkcs7_encrypted by default: this was wrong since this should almost
17529 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17530 the encrypted data type: this is a more sensible place to put it and it
17531 allows the PKCS#12 code to be tidied up that duplicated this
17536 * Changed obj_dat.pl script so it takes its input and output files on
17537 the command line. This should avoid shell escape redirection problems
17542 * Initial support for certificate extension requests, these are included
17543 in things like Xenroll certificate requests. Included functions to allow
17544 extensions to be obtained and added.
17548 * -crlf option to s_client and s_server for sending newlines as
17549 CRLF (as required by many protocols).
17553 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17555 * Install libRSAglue.a when OpenSSL is built with RSAref.
17557 *Ralf S. Engelschall*
17559 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17561 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17563 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17568 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17569 DH parameters/keys (q is lost during that conversion, but the resulting
17570 DH parameters contain its length).
17572 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17573 much faster than DH_generate_parameters (which creates parameters
17574 where `p = 2*q + 1`), and also the smaller q makes DH computations
17575 much more efficient (160-bit exponentiation instead of 1024-bit
17576 exponentiation); so this provides a convenient way to support DHE
17577 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17578 utter importance to use
17579 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17581 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17582 when such DH parameters are used, because otherwise small subgroup
17583 attacks may become possible!
17587 * Avoid memory leak in i2d_DHparams.
17591 * Allow the -k option to be used more than once in the enc program:
17592 this allows the same encrypted message to be read by multiple recipients.
17596 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17597 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17598 it will always use the numerical form of the OID, even if it has a short
17603 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17604 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17605 otherwise bn_mod_exp was called. In the case of hardware keys for example
17606 no private key components need be present and it might store extra data
17607 in the RSA structure, which cannot be accessed from bn_mod_exp.
17608 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17609 private key operations.
17613 * Added support for SPARC Linux.
17617 * pem_password_cb function type incompatibly changed from
17618 typedef int pem_password_cb(char *buf, int size, int rwflag);
17620 ....(char *buf, int size, int rwflag, void *userdata);
17621 so that applications can pass data to their callbacks:
17622 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17623 additional void * argument, which is just handed through whenever
17624 the password callback is called.
17626 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17628 New function SSL_CTX_set_default_passwd_cb_userdata.
17630 Compatibility note: As many C implementations push function arguments
17631 onto the stack in reverse order, the new library version is likely to
17632 interoperate with programs that have been compiled with the old
17633 pem_password_cb definition (PEM_whatever takes some data that
17634 happens to be on the stack as its last argument, and the callback
17635 just ignores this garbage); but there is no guarantee whatsoever that
17638 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17639 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17640 problems not only on Windows, but also on some Unix platforms.
17641 To avoid problematic command lines, these definitions are now in an
17642 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17643 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17647 * MIPS III/IV assembler module is reimplemented.
17651 * More DES library cleanups: remove references to srand/rand and
17652 delete an unused file.
17656 * Add support for the free Netwide assembler (NASM) under Win32,
17657 since not many people have MASM (ml) and it can be hard to obtain.
17658 This is currently experimental but it seems to work OK and pass all
17659 the tests. Check out INSTALL.W32 for info.
17663 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17664 without temporary keys kept an extra copy of the server key,
17665 and connections with temporary keys did not free everything in case
17670 * New function RSA_check_key and new openssl rsa option -check
17671 for verifying the consistency of RSA keys.
17673 *Ulf Moeller, Bodo Moeller*
17675 * Various changes to make Win32 compile work:
17676 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17677 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17678 comparison" warnings.
17679 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17683 * Add a debugging option to PKCS#5 v2 key generation function: when
17684 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17685 derived keys are printed to stderr.
17689 * Copy the flags in ASN1_STRING_dup().
17691 *Roman E. Pavlov <pre@mo.msk.ru>*
17693 * The x509 application mishandled signing requests containing DSA
17694 keys when the signing key was also DSA and the parameters didn't match.
17696 It was supposed to omit the parameters when they matched the signing key:
17697 the verifying software was then supposed to automatically use the CA's
17698 parameters if they were absent from the end user certificate.
17700 Omitting parameters is no longer recommended. The test was also
17701 the wrong way round! This was probably due to unusual behaviour in
17702 EVP_cmp_parameters() which returns 1 if the parameters match.
17703 This meant that parameters were omitted when they *didn't* match and
17704 the certificate was useless. Certificates signed with 'ca' didn't have
17707 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17709 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17710 The interface is as follows:
17711 Applications can use
17712 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17713 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17714 "off" is now the default.
17715 The library internally uses
17716 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17717 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17718 to disable memory-checking temporarily.
17720 Some inconsistent states that previously were possible (and were
17721 even the default) are now avoided.
17723 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17724 with each memory chunk allocated; this is occasionally more helpful
17725 than just having a counter.
17727 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17729 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17734 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17735 which largely parallels "options", but is for changing API behaviour,
17736 whereas "options" are about protocol behaviour.
17737 Initial "mode" flags are:
17739 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17740 a single record has been written.
17741 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17742 retries use the same buffer location.
17743 (But all of the contents must be
17748 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17751 * Fix problems with no-hmac etc.
17753 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17755 * New functions RSA_get_default_method(), RSA_set_method() and
17756 RSA_get_method(). These allows replacement of RSA_METHODs without having
17757 to mess around with the internals of an RSA structure.
17761 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17762 Also really enable memory leak checks in openssl.c and in some
17765 *Chad C. Mulligan, Bodo Moeller*
17767 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17768 up the length of negative integers. This has now been simplified to just
17769 store the length when it is first determined and use it later, rather
17770 than trying to keep track of where data is copied and updating it to
17772 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17774 * Add a new function PKCS7_signatureVerify. This allows the verification
17775 of a PKCS#7 signature but with the signing certificate passed to the
17776 function itself. This contrasts with PKCS7_dataVerify which assumes the
17777 certificate is present in the PKCS#7 structure. This isn't always the
17778 case: certificates can be omitted from a PKCS#7 structure and be
17779 distributed by "out of band" means (such as a certificate database).
17783 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17784 function prototypes in pem.h, also change util/mkdef.pl to add the
17785 necessary function names.
17789 * mk1mf.pl (used by Windows builds) did not properly read the
17790 options set by Configure in the top level Makefile, and Configure
17791 was not even able to write more than one option correctly.
17792 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17796 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17797 file to be loaded from a BIO or FILE pointer. The BIO version will
17798 for example allow memory BIOs to contain config info.
17802 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17803 Whoever hopes to achieve shared-library compatibility across versions
17804 must use this, not the compile-time macro.
17805 (Exercise 0.9.4: Which is the minimum library version required by
17807 Note: All this applies only to multi-threaded programs, others don't
17812 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17813 through a BIO pair triggered the default case, i.e.
17814 SSLerr(...,SSL_R_UNKNOWN_STATE).
17818 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17819 can use the SSL library even if none of the specific BIOs is
17824 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17825 for the encoded length.
17827 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17829 * Add initial documentation of the X509V3 functions.
17833 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17834 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17835 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17836 secure PKCS#8 private key format with a high iteration count.
17840 * Fix determination of Perl interpreter: A perl or perl5
17841 *directory* in $PATH was also accepted as the interpreter.
17843 *Ralf S. Engelschall*
17845 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17846 wrong with it but it was very old and did things like calling
17847 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17848 unusual formatting.
17852 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17853 to use the new extension code.
17857 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17858 with macros. This should make it easier to change their form, add extra
17859 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17864 * Add to configuration table a new entry that can specify an alternative
17865 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17866 according to Mark Crispin <MRC@Panda.COM>.
17870 * DES CBC did not update the IV. Weird.
17874 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17875 Changing the behaviour of the former might break existing programs --
17876 where IV updating is needed, des_ncbc_encrypt can be used.
17879 * When bntest is run from "make test" it drives bc to check its
17880 calculations, as well as internally checking them. If an internal check
17881 fails, it needs to cause bc to give a non-zero result or make test carries
17882 on without noticing the failure. Fixed.
17886 * DES library cleanups.
17890 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17891 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17892 ciphers. NOTE: although the key derivation function has been verified
17893 against some published test vectors it has not been extensively tested
17894 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17899 * Instead of "mkdir -p", which is not fully portable, use new
17900 Perl script "util/mkdir-p.pl".
17904 * Rewrite the way password based encryption (PBE) is handled. It used to
17905 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17906 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17907 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17908 the 'parameter' field of the AlgorithmIdentifier is passed to the
17909 underlying key generation function so it must do its own ASN1 parsing.
17910 This has also changed the EVP_PBE_CipherInit() function which now has a
17911 'parameter' argument instead of literal salt and iteration count values
17912 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17916 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17917 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17918 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17919 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17920 value was just used as a "magic string" and not used directly its
17921 value doesn't matter.
17925 * Introduce some semblance of const correctness to BN. Shame C doesn't
17930 * "linux-sparc64" configuration (ultrapenguin).
17932 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17933 "linux-sparc" configuration.
17935 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17937 * config now generates no-xxx options for missing ciphers.
17941 * Support the EBCDIC character set (work in progress).
17942 File ebcdic.c not yet included because it has a different license.
17944 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17946 * Support BS2000/OSD-POSIX.
17948 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17950 * Make callbacks for key generation use `void *` instead of `char *`.
17954 * Make S/MIME samples compile (not yet tested).
17958 * Additional typesafe stacks.
17962 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17966 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17968 * New configuration variant "sco5-gcc".
17970 * Updated some demos.
17972 *Sean O Riordain, Wade Scholine*
17974 * Add missing BIO_free at exit of pkcs12 application.
17978 * Fix memory leak in conf.c.
17982 * Updates for Win32 to assembler version of MD5.
17986 * Set #! path to perl in `apps/der_chop` to where we found it
17987 instead of using a fixed path.
17991 * SHA library changes for irix64-mips4-cc.
17995 * Improvements for VMS support.
17999 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18001 * Bignum library bug fix. IRIX 6 passes "make test" now!
18002 This also avoids the problems with SC4.2 and unpatched SC5.
18004 *Andy Polyakov <appro@fy.chalmers.se>*
18006 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18007 These are required because of the typesafe stack would otherwise break
18008 existing code. If old code used a structure member which used to be STACK
18009 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18010 sk_num or sk_value it would produce an error because the num, data members
18011 are not present in STACK_OF. Now it just produces a warning. sk_set
18012 replaces the old method of assigning a value to sk_value
18013 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18014 that does this will no longer work (and should use sk_set instead) but
18015 this could be regarded as a "questionable" behaviour anyway.
18019 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18020 correctly handle encrypted S/MIME data.
18024 * Change type of various DES function arguments from des_cblock
18025 (which means, in function argument declarations, pointer to char)
18026 to des_cblock * (meaning pointer to array with 8 char elements),
18027 which allows the compiler to do more typechecking; it was like
18028 that back in SSLeay, but with lots of ugly casts.
18030 Introduce new type const_des_cblock.
18034 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18035 problems: find RecipientInfo structure that matches recipient certificate
18036 and initialise the ASN1 structures properly based on passed cipher.
18040 * Belatedly make the BN tests actually check the results.
18044 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18045 to and from BNs: it was completely broken. New compilation option
18046 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18047 key elements as negative integers.
18051 * Reorganize and speed up MD5.
18053 *Andy Polyakov <appro@fy.chalmers.se>*
18057 *Richard Levitte <richard@levitte.org>*
18059 * New option -out to asn1parse to allow the parsed structure to be
18060 output to a file. This is most useful when combined with the -strparse
18061 option to examine the output of things like OCTET STRINGS.
18065 * Make SSL library a little more fool-proof by not requiring any longer
18066 that `SSL_set_{accept,connect}_state` be called before
18067 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18068 in many applications because usually everything *appeared* to work as
18069 intended anyway -- now it really works as intended).
18073 * Move openssl.cnf out of lib/.
18077 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18078 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18079 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18081 *Ralf S. Engelschall*
18083 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18084 handle PKCS#7 enveloped data properly.
18086 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18088 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18089 copying pointers. The cert_st handling is changed by this in
18090 various ways (and thus what used to be known as ctx->default_cert
18091 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18092 any longer when s->cert does not give us what we need).
18093 ssl_cert_instantiate becomes obsolete by this change.
18094 As soon as we've got the new code right (possibly it already is?),
18095 we have solved a couple of bugs of the earlier code where s->cert
18096 was used as if it could not have been shared with other SSL structures.
18098 Note that using the SSL API in certain dirty ways now will result
18099 in different behaviour than observed with earlier library versions:
18100 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18101 does not influence s as it used to.
18103 In order to clean up things more thoroughly, inside SSL_SESSION
18104 we don't use CERT any longer, but a new structure SESS_CERT
18105 that holds per-session data (if available); currently, this is
18106 the peer's certificate chain and, for clients, the server's certificate
18107 and temporary key. CERT holds only those values that can have
18108 meaningful defaults in an SSL_CTX.
18112 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18113 from the internal representation. Various PKCS#7 fixes: remove some
18114 evil casts and set the enc_dig_alg field properly based on the signing
18119 * Allow PKCS#12 password to be set from the command line or the
18120 environment. Let 'ca' get its config file name from the environment
18121 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18126 * Allow certificate policies extension to use an IA5STRING for the
18127 organization field. This is contrary to the PKIX definition but
18128 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18133 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18134 without disallowing inline assembler and the like for non-pedantic builds.
18138 * Support Borland C++ builder.
18140 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18146 * SHA-1 cleanups and performance enhancements.
18148 *Andy Polyakov <appro@fy.chalmers.se>*
18150 * Sparc v8plus assembler for the bignum library.
18152 *Andy Polyakov <appro@fy.chalmers.se>*
18154 * Accept any -xxx and +xxx compiler options in Configure.
18158 * Update HPUX configuration.
18162 * Add missing `sk_<type>_unshift()` function to safestack.h
18164 *Ralf S. Engelschall*
18166 * New function SSL_CTX_use_certificate_chain_file that sets the
18167 "extra_cert"s in addition to the certificate. (This makes sense
18168 only for "PEM" format files, as chains as a whole are not
18173 * Support verify_depth from the SSL API.
18174 x509_vfy.c had what can be considered an off-by-one-error:
18175 Its depth (which was not part of the external interface)
18176 was actually counting the number of certificates in a chain;
18177 now it really counts the depth.
18181 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18182 instead of X509err, which often resulted in confusing error
18183 messages since the error codes are not globally unique
18184 (e.g. an alleged error in ssl3_accept when a certificate
18185 didn't match the private key).
18187 * New function SSL_CTX_set_session_id_context that allows to set a default
18188 value (so that you don't need SSL_set_session_id_context for each
18189 connection using the SSL_CTX).
18193 * OAEP decoding bug fix.
18197 * Support INSTALL_PREFIX for package builders, as proposed by
18202 * New Configure options "threads" and "no-threads". For systems
18203 where the proper compiler options are known (currently Solaris
18204 and Linux), "threads" is the default.
18208 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18212 * Install various scripts to $(OPENSSLDIR)/misc, not to
18213 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18214 such as /usr/local/bin.
18218 * "make linux-shared" to build shared libraries.
18220 *Niels Poppe <niels@netbox.org>*
18222 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18226 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18227 extension adding in x509 utility.
18231 * Remove NOPROTO sections and error code comments.
18235 * Partial rewrite of the DEF file generator to now parse the ANSI
18240 * New Configure options --prefix=DIR and --openssldir=DIR.
18244 * Complete rewrite of the error code script(s). It is all now handled
18245 by one script at the top level which handles error code gathering,
18246 header rewriting and C source file generation. It should be much better
18247 than the old method: it now uses a modified version of Ulf's parser to
18248 read the ANSI prototypes in all header files (thus the old K&R definitions
18249 aren't needed for error creation any more) and do a better job of
18250 translating function codes into names. The old 'ASN1 error code embedded
18251 in a comment' is no longer necessary and it doesn't use .err files which
18252 have now been deleted. Also the error code call doesn't have to appear all
18253 on one line (which resulted in some large lines...).
18257 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18261 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18262 0 (which usually indicates a closed connection), but continue reading.
18266 * Fix some race conditions.
18270 * Add support for CRL distribution points extension. Add Certificate
18271 Policies and CRL distribution points documentation.
18275 * Move the autogenerated header file parts to crypto/opensslconf.h.
18279 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18280 8 of keying material. Merlin has also confirmed interop with this fix
18281 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18283 *Merlin Hughes <merlin@baltimore.ie>*
18285 * Fix lots of warnings.
18287 *Richard Levitte <levitte@stacken.kth.se>*
18289 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18290 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18292 *Richard Levitte <levitte@stacken.kth.se>*
18294 * Fix problems with sizeof(long) == 8.
18296 *Andy Polyakov <appro@fy.chalmers.se>*
18298 * Change functions to ANSI C.
18302 * Fix typos in error codes.
18304 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18306 * Remove defunct assembler files from Configure.
18310 * SPARC v8 assembler BIGNUM implementation.
18312 *Andy Polyakov <appro@fy.chalmers.se>*
18314 * Support for Certificate Policies extension: both print and set.
18315 Various additions to support the r2i method this uses.
18319 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18320 return a const string when you are expecting an allocated buffer.
18324 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18325 types DirectoryString and DisplayText.
18329 * Add code to allow r2i extensions to access the configuration database,
18330 add an LHASH database driver and add several ctx helper functions.
18334 * Fix an evil bug in bn_expand2() which caused various BN functions to
18335 fail when they extended the size of a BIGNUM.
18339 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18340 support typesafe stack.
18344 * Fix typo in SSL_[gs]et_options().
18346 *Nils Frostberg <nils@medcom.se>*
18348 * Delete various functions and files that belonged to the (now obsolete)
18349 old X509V3 handling code.
18353 * New Configure option "rsaref".
18357 * Don't auto-generate pem.h.
18361 * Introduce type-safe ASN.1 SETs.
18365 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18367 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18369 * Introduce type-safe STACKs. This will almost certainly break lots of code
18370 that links with OpenSSL (well at least cause lots of warnings), but fear
18371 not: the conversion is trivial, and it eliminates loads of evil casts. A
18372 few STACKed things have been converted already. Feel free to convert more.
18373 In the fullness of time, I'll do away with the STACK type altogether.
18377 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18378 specified in `<certfile>` by updating the entry in the index.txt file.
18379 This way one no longer has to edit the index.txt file manually for
18380 revoking a certificate. The -revoke option does the gory details now.
18382 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18384 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18385 `-text` option at all and this way the `-noout -text` combination was
18386 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18388 *Ralf S. Engelschall*
18390 * Make sure a corresponding plain text error message exists for the
18391 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18392 verify callback function determined that a certificate was revoked.
18394 *Ralf S. Engelschall*
18396 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18397 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18398 all available ciphers including rc5, which was forgotten until now.
18399 In order to let the testing shell script know which algorithms
18400 are available, a new (up to now undocumented) command
18401 `openssl list-cipher-commands` is used.
18405 * Bugfix: s_client occasionally would sleep in select() when
18406 it should have checked SSL_pending() first.
18410 * New functions DSA_do_sign and DSA_do_verify to provide access to
18411 the raw DSA values prior to ASN.1 encoding.
18415 * Tweaks to Configure
18417 *Niels Poppe <niels@netbox.org>*
18419 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18424 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18428 * New config option to avoid instructions that are illegal on the 80386.
18429 The default code is faster, but requires at least a 486.
18433 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18434 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18435 same as SSL2_VERSION anyway.
18439 * New "-showcerts" option for s_client.
18443 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18444 application. Various cleanups and fixes.
18448 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18449 modify error routines to work internally. Add error codes and PBE init
18450 to library startup routines.
18454 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18455 packing functions to asn1 and evp. Changed function names and error
18456 codes along the way.
18460 * PKCS12 integration: and so it begins... First of several patches to
18461 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18462 objects to objects.h
18466 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18467 and display support for Thawte strong extranet extension.
18471 * Add LinuxPPC support.
18473 *Jeff Dubrule <igor@pobox.org>*
18475 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18476 bn_div_words in alpha.s.
18478 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18480 * Make sure the RSA OAEP test is skipped under -DRSAref because
18481 OAEP isn't supported when OpenSSL is built with RSAref.
18483 *Ulf Moeller <ulf@fitug.de>*
18485 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18486 so they no longer are missing under -DNOPROTO.
18488 *Soren S. Jorvang <soren@t.dk>*
18490 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18492 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18493 doesn't work when the session is reused. Coming soon!
18497 * Fix a security hole, that allows sessions to be reused in the wrong
18498 context thus bypassing client cert protection! All software that uses
18499 client certs and session caches in multiple contexts NEEDS PATCHING to
18500 allow session reuse! A fuller solution is in the works.
18502 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18504 * Some more source tree cleanups (removed obsolete files
18505 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18506 permission on "config" script to be executable) and a fix for the INSTALL
18509 *Ulf Moeller <ulf@fitug.de>*
18511 * Remove some legacy and erroneous uses of malloc, free instead of
18514 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18516 * Make rsa_oaep_test return non-zero on error.
18518 *Ulf Moeller <ulf@fitug.de>*
18520 * Add support for native Solaris shared libraries. Configure
18521 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18522 if someone would make that last step automatic.
18524 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18526 * ctx_size was not built with the right compiler during "make links". Fixed.
18530 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18531 except NULL ciphers". This means the default cipher list will no longer
18532 enable NULL ciphers. They need to be specifically enabled e.g. with
18533 the string "DEFAULT:eNULL".
18537 * Fix to RSA private encryption routines: if p < q then it would
18538 occasionally produce an invalid result. This will only happen with
18539 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18543 * Be less restrictive and allow also `perl util/perlpath.pl
18544 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18545 because this way one can also use an interpreter named `perl5` (which is
18546 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18547 installed as `perl`).
18549 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18551 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18553 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18555 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18556 advapi32.lib to Win32 build and change the pem test comparison
18557 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18558 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18559 and crypto/des/ede_cbcm_enc.c.
18563 * DES quad checksum was broken on big-endian architectures. Fixed.
18567 * Comment out two functions in bio.h that aren't implemented. Fix up the
18568 Win32 test batch file so it (might) work again. The Win32 test batch file
18569 is horrible: I feel ill....
18573 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18574 in e_os.h. Audit of header files to check ANSI and non ANSI
18575 sections: 10 functions were absent from non ANSI section and not exported
18576 from Windows DLLs. Fixed up libeay.num for new functions.
18580 * Make `openssl version` output lines consistent.
18582 *Ralf S. Engelschall*
18584 * Fix Win32 symbol export lists for BIO functions: Added
18585 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18586 to ms/libeay{16,32}.def.
18588 *Ralf S. Engelschall*
18590 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18591 fine under Unix and passes some trivial tests I've now added. But the
18592 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18593 added to make sure no one expects that this stuff really works in the
18594 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18595 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18598 *Ralf S. Engelschall*
18600 * Fix the generation of two part addresses in perl.
18602 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18604 * Add config entry for Linux on MIPS.
18606 *John Tobey <jtobey@channel1.com>*
18608 * Make links whenever Configure is run, unless we are on Windoze.
18612 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18613 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18618 * Add a useful kludge to allow package maintainers to specify compiler and
18619 other platforms details on the command line without having to patch the
18620 Configure script every time: One now can use
18621 `perl Configure <id>:<details>`,
18622 i.e. platform ids are allowed to have details appended
18623 to them (separated by colons). This is treated as there would be a static
18624 pre-configured entry in Configure's %table under key `<id>` with value
18625 `<details>` and `perl Configure <id>` is called. So, when you want to
18626 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18627 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18628 now, which overrides the FreeBSD-elf entry on-the-fly.
18630 *Ralf S. Engelschall*
18632 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18636 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18637 on the `perl Configure ...` command line. This way one can compile
18638 OpenSSL libraries with Position Independent Code (PIC) which is needed
18639 for linking it into DSOs.
18641 *Ralf S. Engelschall*
18643 * Remarkably, export ciphers were totally broken and no-one had noticed!
18648 * Cleaned up the LICENSE document: The official contact for any license
18649 questions now is the OpenSSL core team under openssl-core@openssl.org.
18650 And add a paragraph about the dual-license situation to make sure people
18651 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18652 to the OpenSSL toolkit.
18654 *Ralf S. Engelschall*
18656 * General source tree makefile cleanups: Made `making xxx in yyy...`
18657 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18658 Additionally cleaned up the `make links` target: Remove unnecessary
18659 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18660 to speed processing and no longer clutter the display with confusing
18661 stuff. Instead only the actually done links are displayed.
18663 *Ralf S. Engelschall*
18665 * Permit null encryption ciphersuites, used for authentication only. It used
18666 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18667 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18672 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18673 signed attributes when verifying signatures (this would break them),
18674 the detached data encoding was wrong and public keys obtained using
18675 X509_get_pubkey() weren't freed.
18679 * Add text documentation for the BUFFER functions. Also added a work around
18680 to a Win95 console bug. This was triggered by the password read stuff: the
18681 last character typed gets carried over to the next fread(). If you were
18682 generating a new cert request using 'req' for example then the last
18683 character of the passphrase would be CR which would then enter the first
18688 * Added the new 'Includes OpenSSL Cryptography Software' button as
18689 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18690 button and can be used by applications based on OpenSSL to show the
18691 relationship to the OpenSSL project.
18693 *Ralf S. Engelschall*
18695 * Remove confusing variables in function signatures in files
18696 ssl/ssl_lib.c and ssl/ssl.h.
18698 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18700 * Don't install bss_file.c under PREFIX/include/
18702 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18704 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18705 functions that return function pointers and has support for NT specific
18706 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18707 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18708 unsigned to signed types: this was killing the Win32 compile.
18712 * Add new certificate file to stack functions,
18713 SSL_add_dir_cert_subjects_to_stack() and
18714 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18715 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18716 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18717 This means that Apache-SSL and similar packages don't have to mess around
18718 to add as many CAs as they want to the preferred list.
18722 * Experiment with doxygen documentation. Currently only partially applied to
18724 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18725 openssl.doxy as the configuration file.
18729 * Get rid of remaining C++-style comments which strict C compilers hate.
18731 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18733 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18734 compiled in by default: it has problems with large keys.
18738 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18739 DH private keys and/or callback functions which directly correspond to
18740 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18741 is needed for applications which have to configure certificates on a
18742 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18744 For the RSA certificate situation is makes no difference, but
18745 for the DSA certificate situation this fixes the "no shared cipher"
18746 problem where the OpenSSL cipher selection procedure failed because the
18747 temporary keys were not overtaken from the context and the API provided
18748 no way to reconfigure them.
18749 The new functions now let applications reconfigure the stuff and they
18750 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18751 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18752 non-public-API function ssl_cert_instantiate() is used as a helper
18753 function and also to reduce code redundancy inside ssl_rsa.c.
18755 *Ralf S. Engelschall*
18757 * Move s_server -dcert and -dkey options out of the undocumented feature
18758 area because they are useful for the DSA situation and should be
18759 recognized by the users.
18761 *Ralf S. Engelschall*
18763 * Fix the cipher decision scheme for export ciphers: the export bits are
18764 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18765 SSL_EXP_MASK. So, the original variable has to be used instead of the
18766 already masked variable.
18768 *Richard Levitte <levitte@stacken.kth.se>*
18770 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18772 *Richard Levitte <levitte@stacken.kth.se>*
18774 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18775 from `int` to `unsigned int` because it is a length and initialized by
18776 EVP_DigestFinal() which expects an `unsigned int *`.
18778 *Richard Levitte <levitte@stacken.kth.se>*
18780 * Don't hard-code path to Perl interpreter on shebang line of Configure
18781 script. Instead use the usual Shell->Perl transition trick.
18783 *Ralf S. Engelschall*
18785 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18786 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18787 -noout -modulus` as it's already the case for `openssl rsa -noout
18788 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18789 currently the public key is printed (a decision which was already done by
18790 `openssl dsa -modulus` in the past) which serves a similar purpose.
18791 Additionally the NO_RSA no longer completely removes the whole -modulus
18792 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18795 *Ralf S. Engelschall*
18797 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18798 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18800 *Arne Ansper <arne@ats.cyber.ee>*
18802 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18803 to be added. Now both 'req' and 'ca' can use new objects defined in the
18808 * Add cool BIO that does syslog (or event log on NT).
18810 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18812 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18813 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18814 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18815 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18819 * Add preliminary config info for new extension code.
18823 * Make RSA_NO_PADDING really use no padding.
18825 *Ulf Moeller <ulf@fitug.de>*
18827 * Generate errors when private/public key check is done.
18831 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18832 for some CRL extensions and new objects added.
18836 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18837 key usage extension and fuller support for authority key id.
18841 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18842 padding method for RSA, which is recommended for new applications in PKCS
18843 #1 v2.0 (RFC 2437, October 1998).
18844 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18845 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18846 against Bleichbacher's attack on RSA.
18847 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18850 * Updates to the new SSL compression code
18852 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18854 * Fix so that the version number in the master secret, when passed
18855 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18856 (because the server will not accept higher), that the version number
18857 is 0x03,0x01, not 0x03,0x00
18859 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18861 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18862 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18863 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18867 * Support for RAW extensions where an arbitrary extension can be
18868 created by including its DER encoding. See `apps/openssl.cnf` for
18873 * Make sure latest Perl versions don't interpret some generated C array
18874 code as Perl array code in the crypto/err/err_genc.pl script.
18876 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18878 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18879 not many people have the assembler. Various Win32 compilation fixes and
18880 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18881 build instructions.
18885 * Modify configure script 'Configure' to automatically create crypto/date.h
18886 file under Win32 and also build pem.h from pem.org. New script
18887 util/mkfiles.pl to create the MINFO file on environments that can't do a
18888 'make files': perl util/mkfiles.pl >MINFO should work.
18892 * Major rework of DES function declarations, in the pursuit of correctness
18893 and purity. As a result, many evil casts evaporated, and some weirdness,
18894 too. You may find this causes warnings in your code. Zapping your evil
18895 casts will probably fix them. Mostly.
18899 * Fix for a typo in asn1.h. Bug fix to object creation script
18900 obj_dat.pl. It considered a zero in an object definition to mean
18901 "end of object": none of the objects in objects.h have any zeros
18902 so it wasn't spotted.
18904 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18906 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18907 Masking (CBCM). In the absence of test vectors, the best I have been able
18908 to do is check that the decrypt undoes the encrypt, so far. Send me test
18909 vectors if you have them.
18913 * Correct calculation of key length for export ciphers (too much space was
18914 allocated for null ciphers). This has not been tested!
18918 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18919 message is now correct (it understands "crypto" and "ssl" on its
18920 command line). There is also now an "update" option. This will update
18921 the util/ssleay.num and util/libeay.num files with any new functions.
18923 perl util/mkdef.pl crypto ssl update
18924 it will update them.
18928 * Overhauled the Perl interface:
18929 - ported BN stuff to OpenSSL's different BN library
18930 - made the perl/ source tree CVS-aware
18931 - renamed the package from SSLeay to OpenSSL (the files still contain
18932 their history because I've copied them in the repository)
18933 - removed obsolete files (the test scripts will be replaced
18934 by better Test::Harness variants in the future)
18936 *Ralf S. Engelschall*
18938 * First cut for a very conservative source tree cleanup:
18939 1. merge various obsolete readme texts into doc/ssleay.txt
18940 where we collect the old documents and readme texts.
18941 2. remove the first part of files where I'm already sure that we no
18942 longer need them because of three reasons: either they are just temporary
18943 files which were left by Eric or they are preserved original files where
18944 I've verified that the diff is also available in the CVS via "cvs diff
18945 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18946 the crypto/md/ stuff).
18948 *Ralf S. Engelschall*
18950 * More extension code. Incomplete support for subject and issuer alt
18951 name, issuer and authority key id. Change the i2v function parameters
18952 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18953 what that's for :-) Fix to ASN1 macro which messed up
18954 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18958 * Preliminary support for ENUMERATED type. This is largely copied from the
18963 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18965 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18967 * Make sure `make rehash` target really finds the `openssl` program.
18969 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18971 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18972 like to hear about it if this slows down other processors.
18976 * Add CygWin32 platform information to Configure script.
18978 *Alan Batie <batie@aahz.jf.intel.com>*
18980 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18982 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18984 * New program nseq to manipulate netscape certificate sequences
18988 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18993 * Fixes to BN code. Previously the default was to define BN_RECURSION
18994 but the BN code had some problems that would cause failures when
18995 doing certificate verification and some other functions.
18997 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18999 * Add ASN1 and PEM code to support netscape certificate sequences.
19003 * Add ASN1 and PEM code to support netscape certificate sequences.
19007 * Add several PKIX and private extended key usage OIDs.
19011 * Modify the 'ca' program to handle the new extension code. Modify
19012 openssl.cnf for new extension format, add comments.
19016 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19017 and add a sample to openssl.cnf so req -x509 now adds appropriate
19022 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19023 error code, add initial support to X509_print() and x509 application.
19027 * Takes a deep breath and start adding X509 V3 extension support code. Add
19028 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19029 stuff is currently isolated and isn't even compiled yet.
19033 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19034 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19035 Removed the versions check from X509 routines when loading extensions:
19036 this allows certain broken certificates that don't set the version
19037 properly to be processed.
19041 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19042 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19043 can still be regenerated with "make depend".
19047 * Spelling mistake in C version of CAST-128.
19049 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19051 * Changes to the error generation code. The perl script err-code.pl
19052 now reads in the old error codes and retains the old numbers, only
19053 adding new ones if necessary. It also only changes the .err files if new
19054 codes are added. The makefiles have been modified to only insert errors
19055 when needed (to avoid needlessly modifying header files). This is done
19056 by only inserting errors if the .err file is newer than the auto generated
19057 C file. To rebuild all the error codes from scratch (the old behaviour)
19058 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19059 or delete all the .err files.
19063 * CAST-128 was incorrectly implemented for short keys. The C version has
19064 been fixed, but is untested. The assembler versions are also fixed, but
19065 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19066 to regenerate it if needed.
19067 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19068 Hagino <itojun@kame.net>*
19070 * File was opened incorrectly in randfile.c.
19072 *Ulf Möller <ulf@fitug.de>*
19074 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19075 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19076 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19077 al: it's just almost always a UTCTime. Note this patch adds new error
19078 codes so do a "make errors" if there are problems.
19082 * Correct Linux 1 recognition in config.
19084 *Ulf Möller <ulf@fitug.de>*
19086 * Remove pointless MD5 hash when using DSA keys in ca.
19088 *Anonymous <nobody@replay.com>*
19090 * Generate an error if given an empty string as a cert directory. Also
19091 generate an error if handed NULL (previously returned 0 to indicate an
19092 error, but didn't set one).
19094 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19096 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19100 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19101 parameters. This was causing a warning which killed off the Win32 compile.
19105 * Remove C++ style comments from crypto/bn/bn_local.h.
19107 *Neil Costigan <neil.costigan@celocom.com>*
19109 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19110 based on a text string, looking up short and long names and finally
19111 "dot" format. The "dot" format stuff didn't work. Added new function
19112 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19113 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19114 OID is not part of the table.
19118 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19119 X509_LOOKUP_by_alias().
19123 * Sort openssl functions by name.
19127 * Get the `gendsa` command working and add it to the `list` command. Remove
19128 encryption from sample DSA keys (in case anyone is interested the password
19133 * Make *all* `*_free` functions accept a NULL pointer.
19135 *Frans Heymans <fheymans@isaserver.be>*
19137 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19140 *Anonymous <nobody@replay.com>*
19142 * s_server should send the CAfile as acceptable CAs, not its own cert.
19144 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19146 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19148 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19150 * Temp key "for export" tests were wrong in s3_srvr.c.
19152 *Anonymous <nobody@replay.com>*
19154 * Add prototype for temp key callback functions
19155 SSL_CTX_set_tmp_{rsa,dh}_callback().
19159 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19160 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19164 * X509_name_add_entry() freed the wrong thing after an error.
19166 *Arne Ansper <arne@ats.cyber.ee>*
19168 * rsa_eay.c would attempt to free a NULL context.
19170 *Arne Ansper <arne@ats.cyber.ee>*
19172 * BIO_s_socket() had a broken should_retry() on Windoze.
19174 *Arne Ansper <arne@ats.cyber.ee>*
19176 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19178 *Arne Ansper <arne@ats.cyber.ee>*
19180 * Make sure the already existing X509_STORE->depth variable is initialized
19181 in X509_STORE_new(), but document the fact that this variable is still
19182 unused in the certificate verification process.
19184 *Ralf S. Engelschall*
19186 * Fix the various library and `apps/` files to free up pkeys obtained from
19187 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19191 * Fix reference counting in X509_PUBKEY_get(). This makes
19192 demos/maurice/example2.c work, amongst others, probably.
19194 *Steve Henson and Ben Laurie*
19196 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19197 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19198 are no longer created. This way we have a single and consistent command
19199 line interface `openssl <command>`, similar to `cvs <command>`.
19201 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19203 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19204 BIT STRING wrapper always have zero unused bits.
19208 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19212 * Make the top-level INSTALL documentation easier to understand.
19216 * Makefiles updated to exit if an error occurs in a sub-directory
19217 make (including if user presses ^C) [Paul Sutton]
19219 * Make Montgomery context stuff explicit in RSA data structure.
19223 * Fix build order of pem and err to allow for generated pem.h.
19227 * Fix renumbering bug in X509_NAME_delete_entry().
19231 * Enhanced the err-ins.pl script so it makes the error library number
19232 global and can add a library name. This is needed for external ASN1 and
19233 other error libraries.
19237 * Fixed sk_insert which never worked properly.
19241 * Fix ASN1 macros so they can handle indefinite length constructed
19242 EXPLICIT tags. Some non standard certificates use these: they can now
19247 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19248 into a single doc/ssleay.txt bundle. This way the information is still
19249 preserved but no longer messes up this directory. Now it's new room for
19250 the new set of documentation files.
19252 *Ralf S. Engelschall*
19254 * SETs were incorrectly DER encoded. This was a major pain, because they
19255 shared code with SEQUENCEs, which aren't coded the same. This means that
19256 almost everything to do with SETs or SEQUENCEs has either changed name or
19257 number of arguments.
19259 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19261 * Fix test data to work with the above.
19265 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19266 was already fixed by Eric for 0.9.1 it seems.
19268 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19270 * Autodetect FreeBSD3.
19274 * Fix various bugs in Configure. This affects the following platforms:
19278 unixware-2.0-pentium
19283 * Eliminate generated files from CVS. Reorder tests to regenerate files
19284 before they are needed.
19288 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19292 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19294 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19295 changed SSLeay to OpenSSL in version strings.
19297 *Ralf S. Engelschall*
19299 * Some fixups to the top-level documents.
19303 * Fixed the nasty bug where rsaref.h was not found under compile-time
19304 because the symlink to include/ was missing.
19306 *Ralf S. Engelschall*
19308 * Incorporated the popular no-RSA/DSA-only patches
19309 which allow to compile an RSA-free SSLeay.
19311 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19313 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19314 when "ssleay" is still not found.
19316 *Ralf S. Engelschall*
19318 * Added more platforms to Configure: Cray T3E, HPUX 11,
19320 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19322 * Updated the README file.
19324 *Ralf S. Engelschall*
19326 * Added various .cvsignore files in the CVS repository subdirs
19327 to make a "cvs update" really silent.
19329 *Ralf S. Engelschall*
19331 * Recompiled the error-definition header files and added
19332 missing symbols to the Win32 linker tables.
19334 *Ralf S. Engelschall*
19336 * Cleaned up the top-level documents;
19337 o new files: CHANGES and LICENSE
19338 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19339 o merged COPYRIGHT into LICENSE
19340 o removed obsolete TODO file
19341 o renamed MICROSOFT to INSTALL.W32
19343 *Ralf S. Engelschall*
19345 * Removed dummy files from the 0.9.1b source tree:
19346 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19347 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19348 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19349 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19350 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19352 *Ralf S. Engelschall*
19354 * Added various platform portability fixes.
19358 * The Genesis of the OpenSSL rpject:
19359 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19360 Young and Tim J. Hudson created while they were working for C2Net until
19363 *The OpenSSL Project*
19365 ### Changes between 0.9.0b and 0.9.1b [not released]
19367 * Updated a few CA certificates under certs/
19371 * Changed some BIGNUM api stuff.
19375 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19376 DGUX x86, Linux Alpha, etc.
19380 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19381 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19386 * Add -strparse option to asn1pars program which parses nested
19389 *Dr Stephen Henson <shenson@bigfoot.com>*
19391 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19395 * DSA fix for "ca" program.
19399 * Added "-genkey" option to "dsaparam" program.
19403 * Added RIPE MD160 (rmd160) message digest.
19407 * Added -a (all) option to "ssleay version" command.
19411 * Added PLATFORM define which is the id given to Configure.
19415 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19419 * Extended the ASN.1 parser routines.
19423 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19427 * Added a BN_CTX to the BN library.
19431 * Fixed the weak key values in DES library
19435 * Changed API in EVP library for cipher aliases.
19439 * Added support for RC2/64bit cipher.
19443 * Converted the lhash library to the crypto/mem.c functions.
19447 * Added more recognized ASN.1 object ids.
19451 * Added more RSA padding checks for SSL/TLS.
19455 * Added BIO proxy/filter functionality.
19459 * Added extra_certs to SSL_CTX which can be used
19460 send extra CA certificates to the client in the CA cert chain sending
19461 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19465 * Now Fortezza is denied in the authentication phase because
19466 this is key exchange mechanism is not supported by SSLeay at all.
19470 * Additional PKCS1 checks.
19474 * Support the string "TLSv1" for all TLS v1 ciphers.
19478 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19479 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19483 * Fixed a few memory leaks.
19487 * Fixed various code and comment typos.
19491 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19492 bytes sent in the client random.
19494 *Edward Bishop <ebishop@spyglass.com>*
19498 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19499 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19500 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19501 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19502 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19503 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19504 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19505 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19506 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19507 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19508 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19509 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19510 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19511 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19512 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19513 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19514 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19515 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19516 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19517 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19518 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19519 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19520 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19521 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19522 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19523 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19524 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19525 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19526 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19527 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19528 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19529 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19530 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19531 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19532 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19533 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19534 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19535 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19536 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19537 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19538 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19539 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19540 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19541 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19542 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19543 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19544 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19545 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19546 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19547 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19548 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19549 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19550 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19551 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19552 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19553 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19554 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19555 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19556 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19557 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19558 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19559 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19560 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19561 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19562 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19563 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19564 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19565 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19566 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19567 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19568 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19569 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19570 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19571 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19572 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19573 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19574 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19575 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19576 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19577 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19578 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19579 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19580 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19581 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19582 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19583 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19584 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19585 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19586 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19587 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19588 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19589 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19590 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19591 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19592 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19593 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19594 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19595 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19596 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19597 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19598 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19599 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19600 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19601 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19602 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19603 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19604 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19605 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19606 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19607 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19608 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19609 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19610 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19611 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19612 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19613 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19614 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19615 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19616 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19617 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19618 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19619 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19620 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19621 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19622 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19623 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19624 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19625 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19626 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19627 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19628 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19629 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19630 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19631 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19632 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19633 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19634 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19635 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19636 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19637 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19638 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19639 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19640 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19641 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19642 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19643 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19644 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19645 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19646 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19647 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19648 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19649 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19650 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19651 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19652 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19653 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19654 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19655 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19656 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19657 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19658 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19659 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19660 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19661 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655