4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.2](#openssl-32)
14 - [OpenSSL 3.1](#openssl-31)
15 - [OpenSSL 3.0](#openssl-30)
16 - [OpenSSL 1.1.1](#openssl-111)
17 - [OpenSSL 1.1.0](#openssl-110)
18 - [OpenSSL 1.0.2](#openssl-102)
19 - [OpenSSL 1.0.1](#openssl-101)
20 - [OpenSSL 1.0.0](#openssl-100)
21 - [OpenSSL 0.9.x](#openssl-09x)
26 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
28 * Added support for Hybrid Public Key Encryption (HPKE) as defined
29 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
30 Message Layer Security (MLS) and other IETF specifications.
31 HPKE can also be used by other applications that require
32 encrypting "to" an ECDH public key. External APIs are defined in
33 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
37 * Add support for certificate compression (RFC8879), including
38 library support for Brotli and Zstandard compression.
42 * Add the ability to add custom attributes to PKCS12 files. Add a new API
43 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
44 for a user specified callback and optional argument.
45 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
46 added to the existing STACK_OF attrs.
50 * Major refactor of the libssl record layer
54 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
58 * Add a mac salt length option for the pkcs12 command.
62 * Add more SRTP protection profiles from RFC8723 and RFC8269.
66 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
68 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
70 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
71 supported and enabled.
75 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
76 to the list of ciphersuites providing Perfect Forward Secrecy as
77 required by SECLEVEL >= 3.
79 *Dmitry Belyavskiy, Nicola Tuveri*
81 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
82 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
83 SSL_get0_iana_groups() function-like macro, retrieves the list of
84 supported groups sent by the peer.
85 The function SSL_client_hello_get_extension_order() populates
86 a caller-supplied array with the list of extension types present in the
87 ClientHello, in order of appearance.
91 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
92 to make it possible to use empty passphrase strings.
96 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
97 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
98 of 160 bits and above and less than 224 bits were previously accepted by
99 default but are now no longer allowed. By default TLS compression was
100 already disabled in previous OpenSSL versions. At security level 2 it cannot
105 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
110 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
111 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
112 will need to load the legacy crypto provider.
115 * CCM8 cipher suites in TLS have been downgraded to security level zero
116 because they use a short authentication tag which lowers their strength.
120 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
125 * Add X.509 certificate codeSigning purpose and related checks on key usage and
126 extended key usage of the leaf certificate according to the CA/Browser Forum.
130 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
131 such as adding a trace facility for debugging certificate chain building.
135 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
136 in particular supporting requests for central key generation, generalized
137 polling, and various types of genm/genp exchanges defined in CMP Updates.
141 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
142 like correcting the TLS and proxy support and adding tracing for debugging.
146 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
150 * Fixed and extended `util/check-format.pl` for checking adherence to the
151 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
152 The checks are meanwhile more complete and yield fewer false positives.
156 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
157 sending and receiving multiple messages in a single call. An implementation
158 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
162 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
163 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
164 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
165 paths which are searched for root certificates.
167 The existing `SSL_CERT_DIR` environment variable is deprecated.
168 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
169 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
170 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
171 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
172 for the purposes of determining root certificate stores.
176 * Support for loading root certificates from the Windows certificate store
177 has been added. The support is in the form of a store which recognises the
178 URI string of `org.openssl.winstore://`. This store is enabled by default and
179 can be disabled using the new compile-time option `no-winstore`.
183 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
184 kernel versions that support KTLS have a known bug in CCM processing. That
185 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
186 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
191 * Zerocopy KTLS sendfile() support on Linux.
198 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
200 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
201 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
202 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
203 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
204 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
205 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
206 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
207 for legacy EC and SM2 keys is also changed similarly to honor the
208 equivalent conversion format flag as specified in the underlying
209 `EC_KEY` object being exported to a provider, when this function is
210 called through `EVP_PKEY_export()`.
214 * RNDR and RNDRRS support in provider functions to provide
215 random number generation for Arm CPUs (aarch64).
219 * s_client and s_server apps now explicitly say when the TLS version
220 does not include the renegotiation mechanism. This avoids confusion
221 between that scenario versus when the TLS version includes secure
222 renegotiation but the peer lacks support for it.
226 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
228 *Tomasz Kantecki, Andrey Matyukov*
230 * The various OBJ_* functions have been made thread safe.
234 * Parallel dual-prime 1536/2048-bit modular exponentiation for
235 AVX512_IFMA capable processors.
237 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
239 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
240 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
241 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
242 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
243 `OPENSSL_NO_DEPRECATED_3_1`.
245 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
246 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
247 definitions for these functions regardless of whether
248 `OPENSSL_NO_DEPRECATED_3_1` is defined.
250 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
251 functions regardless of whether they are using them. It is recommended that
252 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
256 * When generating safe-prime DH parameters set the recommended private key
257 length equivalent to minimum key lengths as in RFC 7919.
264 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
265 listed here are only a brief description.
266 The migration guide contains more detailed information related to new features,
267 breaking changes, and mappings for the large list of deprecated functions.
269 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
271 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
273 * Fixed two buffer overflows in punycode decoding functions.
275 A buffer overrun can be triggered in X.509 certificate verification,
276 specifically in name constraint checking. Note that this occurs after
277 certificate chain signature verification and requires either a CA to
278 have signed the malicious certificate or for the application to continue
279 certificate verification despite failure to construct a path to a trusted
282 In a TLS client, this can be triggered by connecting to a malicious
283 server. In a TLS server, this can be triggered if the server requests
284 client authentication and a malicious client connects.
286 An attacker can craft a malicious email address to overflow
287 an arbitrary number of bytes containing the `.` character (decimal 46)
288 on the stack. This buffer overflow could result in a crash (causing a
292 An attacker can craft a malicious email address to overflow four
293 attacker-controlled bytes on the stack. This buffer overflow could
294 result in a crash (causing a denial of service) or potentially remote code
295 execution depending on stack layout for any given platform/compiler.
300 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
301 parameters in OpenSSL code.
302 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
303 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
304 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
305 Using these invalid names may cause algorithms to use slower methods
306 that ignore the CRT parameters.
310 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
315 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
316 data to be signed before signing the certificate.
320 * Added RIPEMD160 to the default provider.
324 * Ensured that the key share group sent or accepted for the key exchange
325 is allowed for the protocol version.
329 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
331 * OpenSSL supports creating a custom cipher via the legacy
332 EVP_CIPHER_meth_new() function and associated function calls. This function
333 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
334 to use the new provider mechanism in order to implement custom ciphers.
336 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
337 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
338 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
339 and decryption initialisation functions). Instead of using the custom cipher
340 directly it incorrectly tries to fetch an equivalent cipher from the
341 available providers. An equivalent cipher is found based on the NID passed to
342 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
343 given cipher. However it is possible for an application to incorrectly pass
344 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
345 is used in this way the OpenSSL encryption/decryption initialisation function
346 will match the NULL cipher as being equivalent and will fetch this from the
347 available providers. This will succeed if the default provider has been
348 loaded (or if a third party provider has been loaded that offers this
349 cipher). Using the NULL cipher means that the plaintext is emitted as the
352 Applications are only affected by this issue if they call
353 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
354 encryption/decryption initialisation function. Applications that only use
355 SSL/TLS are not impacted by this issue.
360 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
365 * Fixed the linux-mips64 Configure target which was missing the
366 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
371 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
376 * Correctly handle a retransmitted ClientHello in DTLS
380 * Fixed detection of ktls support in cross-compile environment on Linux
384 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
389 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
390 report correct results in some cases
394 * Fix UWP builds by defining VirtualLock
398 * For known safe primes use the minimum key length according to RFC 7919.
399 Longer private key sizes unnecessarily raise the cycles needed to compute the
400 shared secret without any increase of the real security. This fixes a
401 regression from 1.1.1 where these shorter keys were generated for the known
406 * Added the loongarch64 target
410 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
411 only passed to the FIPS provider and not to the default or legacy provider.
415 * Fixed reported performance degradation on aarch64. Restored the
416 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
417 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
418 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
419 The new algorithm is still used for 32 bit targets.
423 * Added a missing header for memcmp that caused compilation failure on some
428 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
430 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
431 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
432 This issue makes the RSA implementation with 2048 bit private keys
433 incorrect on such machines and memory corruption will happen during
434 the computation. As a consequence of the memory corruption an attacker
435 may be able to trigger a remote code execution on the machine performing
438 SSL/TLS servers or other servers using 2048 bit RSA private keys running
439 on machines supporting AVX512IFMA instructions of the X86_64 architecture
440 are affected by this issue.
445 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
446 implementation would not encrypt the entirety of the data under some
447 circumstances. This could reveal sixteen bytes of data that was
448 preexisting in the memory that wasn't written. In the special case of
449 "in place" encryption, sixteen bytes of the plaintext would be revealed.
451 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
452 they are both unaffected.
455 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
457 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
459 * In addition to the c_rehash shell command injection identified in
460 CVE-2022-1292, further bugs where the c_rehash script does not
461 properly sanitise shell metacharacters to prevent command injection have been
464 When the CVE-2022-1292 was fixed it was not discovered that there
465 are other places in the script where the file names of certificates
466 being hashed were possibly passed to a command executed through the shell.
468 This script is distributed by some operating systems in a manner where
469 it is automatically executed. On such operating systems, an attacker
470 could execute arbitrary commands with the privileges of the script.
472 Use of the c_rehash script is considered obsolete and should be replaced
473 by the OpenSSL rehash command line tool.
476 *Daniel Fiala, Tomáš Mráz*
478 * Case insensitive string comparison no longer uses locales. It has instead
479 been directly implemented.
483 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
485 * Case insensitive string comparison is reimplemented via new locale-agnostic
486 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
487 comparison. The previous implementation had problems when the Turkish locale
492 * Fixed a bug in the c_rehash script which was not properly sanitising shell
493 metacharacters to prevent command injection. This script is distributed by
494 some operating systems in a manner where it is automatically executed. On
495 such operating systems, an attacker could execute arbitrary commands with the
496 privileges of the script.
498 Use of the c_rehash script is considered obsolete and should be replaced
499 by the OpenSSL rehash command line tool.
504 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
505 certificate on an OCSP response. The bug caused the function in the case
506 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
507 response (meaning a successful verification) even in the case where the
508 response signing certificate fails to verify.
510 It is anticipated that most users of `OCSP_basic_verify` will not use the
511 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
512 a negative value (indicating a fatal error) in the case of a certificate
513 verification failure. The normal expected return value in this case would be
516 This issue also impacts the command line OpenSSL "ocsp" application. When
517 verifying an ocsp response with the "-no_cert_checks" option the command line
518 application will report that the verification is successful even though it
519 has in fact failed. In this case the incorrect successful response will also
520 be accompanied by error messages showing the failure and contradicting the
521 apparently successful result.
526 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
527 AAD data as the MAC key. This made the MAC key trivially predictable.
529 An attacker could exploit this issue by performing a man-in-the-middle attack
530 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
531 that the modified data would still pass the MAC integrity check.
533 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
534 endpoint will always be rejected by the recipient and the connection will
535 fail at that point. Many application protocols require data to be sent from
536 the client to the server first. Therefore, in such a case, only an OpenSSL
537 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
539 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
540 sent in both directions. In this case both clients and servers could be
541 affected, regardless of the application protocol.
543 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
544 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
545 the handshake when using this ciphersuite.
547 The confidentiality of data is not impacted by this issue, i.e. an attacker
548 cannot decrypt data that has been encrypted using this ciphersuite - they can
551 In order for this attack to work both endpoints must legitimately negotiate
552 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
553 OpenSSL 3.0, and is not available within the default provider or the default
554 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
555 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
556 following must have occurred:
558 1) OpenSSL must have been compiled with the (non-default) compile time option
559 enable-weak-ssl-ciphers
561 2) OpenSSL must have had the legacy provider explicitly loaded (either
562 through application code or via configuration)
564 3) The ciphersuite must have been explicitly added to the ciphersuite list
566 4) The libssl security level must have been set to 0 (default is 1)
568 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
570 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
571 others that both endpoints have in common
576 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
577 occupied by the removed hash table entries.
579 This function is used when decoding certificates or keys. If a long lived
580 process periodically decodes certificates or keys its memory usage will
581 expand without bounds and the process might be terminated by the operating
582 system causing a denial of service. Also traversing the empty hash table
583 entries will take increasingly more time.
585 Typically such long lived processes might be TLS clients or TLS servers
586 configured to accept client certificate authentication.
589 *Hugo Landau, Aliaksei Levin*
591 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
592 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
593 statistics are no longer supported. For compatibility, these statistics are
594 still listed in the output but are now always reported as zero.
598 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
600 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
601 for non-prime moduli.
603 Internally this function is used when parsing certificates that contain
604 elliptic curve public keys in compressed form or explicit elliptic curve
605 parameters with a base point encoded in compressed form.
607 It is possible to trigger the infinite loop by crafting a certificate that
608 has invalid explicit curve parameters.
610 Since certificate parsing happens prior to verification of the certificate
611 signature, any process that parses an externally supplied certificate may thus
612 be subject to a denial of service attack. The infinite loop can also be
613 reached when parsing crafted private keys as they can contain explicit
614 elliptic curve parameters.
616 Thus vulnerable situations include:
618 - TLS clients consuming server certificates
619 - TLS servers consuming client certificates
620 - Hosting providers taking certificates or private keys from customers
621 - Certificate authorities parsing certification requests from subscribers
622 - Anything else which parses ASN.1 elliptic curve parameters
624 Also any other applications that use the BN_mod_sqrt() where the attacker
625 can control the parameter values are vulnerable to this DoS issue.
630 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
631 to the list of ciphersuites providing Perfect Forward Secrecy as
632 required by SECLEVEL >= 3.
634 *Dmitry Belyavskiy, Nicola Tuveri*
636 * Made the AES constant time code for no-asm configurations
637 optional due to the resulting 95% performance degradation.
638 The AES constant time code can be enabled, for no assembly
639 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
643 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
648 * The negative return value handling of the certificate verification callback
649 was reverted. The replacement is to set the verification retry state with
650 the SSL_set_retry_verify() function.
654 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
656 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
657 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
658 verify a certificate supplied by a server. That function may return a
659 negative return value to indicate an internal error (for example out of
660 memory). Such a negative return value is mishandled by OpenSSL and will cause
661 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
662 success and a subsequent call to SSL_get_error() to return the value
663 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
664 returned by OpenSSL if the application has previously called
665 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
666 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
667 totally unexpected and applications may not behave correctly as a result. The
668 exact behaviour will depend on the application but it could result in
669 crashes, infinite loops or other similar incorrect responses.
671 This issue is made more serious in combination with a separate bug in OpenSSL
672 3.0 that will cause X509_verify_cert() to indicate an internal error when
673 processing a certificate chain. This will occur where a certificate does not
674 include the Subject Alternative Name extension but where a Certificate
675 Authority has enforced name constraints. This issue can occur even with valid
681 * Corrected a few file name and file reference bugs in the build,
682 installation and setup scripts, which lead to installation verification
683 failures. Slightly enhanced the installation verification script.
687 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
692 * Fixed PVK encoder to properly query for the passphrase.
696 * Multiple fixes in the OSSL_HTTP API functions.
700 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
701 OSSL_PARAM_INTEGER data type and return error on negative numbers
702 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
703 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
707 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
711 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
715 * Multiple threading fixes.
719 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
723 * Allow fetching an operation from the provider that owns an unexportable key
724 as a fallback if that is still allowed by the property query.
728 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
730 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
735 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
736 S390X capability vector to zero. This simplifies testing of different code
737 paths on S390X architecture.
741 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
742 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
743 SP 800-38D". The communication will fail at this point.
747 * The EC_GROUP_clear_free() function is deprecated as there is nothing
748 confidential in EC_GROUP data.
752 * The byte order mark (BOM) character is ignored if encountered at the
753 beginning of a PEM-formatted file.
757 * Added CMS support for the Russian GOST algorithms.
761 * Due to move of the implementation of cryptographic operations
762 to the providers, validation of various operation parameters can
763 be postponed until the actual operation is executed where previously
764 it happened immediately when an operation parameter was set.
766 For example when setting an unsupported curve with
767 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
768 fail but later keygen operations with the EVP_PKEY_CTX will fail.
770 *OpenSSL team members and many third party contributors*
772 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
773 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
774 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
775 instead to retrieve these algorithms from a provider.
779 * On build targets where the multilib postfix is set in the build
780 configuration the libdir directory was changing based on whether
781 the lib directory with the multilib postfix exists on the system
782 or not. This unpredictable behavior was removed and eventual
783 multilib postfix is now always added to the default libdir. Use
784 `--libdir=lib` to override the libdir if adding the postfix is
789 * The triple DES key wrap functionality now conforms to RFC 3217 but is
790 no longer interoperable with OpenSSL 1.1.1.
794 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
795 function codes, this function can only cause problems for calling
800 * Add a configurable flag to output date formats as ISO 8601. Does not
801 change the default date format.
805 * Version of MSVC earlier than 1300 could get link warnings, which could
806 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
807 Support for this flag has been removed.
811 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
812 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
813 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
814 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
815 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
819 * The signatures of the functions to get and set options on SSL and
820 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
821 Some source code changes may be required.
825 * The public definitions of conf_method_st and conf_st have been
826 deprecated. They will be made opaque in a future release.
828 *Rich Salz and Tomáš Mráz*
830 * Client-initiated renegotiation is disabled by default. To allow it, use
831 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
832 flag, or the "ClientRenegotiation" config parameter as appropriate.
836 * Add "abspath" and "includedir" pragma's to config files, to prevent,
837 or modify relative pathname inclusion.
841 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
842 validated. Please consult the README-FIPS and
843 README-PROVIDERS files, as well as the migration guide.
845 *OpenSSL team members and many third party contributors*
847 * For the key types DH and DHX the allowed settable parameters are now different.
851 * The openssl commands that read keys, certificates, and CRLs now
852 automatically detect the PEM or DER format of the input files.
854 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
856 * Added enhanced PKCS#12 APIs which accept a library context.
860 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
864 * Added support for Kernel TLS (KTLS).
866 *Boris Pismenny, John Baldwin and Andrew Gallatin*
868 * Support for RFC 5746 secure renegotiation is now required by default for
869 SSL or TLS connections to succeed.
873 * The signature of the `copy` functional parameter of the
874 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
875 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
876 the signature of the `pub_decode` functional parameter of the
877 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
878 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
882 * The error return values from some control calls (ctrl) have changed.
886 * A public key check is now performed during EVP_PKEY_derive_set_peer().
890 * Many functions in the EVP_ namespace that are getters of values from
891 implementations or contexts were renamed to include get or get0 in their
892 names. Old names are provided as macro aliases for compatibility and
897 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
898 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
899 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
904 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
907 * The output from the command line applications may have minor
912 * The output from numerous "printing" may have minor changes.
916 * Windows thread synchronization uses read/write primitives (SRWLock) when
917 supported by the OS, otherwise CriticalSection continues to be used.
921 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
922 work on read only BIO source/sinks that do not support these functions.
923 This allows piping or redirection of a file BIO using stdin to be buffered
924 into memory. This is used internally in OSSL_DECODER_from_bio().
928 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
929 this function would return one of the values OSSL_STORE_INFO_NAME,
930 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
931 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
932 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
933 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
934 using this function should be amended to handle the changed return value.
938 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
939 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
940 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
941 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
942 contains more than one certificate identifier: This means that all
943 certificates referenced there MUST be part of the validation chain.
947 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
948 RC5, DESX and DES have been moved to the legacy provider.
952 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
953 RIPEMD-160 have been moved to the legacy provider.
957 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
962 * The deprecated functions EVP_PKEY_get0_RSA(),
963 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
964 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
965 well as the similarly named "get1" functions behave differently in
970 * A number of functions handling low-level keys or engines were deprecated
971 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
972 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
973 EVP_PKEY_get0_siphash().
977 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
978 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
979 will need to load the legacy crypto provider. This includes these PBE
980 algorithms which use this KDF:
981 - NID_pbeWithMD2AndDES_CBC
982 - NID_pbeWithMD5AndDES_CBC
983 - NID_pbeWithSHA1AndRC2_CBC
984 - NID_pbeWithMD2AndRC2_CBC
985 - NID_pbeWithMD5AndRC2_CBC
986 - NID_pbeWithSHA1AndDES_CBC
990 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
991 BIO_debug_callback() functions.
995 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
996 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1000 * The RAND_METHOD APIs have been deprecated.
1004 * The SRP APIs have been deprecated.
1008 * Add a compile time option to prevent the caching of provider fetched
1009 algorithms. This is enabled by including the no-cached-fetch option
1010 at configuration time.
1014 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1015 count of PKCS12_DEFAULT_ITER.
1017 *Tomáš Mráz and Sahana Prasad*
1019 * The openssl speed command does not use low-level API calls anymore.
1023 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1026 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1028 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1032 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1033 providers may supply their own group implementations (using either the "key
1034 exchange" or the "key encapsulation" methods) which will automatically be
1035 detected and used by libssl.
1037 *Matt Caswell, Nicola Tuveri*
1039 * The undocumented function X509_certificate_type() has been deprecated;
1043 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1047 * Removed RSA padding mode for SSLv23 (which was only used for
1048 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1049 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1054 * Deprecated the obsolete X9.31 RSA key generation related functions.
1056 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1057 is not allowed to return a value > 1, this is no more taken as failure.
1059 *Viktor Dukhovni and David von Oheimb*
1061 * Deprecated the obsolete X9.31 RSA key generation related functions
1062 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1063 BN_X931_generate_prime_ex().
1067 * The default key generation method for the regular 2-prime RSA keys was
1068 changed to the FIPS 186-4 B.3.6 method.
1072 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1076 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1080 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1081 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1083 *Rich Salz, Richard Levitte, and David von Oheimb*
1085 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1089 * Deprecated `OCSP_parse_url()`.
1093 * Validation of SM2 keys has been separated from the validation of regular EC
1098 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1099 switches: a validation failure triggers an early exit, returning a failure
1100 exit status to the parent process.
1104 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1105 to ignore unknown ciphers.
1109 * The `-cipher-commands` and `-digest-commands` options
1110 of the command line utility `list` have been deprecated.
1111 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1115 * Added convenience functions for generating asymmetric key pairs:
1116 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1117 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1121 * All of the low-level EC_KEY functions have been deprecated.
1123 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1125 * Deprecated all the libcrypto and libssl error string loading
1130 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1131 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1136 * The `-crypt` option to the `passwd` command line tool has been removed.
1140 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1145 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1149 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1150 EVP_PKEY_get1_tls_encodedpoint().
1154 * The security callback, which can be customised by application code, supports
1155 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1156 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1160 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1161 interface. Their functionality remains unchanged.
1165 * Added new option for 'openssl list', '-providers', which will display the
1166 list of loaded providers, their names, version and status. It optionally
1167 displays their gettable parameters.
1171 * Removed EVP_PKEY_set_alias_type().
1175 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1176 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1180 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1181 parameters are still checked for type safety at compile time via helper
1186 * Remove the RAND_DRBG API
1188 *Paul Dale and Matthias St. Pierre*
1190 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1191 as well as actual hostnames.
1195 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1196 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1197 conversely, silently ignore DTLS protocol version bounds when configuring
1198 TLS-based contexts. The commands can be repeated to set bounds of both
1199 types. The same applies with the corresponding "min_protocol" and
1200 "max_protocol" command-line switches, in case some application uses both TLS
1203 SSL_CTX instances that are created for a fixed protocol version (e.g.
1204 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1205 attempts to apply bounds to these protocol versions would result in an
1206 error. Now only the "version-flexible" SSL_CTX instances are subject to
1207 limits in configuration files in command-line options.
1211 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1216 * Reworked the recorded ERR codes to make better space for system errors.
1217 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1218 given code is a system error (true) or an OpenSSL error (false).
1222 * Reworked the test perl framework to better allow parallel testing.
1224 *Nicola Tuveri and David von Oheimb*
1226 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1227 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1231 * 'Configure' has been changed to figure out the configuration target if
1232 none is given on the command line. Consequently, the 'config' script is
1233 now only a mere wrapper. All documentation is changed to only mention
1236 *Rich Salz and Richard Levitte*
1238 * Added a library context `OSSL_LIB_CTX` that applications as well as
1239 other libraries can use to form a separate context within which
1240 libcrypto operations are performed.
1244 * Added various `_ex` functions to the OpenSSL API that support using
1245 a non-default `OSSL_LIB_CTX`.
1249 * Handshake now fails if Extended Master Secret extension is dropped
1254 * Dropped interactive mode from the `openssl` program.
1258 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1260 *David von Oheimb and Shane Lontis*
1262 * Deprecated `EC_METHOD_get_field_type()`.
1266 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1267 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1268 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1272 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1276 * Add CAdES-BES signature verification support, mostly derived
1277 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1279 *Filipe Raimundo da Silva*
1281 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1285 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1286 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1290 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1294 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1295 EC_KEY_precompute_mult().
1299 * Deprecated EC_POINTs_mul().
1303 * Removed FIPS_mode() and FIPS_mode_set().
1307 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1311 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1312 EC_POINT_get_Jprojective_coordinates_GFp().
1316 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1317 arrays to be more easily constructed via a series of utility functions.
1318 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1319 the various push functions and finally convert to a passable OSSL_PARAM
1320 array using OSSL_PARAM_BLD_to_param().
1324 * The security strength of SHA1 and MD5 based signatures in TLS has been
1329 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1330 contain a provider side internal key.
1334 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1338 * Project text documents not yet having a proper file name extension
1339 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1340 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1344 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1345 have been converted to Markdown with the goal to produce documents
1346 which not only look pretty when viewed online in the browser, but
1347 remain well readable inside a plain text editor.
1349 To achieve this goal, a 'minimalistic' Markdown style has been applied
1350 which avoids formatting elements that interfere too much with the
1351 reading flow in the text file. For example, it
1353 * avoids [ATX headings][] and uses [setext headings][] instead
1354 (which works for `<h1>` and `<h2>` headings only).
1355 * avoids [inline links][] and uses [reference links][] instead.
1356 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1358 [ATX headings]: https://github.github.com/gfm/#atx-headings
1359 [setext headings]: https://github.github.com/gfm/#setext-headings
1360 [inline links]: https://github.github.com/gfm/#inline-link
1361 [reference links]: https://github.github.com/gfm/#reference-link
1362 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1363 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1365 *Matthias St. Pierre*
1367 * The test suite is changed to preserve results of each test recipe.
1368 A new directory test-runs/ with subdirectories named like the
1369 test recipes are created in the build tree for this purpose.
1373 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1374 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1375 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1377 *David von Oheimb, Martin Peylo*
1379 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1380 It supports arbitrary request and response content types, GET redirection,
1381 TLS, connections via HTTP(S) proxies, connections and exchange via
1382 user-defined BIOs (allowing implicit connections), persistent connections,
1383 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1384 The legacy OCSP-focused (and only partly documented) API
1385 is retained for backward compatibility, while most of it is deprecated.
1389 * Added `util/check-format.pl`, a tool for checking adherence to the
1390 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1391 The checks performed are incomplete and yield some false positives.
1392 Still the tool should be useful for detecting most typical glitches.
1396 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1397 If domain name resolution yields multiple IP addresses all of them are tried
1398 after `connect()` failures.
1402 * All of the low-level RSA functions have been deprecated.
1406 * X509 certificates signed using SHA1 are no longer allowed at security
1411 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1412 modified to use PKEY APIs. These commands are now in maintenance mode
1413 and no new features will be added to them.
1417 * The command line utility rsautl has been deprecated.
1421 * The command line utilities genrsa and rsa have been modified to use PKEY
1422 APIs. They now write PKCS#8 keys by default. These commands are now in
1423 maintenance mode and no new features will be added to them.
1427 * All of the low-level DH functions have been deprecated.
1429 *Paul Dale and Matt Caswell*
1431 * All of the low-level DSA functions have been deprecated.
1435 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1436 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1440 * Deprecated low-level ECDH and ECDSA functions.
1444 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1448 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1449 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1450 a new formulation to include all the things it can be used for,
1451 as well as words of caution.
1455 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1459 * All of the low-level HMAC functions have been deprecated.
1461 *Paul Dale and David von Oheimb*
1463 * Over two thousand fixes were made to the documentation, including:
1464 - Common options (such as -rand/-writerand, TLS version control, etc)
1465 were refactored and point to newly-enhanced descriptions in openssl.pod.
1466 - Added style conformance for all options (with help from Richard Levitte),
1467 documented all reported missing options, added a CI build to check
1468 that all options are documented and that no unimplemented options
1470 - Documented some internals, such as all use of environment variables.
1471 - Addressed all internal broken L<> references.
1475 * All of the low-level CMAC functions have been deprecated.
1479 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1480 functions have been deprecated.
1482 *Paul Dale and David von Oheimb*
1484 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1485 set of functions. The documentation mentioned negative values for some
1486 errors, but this was never the case, so the mention of negative values
1489 Code that followed the documentation and thereby check with something
1490 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1494 * All of the low-level cipher functions have been deprecated.
1496 *Matt Caswell and Paul Dale*
1498 * Removed include/openssl/opensslconf.h.in and replaced it with
1499 include/openssl/configuration.h.in, which differs in not including
1500 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1501 was added to include both.
1503 This allows internal hacks where one might need to modify the set
1504 of configured macros, for example this if deprecated symbols are
1505 still supposed to be available internally:
1507 #include <openssl/configuration.h>
1509 #undef OPENSSL_NO_DEPRECATED
1510 #define OPENSSL_SUPPRESS_DEPRECATED
1512 #include <openssl/macros.h>
1514 This should not be used by applications that use the exported
1515 symbols, as that will lead to linking errors.
1519 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1520 used in exponentiation with 512-bit moduli. No EC algorithms are
1521 affected. Analysis suggests that attacks against 2-prime RSA1024,
1522 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1523 difficult to perform and are not believed likely. Attacks against DH512
1524 are considered just feasible. However, for an attack the target would
1525 have to re-use the DH512 private key, which is not recommended anyway.
1526 Also applications directly using the low-level API BN_mod_exp may be
1527 affected if they use BN_FLG_CONSTTIME.
1532 * Most memory-debug features have been deprecated, and the functionality
1533 replaced with no-ops.
1537 * Added documentation for the STACK API.
1541 * Introduced a new method type and API, OSSL_ENCODER, to represent
1542 generic encoders. These do the same sort of job that PEM writers
1543 and d2i functions do, but with support for methods supplied by
1544 providers, and the possibility for providers to support other
1549 * Introduced a new method type and API, OSSL_DECODER, to represent
1550 generic decoders. These do the same sort of job that PEM readers
1551 and i2d functions do, but with support for methods supplied by
1552 providers, and the possibility for providers to support other
1557 * Added a .pragma directive to the syntax of configuration files, to
1558 allow varying behavior in a supported and predictable manner.
1559 Currently added pragma:
1563 This allows dollar signs to be a keyword character unless it's
1564 followed by a opening brace or parenthesis. This is useful for
1565 platforms where dollar signs are commonly used in names, such as
1566 volume names and system directory names on VMS.
1570 * Added functionality to create an EVP_PKEY from user data.
1574 * Change the interpretation of the '--api' configuration option to
1575 mean that this is a desired API compatibility level with no
1576 further meaning. The previous interpretation, that this would
1577 also mean to remove all deprecated symbols up to and including
1578 the given version, no requires that 'no-deprecated' is also used
1579 in the configuration.
1581 When building applications, the desired API compatibility level
1582 can be set with the OPENSSL_API_COMPAT macro like before. For
1583 API compatibility version below 3.0, the old style numerical
1584 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1585 For version 3.0 and on, the value is expected to be the decimal
1586 value calculated from the major and minor version like this:
1588 MAJOR * 10000 + MINOR * 100
1592 -DOPENSSL_API_COMPAT=30000 For 3.0
1593 -DOPENSSL_API_COMPAT=30200 For 3.2
1595 To hide declarations that are deprecated up to and including the
1596 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1597 given when building the application as well.
1601 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1602 access to certificate and CRL stores via URIs and OSSL_STORE
1605 This adds the following functions:
1607 - X509_LOOKUP_store()
1608 - X509_STORE_load_file()
1609 - X509_STORE_load_path()
1610 - X509_STORE_load_store()
1611 - SSL_add_store_cert_subjects_to_stack()
1612 - SSL_CTX_set_default_verify_store()
1613 - SSL_CTX_load_verify_file()
1614 - SSL_CTX_load_verify_dir()
1615 - SSL_CTX_load_verify_store()
1619 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1620 The presence of this system service is determined at run-time.
1624 * Added functionality to create an EVP_PKEY context based on data
1625 for methods from providers. This takes an algorithm name and a
1626 property query string and simply stores them, with the intent
1627 that any operation that uses this context will use those strings
1628 to fetch the needed methods implicitly, thereby making the port
1629 of application written for pre-3.0 OpenSSL easier.
1633 * The undocumented function NCONF_WIN32() has been deprecated; for
1634 conversion details see the HISTORY section of doc/man5/config.pod
1638 * Introduced the new functions EVP_DigestSignInit_ex() and
1639 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1640 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1641 pages for further details.
1645 * Over two thousand fixes were made to the documentation, including:
1646 adding missing command flags, better style conformance, documentation
1649 *Rich Salz, Richard Levitte*
1651 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1652 X25519, X448, Ed25519 and Ed448.
1656 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1661 * Deprecated the public definition of `ERR_STATE` as well as the function
1662 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1667 * Added ERR functionality to give callers access to the stored function
1668 names that have replaced the older function code based functions.
1670 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1671 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1672 ERR_peek_error_all() and ERR_peek_last_error_all().
1674 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1675 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1676 ERR_func_error_string().
1680 * Extended testing to be verbose for failing tests only. The make variables
1681 VERBOSE_FAILURE or VF can be used to enable this:
1683 $ make VF=1 test # Unix
1684 $ mms /macro=(VF=1) test ! OpenVMS
1685 $ nmake VF=1 test # Windows
1689 * Added the `-copy_extensions` option to the `x509` command for use with
1690 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1691 all extensions in the request are copied to the certificate or vice versa.
1693 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1695 * Added the `-copy_extensions` option to the `req` command for use with
1696 `-x509`. When given with the `copy` or `copyall` argument,
1697 all extensions in the certification request are copied to the certificate.
1701 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1702 they generate are by default RFC 5280 compliant in the following sense:
1703 There is a subjectKeyIdentifier extension with a hash value of the public key
1704 and for not self-signed certs there is an authorityKeyIdentifier extension
1705 with a keyIdentifier field or issuer information identifying the signing key.
1706 This is done unless some configuration overrides the new default behavior,
1707 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1711 * Added several checks to `X509_verify_cert()` according to requirements in
1712 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1713 (which may be done by using the CLI option `-x509_strict`):
1714 * The basicConstraints of CA certificates must be marked critical.
1715 * CA certificates must explicitly include the keyUsage extension.
1716 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1717 * The issuer name of any certificate must not be empty.
1718 * The subject name of CA certs, certs with keyUsage crlSign,
1719 and certs without subjectAlternativeName must not be empty.
1720 * If a subjectAlternativeName extension is given it must not be empty.
1721 * The signatureAlgorithm field and the cert signature must be consistent.
1722 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1723 must not be marked critical.
1724 * The authorityKeyIdentifier must be given for X.509v3 certs
1725 unless they are self-signed.
1726 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1730 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1731 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1735 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1736 used even when parsing explicit parameters, when loading a encoded key
1737 or calling `EC_GROUP_new_from_ecpkparameters()`/
1738 `EC_GROUP_new_from_ecparameters()`.
1739 This prevents bypass of security hardening and performance gains,
1740 especially for curves with specialized EC_METHODs.
1741 By default, if a key encoded with explicit parameters is loaded and later
1742 encoded, the output is still encoded with explicit parameters, even if
1743 internally a "named" EC_GROUP is used for computation.
1747 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1748 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1749 NULL. After this change, only the cofactor parameter can be NULL. It also
1750 does some minimal sanity checks on the passed order.
1755 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1756 An attack is simple, if the first CMS_recipientInfo is valid but the
1757 second CMS_recipientInfo is chosen ciphertext. If the second
1758 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1759 encryption key will be replaced by garbage, and the message cannot be
1760 decoded, but if the RSA decryption fails, the correct encryption key is
1761 used and the recipient will not notice the attack.
1762 As a work around for this potential attack the length of the decrypted
1763 key must be equal to the cipher default key length, in case the
1764 certificate is not given and all recipientInfo are tried out.
1765 The old behaviour can be re-enabled in the CMS code by setting the
1766 CMS_DEBUG_DECRYPT flag.
1770 * Early start up entropy quality from the DEVRANDOM seed source has been
1771 improved for older Linux systems. The RAND subsystem will wait for
1772 /dev/random to be producing output before seeding from /dev/urandom.
1773 The seeded state is stored for future library initialisations using
1774 a system global shared memory segment. The shared memory identifier
1775 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1776 the desired value. The default identifier is 114.
1780 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1781 when primes for RSA keys are computed.
1782 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1783 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1784 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1785 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1786 This avoids possible fingerprinting of newly generated RSA modules.
1790 * Correct the extended master secret constant on EBCDIC systems. Without this
1791 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1792 negotiate EMS will fail. Unfortunately this also means that TLS connections
1793 between EBCDIC systems with this fix, and EBCDIC systems without this
1794 fix will fail if they negotiate EMS.
1798 * Changed the library initialisation so that the config file is now loaded
1799 by default. This was already the case for libssl. It now occurs for both
1800 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1801 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1805 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1806 where the former acts as a replacement for `ERR_put_error()`, and the
1807 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1808 `ERR_raise_data()` adds more flexibility by taking a format string and
1809 an arbitrary number of arguments following it, to be processed with
1814 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1815 to check if a named provider is loaded and available. When called, it
1816 will also activate all fallback providers if such are still present.
1820 * Enforce a minimum DH modulus size of 512 bits.
1824 * Changed DH parameters to generate the order q subgroup instead of 2q.
1825 Previously generated DH parameters are still accepted by DH_check
1826 but DH_generate_key works around that by clearing bit 0 of the
1827 private key for those. This avoids leaking bit 0 of the private key.
1831 * Significantly reduce secure memory usage by the randomness pools.
1835 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1840 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1841 algorithms. An implementation of a key exchange algorithm can be obtained
1842 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1843 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1844 the older EVP_PKEY_derive_init() function. See the man pages for the new
1845 functions for further details.
1849 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1853 * Removed the function names from error messages and deprecated the
1858 * Removed NextStep support and the macro OPENSSL_UNISTD
1862 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1863 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1864 Also removed "export var as function" capability; we do not export
1865 variables, only functions.
1869 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1870 an error and 1 indicating success. In previous versions of OpenSSL this
1871 was a void type. If a key was set longer than the maximum possible this
1876 * Support SM2 signing and verification schemes with X509 certificate.
1880 * Use SHA256 as the default digest for TS query in the `ts` app.
1884 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1888 * Default cipher lists/suites are now available via a function, the
1889 #defines are deprecated.
1893 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1894 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1895 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1899 * Join the directories crypto/x509 and crypto/x509v3
1903 * Added command 'openssl kdf' that uses the EVP_KDF API.
1907 * Added command 'openssl mac' that uses the EVP_MAC API.
1911 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1912 as default directories. Also added the command 'openssl info'
1913 for scripting purposes.
1917 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1922 * Add prediction resistance to the DRBG reseeding process.
1926 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1927 mandated by IEEE Std 1619-2018.
1931 * Added newline escaping functionality to a filename when using openssl dgst.
1932 This output format is to replicate the output format found in the `*sum`
1933 checksum programs. This aims to preserve backward compatibility.
1935 *Matt Eaton, Richard Levitte, and Paul Dale*
1937 * Removed the heartbeat message in DTLS feature, as it has very
1938 little usage and doesn't seem to fulfill a valuable purpose.
1939 The configuration option is now deprecated.
1943 * Changed the output of 'openssl {digestname} < file' to display the
1944 digest name in its output.
1948 * Added a new generic trace API which provides support for enabling
1949 instrumentation through trace output.
1951 *Richard Levitte & Matthias St. Pierre*
1953 * Added build tests for C++. These are generated files that only do one
1954 thing, to include one public OpenSSL head file each. This tests that
1955 the public header files can be usefully included in a C++ application.
1957 This test isn't enabled by default. It can be enabled with the option
1958 'enable-buildtest-c++'.
1962 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1966 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1970 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1974 * Added KMAC to EVP_MAC.
1978 * Added property based algorithm implementation selection framework to
1983 * Added SCA hardening for modular field inversion in EC_GROUP through
1984 a new dedicated field_inv() pointer in EC_METHOD.
1985 This also addresses a leakage affecting conversions from projective
1986 to affine coordinates.
1988 *Billy Bob Brumley, Nicola Tuveri*
1990 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1991 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1992 those algorithms that were already supported through the EVP_PKEY API
1993 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1994 and scrypt are now wrappers that call EVP_KDF.
1998 * Build devcrypto engine as a dynamic engine.
2000 *Eneas U de Queiroz*
2002 * Add keyed BLAKE2 to EVP_MAC.
2006 * Fix a bug in the computation of the endpoint-pair shared secret used
2007 by DTLS over SCTP. This breaks interoperability with older versions
2008 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2009 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2010 interoperability with such broken implementations. However, enabling
2011 this switch breaks interoperability with correct implementations.
2013 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2014 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2018 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2022 * Changed the license to the Apache License v2.0.
2026 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2028 - Major releases (indicated by incrementing the MAJOR release number)
2029 may introduce incompatible API/ABI changes.
2030 - Minor releases (indicated by incrementing the MINOR release number)
2031 may introduce new features but retain API/ABI compatibility.
2032 - Patch releases (indicated by incrementing the PATCH number)
2033 are intended for bug fixes and other improvements of existing
2034 features only (like improving performance or adding documentation)
2035 and retain API/ABI compatibility.
2039 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2043 * Remove the 'dist' target and add a tarball building script. The
2044 'dist' target has fallen out of use, and it shouldn't be
2045 necessary to configure just to create a source distribution.
2049 * Recreate the OS390-Unix config target. It no longer relies on a
2050 special script like it did for OpenSSL pre-1.1.0.
2054 * Instead of having the source directories listed in Configure, add
2055 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2060 * Add GMAC to EVP_MAC.
2064 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2068 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2069 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2070 to facilitate the continued use of MACs through raw private keys in
2071 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2075 * Deprecate ECDH_KDF_X9_62().
2079 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2080 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2081 are retained for backwards compatibility.
2085 * AES-XTS mode now enforces that its two keys are different to mitigate
2086 the attacked described in "Efficient Instantiations of Tweakable
2087 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2088 Details of this attack can be obtained from:
2089 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2093 * Rename the object files, i.e. give them other names than in previous
2094 versions. Their names now include the name of the final product, as
2095 well as its type mnemonic (bin, lib, shlib).
2099 * Added new option for 'openssl list', '-objects', which will display the
2100 list of built in objects, i.e. OIDs with names.
2104 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2105 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2110 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2111 improves application performance by removing data copies and providing
2112 applications with zero-copy system calls such as sendfile and splice.
2116 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2120 * `PKCS12_parse` now maintains the order of the parsed certificates
2121 when outputting them via `*ca` (rather than reversing it).
2125 * Deprecated pthread fork support methods.
2129 * Added support for FFDHE key exchange in TLS 1.3.
2133 * Added a new concept for OpenSSL plugability: providers. This
2134 functionality is designed to replace the ENGINE API and ENGINE
2135 implementations, and to be much more dynamic, allowing provider
2136 authors to introduce new algorithms among other things, as long as
2137 there's an API that supports the algorithm type.
2139 With this concept comes a new core API for interaction between
2140 libcrypto and provider implementations. Public libcrypto functions
2141 that want to use providers do so through this core API.
2143 The main documentation for this core API is found in
2144 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2145 refer to other manuals describing the API specific for supported
2146 algorithm types (also called operations).
2153 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2155 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2157 * Avoid loading of a dynamic engine twice.
2161 * Prioritise DANE TLSA issuer certs over peer certs
2165 * Fixed random API for MacOS prior to 10.12
2167 These MacOS versions don't support the CommonCrypto APIs
2171 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2173 * Fixed an SM2 Decryption Buffer Overflow.
2175 In order to decrypt SM2 encrypted data an application is expected to
2176 call the API function EVP_PKEY_decrypt(). Typically an application will
2177 call this function twice. The first time, on entry, the "out" parameter
2178 can be NULL and, on exit, the "outlen" parameter is populated with the
2179 buffer size required to hold the decrypted plaintext. The application
2180 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2181 again, but this time passing a non-NULL value for the "out" parameter.
2183 A bug in the implementation of the SM2 decryption code means that the
2184 calculation of the buffer size required to hold the plaintext returned
2185 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2186 size required by the second call. This can lead to a buffer overflow
2187 when EVP_PKEY_decrypt() is called by the application a second time with
2188 a buffer that is too small.
2190 A malicious attacker who is able present SM2 content for decryption to
2191 an application could cause attacker chosen data to overflow the buffer
2192 by up to a maximum of 62 bytes altering the contents of other data held
2193 after the buffer, possibly changing application behaviour or causing
2194 the application to crash. The location of the buffer is application
2195 dependent but is typically heap allocated.
2200 * Fixed various read buffer overruns processing ASN.1 strings
2202 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2203 structure which contains a buffer holding the string data and a field
2204 holding the buffer length. This contrasts with normal C strings which
2205 are represented as a buffer for the string data which is terminated
2206 with a NUL (0) byte.
2208 Although not a strict requirement, ASN.1 strings that are parsed using
2209 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2210 well as any string whose value has been set with the ASN1_STRING_set()
2211 function will additionally NUL terminate the byte array in the
2212 ASN1_STRING structure.
2214 However, it is possible for applications to directly construct valid
2215 ASN1_STRING structures which do not NUL terminate the byte array by
2216 directly setting the "data" and "length" fields in the ASN1_STRING
2217 array. This can also happen by using the ASN1_STRING_set0() function.
2219 Numerous OpenSSL functions that print ASN.1 data have been found to
2220 assume that the ASN1_STRING byte array will be NUL terminated, even
2221 though this is not guaranteed for strings that have been directly
2222 constructed. Where an application requests an ASN.1 structure to be
2223 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2224 been directly constructed by the application without NUL terminating
2225 the "data" field, then a read buffer overrun can occur.
2227 The same thing can also occur during name constraints processing
2228 of certificates (for example if a certificate has been directly
2229 constructed by the application instead of loading it via the OpenSSL
2230 parsing functions, and the certificate contains non NUL terminated
2231 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2232 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2234 If a malicious actor can cause an application to directly construct an
2235 ASN1_STRING and then process it through one of the affected OpenSSL
2236 functions then this issue could be hit. This might result in a crash
2237 (causing a Denial of Service attack). It could also result in the
2238 disclosure of private memory contents (such as private keys, or
2239 sensitive plaintext).
2244 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2246 * Fixed a problem with verifying a certificate chain when using the
2247 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2248 the certificates present in a certificate chain. It is not set by default.
2250 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2251 the chain that have explicitly encoded elliptic curve parameters was added
2252 as an additional strict check.
2254 An error in the implementation of this check meant that the result of a
2255 previous check to confirm that certificates in the chain are valid CA
2256 certificates was overwritten. This effectively bypasses the check
2257 that non-CA certificates must not be able to issue other certificates.
2259 If a "purpose" has been configured then there is a subsequent opportunity
2260 for checks that the certificate is a valid CA. All of the named "purpose"
2261 values implemented in libcrypto perform this check. Therefore, where
2262 a purpose is set the certificate chain will still be rejected even when the
2263 strict flag has been used. A purpose is set by default in libssl client and
2264 server certificate verification routines, but it can be overridden or
2265 removed by an application.
2267 In order to be affected, an application must explicitly set the
2268 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2269 for the certificate verification or, in the case of TLS client or server
2270 applications, override the default purpose.
2275 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2276 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2277 renegotiation ClientHello omits the signature_algorithms extension (where it
2278 was present in the initial ClientHello), but includes a
2279 signature_algorithms_cert extension then a NULL pointer dereference will
2280 result, leading to a crash and a denial of service attack.
2282 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2283 (which is the default configuration). OpenSSL TLS clients are not impacted by
2287 *Peter Kästle and Samuel Sapalski*
2289 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2291 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2292 create a unique hash value based on the issuer and serial number data
2293 contained within an X509 certificate. However, it was failing to correctly
2294 handle any errors that may occur while parsing the issuer field (which might
2295 occur if the issuer field is maliciously constructed). This may subsequently
2296 result in a NULL pointer deref and a crash leading to a potential denial of
2302 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2303 padding mode to correctly check for rollback attacks. This is considered a
2304 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2309 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2310 functions. Previously they could overflow the output length argument in some
2311 cases where the input length is close to the maximum permissible length for
2312 an integer on the platform. In such cases the return value from the function
2313 call would be 1 (indicating success), but the output length value would be
2314 negative. This could cause applications to behave incorrectly or crash.
2319 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2320 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2321 could be exploited in a side channel attack to recover the password. Since
2322 the attack is local host only this is outside of the current OpenSSL
2323 threat model and therefore no CVE is assigned.
2325 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2330 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2332 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2333 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2334 If an attacker can control both items being compared then this could lead
2335 to a possible denial of service attack. OpenSSL itself uses the
2336 GENERAL_NAME_cmp function for two purposes:
2337 1) Comparing CRL distribution point names between an available CRL and a
2338 CRL distribution point embedded in an X509 certificate
2339 2) When verifying that a timestamp response token signer matches the
2340 timestamp authority name (exposed via the API functions
2341 TS_RESP_verify_response and TS_RESP_verify_token)
2346 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2348 * Certificates with explicit curve parameters are now disallowed in
2349 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2353 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2354 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2355 conversely, silently ignore DTLS protocol version bounds when configuring
2356 TLS-based contexts. The commands can be repeated to set bounds of both
2357 types. The same applies with the corresponding "min_protocol" and
2358 "max_protocol" command-line switches, in case some application uses both TLS
2361 SSL_CTX instances that are created for a fixed protocol version (e.g.
2362 TLSv1_server_method()) also silently ignore version bounds. Previously
2363 attempts to apply bounds to these protocol versions would result in an
2364 error. Now only the "version-flexible" SSL_CTX instances are subject to
2365 limits in configuration files in command-line options.
2369 * Handshake now fails if Extended Master Secret extension is dropped
2374 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2376 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2378 * Fixed segmentation fault in SSL_check_chain()
2379 Server or client applications that call the SSL_check_chain() function
2380 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2381 dereference as a result of incorrect handling of the
2382 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2383 or unrecognised signature algorithm is received from the peer. This could
2384 be exploited by a malicious peer in a Denial of Service attack.
2389 * Added AES consttime code for no-asm configurations
2390 an optional constant time support for AES was added
2391 when building openssl for no-asm.
2392 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2393 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2394 At this time this feature is by default disabled.
2395 It will be enabled by default in 3.0.
2399 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2401 * Revert the change of EOF detection while reading in libssl to avoid
2402 regressions in applications depending on the current way of reporting
2403 the EOF. As the existing method is not fully accurate the change to
2404 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2405 branch and will be present in the 3.0 release.
2409 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2410 when primes for RSA keys are computed.
2411 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2412 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2413 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2414 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2415 This avoids possible fingerprinting of newly generated RSA modules.
2419 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2421 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2422 while reading in libssl then we would report an error back to the
2423 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2424 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2425 therefore give a hint as to what went wrong.
2429 * Check that ed25519 and ed448 are allowed by the security level. Previously
2430 signature algorithms not using an MD were not being checked that they were
2431 allowed by the security level.
2435 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2436 was not quite right. The behaviour was not consistent between resumption
2437 and normal handshakes, and also not quite consistent with historical
2438 behaviour. The behaviour in various scenarios has been clarified and
2439 it has been updated to make it match historical behaviour as closely as
2444 * *[VMS only]* The header files that the VMS compilers include automatically,
2445 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2446 that the C++ compiler doesn't understand. This is a shortcoming in the
2447 compiler, but can be worked around with `__cplusplus` guards.
2449 C++ applications that use OpenSSL libraries must be compiled using the
2450 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2451 functions. Otherwise, only functions with symbols of less than 31
2452 characters can be used, as the linker will not be able to successfully
2453 resolve symbols with longer names.
2457 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2458 The presence of this system service is determined at run-time.
2462 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2467 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2469 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2470 number generator (RNG). This was intended to include protection in the
2471 event of a fork() system call in order to ensure that the parent and child
2472 processes did not share the same RNG state. However, this protection was not
2473 being used in the default case.
2475 A partial mitigation for this issue is that the output from a high
2476 precision timer is mixed into the RNG state so the likelihood of a parent
2477 and child process sharing state is significantly reduced.
2479 If an application already calls OPENSSL_init_crypto() explicitly using
2480 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2483 *Matthias St. Pierre*
2485 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2486 used even when parsing explicit parameters, when loading a encoded key
2487 or calling `EC_GROUP_new_from_ecpkparameters()`/
2488 `EC_GROUP_new_from_ecparameters()`.
2489 This prevents bypass of security hardening and performance gains,
2490 especially for curves with specialized EC_METHODs.
2491 By default, if a key encoded with explicit parameters is loaded and later
2492 encoded, the output is still encoded with explicit parameters, even if
2493 internally a "named" EC_GROUP is used for computation.
2497 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2498 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2499 NULL. After this change, only the cofactor parameter can be NULL. It also
2500 does some minimal sanity checks on the passed order.
2505 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2506 An attack is simple, if the first CMS_recipientInfo is valid but the
2507 second CMS_recipientInfo is chosen ciphertext. If the second
2508 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2509 encryption key will be replaced by garbage, and the message cannot be
2510 decoded, but if the RSA decryption fails, the correct encryption key is
2511 used and the recipient will not notice the attack.
2512 As a work around for this potential attack the length of the decrypted
2513 key must be equal to the cipher default key length, in case the
2514 certificate is not given and all recipientInfo are tried out.
2515 The old behaviour can be re-enabled in the CMS code by setting the
2516 CMS_DEBUG_DECRYPT flag.
2521 * Early start up entropy quality from the DEVRANDOM seed source has been
2522 improved for older Linux systems. The RAND subsystem will wait for
2523 /dev/random to be producing output before seeding from /dev/urandom.
2524 The seeded state is stored for future library initialisations using
2525 a system global shared memory segment. The shared memory identifier
2526 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2527 the desired value. The default identifier is 114.
2531 * Correct the extended master secret constant on EBCDIC systems. Without this
2532 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2533 negotiate EMS will fail. Unfortunately this also means that TLS connections
2534 between EBCDIC systems with this fix, and EBCDIC systems without this
2535 fix will fail if they negotiate EMS.
2539 * Use Windows installation paths in the mingw builds
2541 Mingw isn't a POSIX environment per se, which means that Windows
2542 paths should be used for installation.
2547 * Changed DH_check to accept parameters with order q and 2q subgroups.
2548 With order 2q subgroups the bit 0 of the private key is not secret
2549 but DH_generate_key works around that by clearing bit 0 of the
2550 private key for those. This avoids leaking bit 0 of the private key.
2554 * Significantly reduce secure memory usage by the randomness pools.
2558 * Revert the DEVRANDOM_WAIT feature for Linux systems
2560 The DEVRANDOM_WAIT feature added a select() call to wait for the
2561 /dev/random device to become readable before reading from the
2562 /dev/urandom device.
2564 It turned out that this change had negative side effects on
2565 performance which were not acceptable. After some discussion it
2566 was decided to revert this feature and leave it up to the OS
2567 resp. the platform maintainer to ensure a proper initialization
2568 during early boot time.
2570 *Matthias St. Pierre*
2572 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2574 * Add build tests for C++. These are generated files that only do one
2575 thing, to include one public OpenSSL head file each. This tests that
2576 the public header files can be usefully included in a C++ application.
2578 This test isn't enabled by default. It can be enabled with the option
2579 'enable-buildtest-c++'.
2583 * Enable SHA3 pre-hashing for ECDSA and DSA.
2587 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2588 This changes the size when using the `genpkey` command when no size is given.
2589 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2590 generation commands to use 2048 bits by default.
2594 * Reorganize the manual pages to consistently have RETURN VALUES,
2595 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2596 util/fix-doc-nits accordingly.
2598 *Paul Yang, Joshua Lock*
2600 * Add the missing accessor EVP_PKEY_get0_engine()
2604 * Have commands like `s_client` and `s_server` output the signature scheme
2605 along with other cipher suite parameters when debugging.
2609 * Make OPENSSL_config() error agnostic again.
2613 * Do the error handling in RSA decryption constant time.
2617 * Prevent over long nonces in ChaCha20-Poly1305.
2619 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2620 for every encryption operation. RFC 7539 specifies that the nonce value
2621 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2622 and front pads the nonce with 0 bytes if it is less than 12
2623 bytes. However it also incorrectly allows a nonce to be set of up to 16
2624 bytes. In this case only the last 12 bytes are significant and any
2625 additional leading bytes are ignored.
2627 It is a requirement of using this cipher that nonce values are
2628 unique. Messages encrypted using a reused nonce value are susceptible to
2629 serious confidentiality and integrity attacks. If an application changes
2630 the default nonce length to be longer than 12 bytes and then makes a
2631 change to the leading bytes of the nonce expecting the new value to be a
2632 new unique nonce then such an application could inadvertently encrypt
2633 messages with a reused nonce.
2635 Additionally the ignored bytes in a long nonce are not covered by the
2636 integrity guarantee of this cipher. Any application that relies on the
2637 integrity of these ignored leading bytes of a long nonce may be further
2638 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2639 is safe because no such use sets such a long nonce value. However user
2640 applications that use this cipher directly and set a non-default nonce
2641 length to be longer than 12 bytes may be vulnerable.
2643 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2649 * Add DEVRANDOM_WAIT feature for Linux systems
2651 On older Linux systems where the getrandom() system call is not available,
2652 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2653 Contrary to getrandom(), the /dev/urandom device will not block during
2654 early boot when the kernel CSPRNG has not been seeded yet.
2656 To mitigate this known weakness, use select() to wait for /dev/random to
2657 become readable before reading from /dev/urandom.
2659 * Ensure that SM2 only uses SM3 as digest algorithm
2663 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2665 * Change the info callback signals for the start and end of a post-handshake
2666 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2667 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2668 confused by this and assume that a TLSv1.2 renegotiation has started. This
2669 can break KeyUpdate handling. Instead we no longer signal the start and end
2670 of a post handshake message exchange (although the messages themselves are
2671 still signalled). This could break some applications that were expecting
2672 the old signals. However without this KeyUpdate is not usable for many
2677 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2679 * Timing vulnerability in DSA signature generation
2681 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2682 timing side channel attack. An attacker could use variations in the signing
2683 algorithm to recover the private key.
2685 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2690 * Timing vulnerability in ECDSA signature generation
2692 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2693 timing side channel attack. An attacker could use variations in the signing
2694 algorithm to recover the private key.
2696 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2701 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2702 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2703 of two gigabytes and the error handling improved.
2705 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2706 categorized as a normal bug, not a security issue, because the DRBG reseeds
2707 automatically and is fully functional even without additional randomness
2708 provided by the application.
2710 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2712 * Add a new ClientHello callback. Provides a callback interface that gives
2713 the application the ability to adjust the nascent SSL object at the
2714 earliest stage of ClientHello processing, immediately after extensions have
2715 been collected but before they have been processed. In particular, this
2716 callback can adjust the supported TLS versions in response to the contents
2721 * Add SM2 base algorithm support.
2725 * s390x assembly pack: add (improved) hardware-support for the following
2726 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2727 aes-cfb/cfb8, aes-ecb.
2731 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2732 parameter is no longer accepted, as it leads to a corrupt table. NULL
2733 pem_str is reserved for alias entries only.
2737 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2738 step for prime curves. The new implementation is based on formulae from
2739 differential addition-and-doubling in homogeneous projective coordinates
2740 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2741 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2742 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2743 to work in projective coordinates.
2745 *Billy Bob Brumley, Nicola Tuveri*
2747 * Change generating and checking of primes so that the error rate of not
2748 being prime depends on the intended use based on the size of the input.
2749 For larger primes this will result in more rounds of Miller-Rabin.
2750 The maximal error rate for primes with more than 1080 bits is lowered
2753 *Kurt Roeckx, Annie Yousar*
2755 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2759 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2760 moving between systems, and to avoid confusion when a Windows build is
2761 done with mingw vs with MSVC. For POSIX installs, there's still a
2762 symlink or copy named 'tsget' to avoid that confusion as well.
2766 * Revert blinding in ECDSA sign and instead make problematic addition
2767 length-invariant. Switch even to fixed-length Montgomery multiplication.
2771 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2772 step for binary curves. The new implementation is based on formulae from
2773 differential addition-and-doubling in mixed Lopez-Dahab projective
2774 coordinates, modified to independently blind the operands.
2776 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2778 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2779 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2780 EC_METHODs to implement their own specialized "ladder step", to take
2781 advantage of more favorable coordinate systems or more efficient
2782 differential addition-and-doubling algorithms.
2784 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2786 * Modified the random device based seed sources to keep the relevant
2787 file descriptors open rather than reopening them on each access.
2788 This allows such sources to operate in a chroot() jail without
2789 the associated device nodes being available. This behaviour can be
2790 controlled using RAND_keep_random_devices_open().
2794 * Numerous side-channel attack mitigations have been applied. This may have
2795 performance impacts for some algorithms for the benefit of improved
2796 security. Specific changes are noted in this change log by their respective
2801 * AIX shared library support overhaul. Switch to AIX "natural" way of
2802 handling shared libraries, which means collecting shared objects of
2803 different versions and bitnesses in one common archive. This allows to
2804 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2805 doesn't affect the way 3rd party applications are linked, only how
2806 multi-version installation is managed.
2810 * Make ec_group_do_inverse_ord() more robust and available to other
2811 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2812 mitigations are applied to the fallback BN_mod_inverse().
2813 When using this function rather than BN_mod_inverse() directly, new
2814 EC cryptosystem implementations are then safer-by-default.
2818 * Add coordinate blinding for EC_POINT and implement projective
2819 coordinate blinding for generic prime curves as a countermeasure to
2820 chosen point SCA attacks.
2822 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2824 * Add blinding to ECDSA and DSA signatures to protect against side channel
2825 attacks discovered by Keegan Ryan (NCC Group).
2829 * Enforce checking in the `pkeyutl` command to ensure that the input
2830 length does not exceed the maximum supported digest length when performing
2831 a sign, verify or verifyrecover operation.
2835 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2836 I/O in combination with something like select() or poll() will hang. This
2837 can be turned off again using SSL_CTX_clear_mode().
2838 Many applications do not properly handle non-application data records, and
2839 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2840 around the problems in those applications, but can also break some.
2841 It's recommended to read the manpages about SSL_read(), SSL_write(),
2842 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2843 SSL_CTX_set_read_ahead() again.
2847 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2848 now allow empty (zero character) pass phrases.
2852 * Apply blinding to binary field modular inversion and remove patent
2853 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2857 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2858 binary and prime elliptic curves.
2862 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2863 constant time fixed point multiplication.
2867 * Revise elliptic curve scalar multiplication with timing attack
2868 defenses: ec_wNAF_mul redirects to a constant time implementation
2869 when computing fixed point and variable point multiplication (which
2870 in OpenSSL are mostly used with secret scalars in keygen, sign,
2871 ECDH derive operations).
2872 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2875 * Updated CONTRIBUTING
2879 * Updated DRBG / RAND to request nonce and additional low entropy
2880 randomness from the system.
2882 *Matthias St. Pierre*
2884 * Updated 'openssl rehash' to use OpenSSL consistent default.
2888 * Moved the load of the ssl_conf module to libcrypto, which helps
2889 loading engines that libssl uses before libssl is initialised.
2893 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2897 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2899 *Ingo Schwarze, Rich Salz*
2901 * Added output of accepting IP address and port for 'openssl s_server'
2905 * Added a new API for TLSv1.3 ciphersuites:
2906 SSL_CTX_set_ciphersuites()
2907 SSL_set_ciphersuites()
2911 * Memory allocation failures consistently add an error to the error
2916 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2917 in libcrypto when run as setuid/setgid.
2921 * Load any config file by default when libssl is used.
2925 * Added new public header file <openssl/rand_drbg.h> and documentation
2926 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2928 *Matthias St. Pierre*
2930 * QNX support removed (cannot find contributors to get their approval
2931 for the license change).
2935 * TLSv1.3 replay protection for early data has been implemented. See the
2936 SSL_read_early_data() man page for further details.
2940 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2941 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2942 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2943 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2944 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2945 configuration has been separated out. See the ciphers man page or the
2946 SSL_CTX_set_ciphersuites() man page for more information.
2950 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2951 in responder mode now supports the new "-multi" option, which
2952 spawns the specified number of child processes to handle OCSP
2953 requests. The "-timeout" option now also limits the OCSP
2954 responder's patience to wait to receive the full client request
2955 on a newly accepted connection. Child processes are respawned
2956 as needed, and the CA index file is automatically reloaded
2957 when changed. This makes it possible to run the "ocsp" responder
2958 as a long-running service, making the OpenSSL CA somewhat more
2959 feature-complete. In this mode, most diagnostic messages logged
2960 after entering the event loop are logged via syslog(3) rather than
2965 * Added support for X448 and Ed448. Heavily based on original work by
2970 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2971 objects loaded. This adds the functions OSSL_STORE_expect() and
2972 OSSL_STORE_find() as well as needed tools to construct searches and
2973 get the search data out of them.
2977 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2978 version of OpenSSL should review their configuration settings to ensure
2979 that they are still appropriate for TLSv1.3. For further information see:
2980 <https://wiki.openssl.org/index.php/TLS1.3>
2984 * Grand redesign of the OpenSSL random generator
2986 The default RAND method now utilizes an AES-CTR DRBG according to
2987 NIST standard SP 800-90Ar1. The new random generator is essentially
2988 a port of the default random generator from the OpenSSL FIPS 2.0
2989 object module. It is a hybrid deterministic random bit generator
2990 using an AES-CTR bit stream and which seeds and reseeds itself
2991 automatically using trusted system entropy sources.
2993 Some of its new features are:
2994 - Support for multiple DRBG instances with seed chaining.
2995 - The default RAND method makes use of a DRBG.
2996 - There is a public and private DRBG instance.
2997 - The DRBG instances are fork-safe.
2998 - Keep all global DRBG instances on the secure heap if it is enabled.
2999 - The public and private DRBG instance are per thread for lock free
3002 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3004 * Changed Configure so it only says what it does and doesn't dump
3005 so much data. Instead, ./configdata.pm should be used as a script
3006 to display all sorts of configuration data.
3010 * Added processing of "make variables" to Configure.
3014 * Added SHA512/224 and SHA512/256 algorithm support.
3018 * The last traces of Netware support, first removed in 1.1.0, have
3023 * Get rid of Makefile.shared, and in the process, make the processing
3024 of certain files (rc.obj, or the .def/.map/.opt files produced from
3025 the ordinal files) more visible and hopefully easier to trace and
3026 debug (or make silent).
3030 * Make it possible to have environment variable assignments as
3031 arguments to config / Configure.
3035 * Add multi-prime RSA (RFC 8017) support.
3039 * Add SM3 implemented according to GB/T 32905-2016
3040 *Jack Lloyd <jack.lloyd@ribose.com>,*
3041 *Ronald Tse <ronald.tse@ribose.com>,*
3042 *Erick Borsboom <erick.borsboom@ribose.com>*
3044 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3045 as documented in RFC6066.
3046 Based on a patch from Tomasz Moń
3048 *Filipe Raimundo da Silva*
3050 * Add SM4 implemented according to GB/T 32907-2016.
3051 *Jack Lloyd <jack.lloyd@ribose.com>,*
3052 *Ronald Tse <ronald.tse@ribose.com>,*
3053 *Erick Borsboom <erick.borsboom@ribose.com>*
3055 * Reimplement -newreq-nodes and ERR_error_string_n; the
3056 original author does not agree with the license change.
3060 * Add ARIA AEAD TLS support.
3064 * Some macro definitions to support VS6 have been removed. Visual
3065 Studio 6 has not worked since 1.1.0
3069 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3070 without clearing the errors.
3074 * Add "atfork" functions. If building on a system that without
3075 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3076 requirements. The RAND facility now uses/requires this.
3084 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3085 not possible to disable entirely. However, it's still possible to
3086 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3089 To disable, configure with 'no-ui-console'. 'no-ui' is still
3090 possible to use as an alias. Check at compile time with the
3091 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3092 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3096 * Add a STORE module, which implements a uniform and URI based reader of
3097 stores that can contain keys, certificates, CRLs and numerous other
3098 objects. The main API is loosely based on a few stdio functions,
3099 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3100 OSSL_STORE_error and OSSL_STORE_close.
3101 The implementation uses backends called "loaders" to implement arbitrary
3102 URI schemes. There is one built in "loader" for the 'file' scheme.
3106 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3107 then adjusted to work on FreeBSD 8.4 as well.
3108 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3109 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3113 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3114 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3115 error code calls like this:
3117 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3119 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3120 that can be encoded in C. For the foreseeable future, this will only
3123 *Richard Levitte and Tim Hudson*
3125 * Removed BSD cryptodev engine.
3129 * Add a build target 'build_all_generated', to build all generated files
3130 and only that. This can be used to prepare everything that requires
3131 things like perl for a system that lacks perl and then move everything
3132 to that system and do the rest of the build there.
3136 * In the UI interface, make it possible to duplicate the user data. This
3137 can be used by engines that need to retain the data for a longer time
3138 than just the call where this user data is passed.
3142 * Ignore the '-named_curve auto' value for compatibility of applications
3145 *Tomáš Mráz <tmraz@fedoraproject.org>*
3147 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3148 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3149 alerts across multiple records (some of which could be empty). In practice
3150 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3151 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3152 support this at all. Supporting it adds significant complexity to the
3153 record layer, and its removal is unlikely to cause interoperability
3158 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3159 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3160 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3165 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3166 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3168 *Richard Levitte, Andy Polyakov*
3170 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3175 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3176 platform rather than 'mingw'.
3180 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3181 success if they are asked to add an object which already exists
3182 in the store. This change cascades to other functions which load
3183 certificates and CRLs.
3187 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3188 facilitate stack unwinding even from assembly subroutines.
3192 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3193 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3197 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3198 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3199 which is the minimum version we support.
3203 * Certificate time validation (X509_cmp_time) enforces stricter
3204 compliance with RFC 5280. Fractional seconds and timezone offsets
3205 are no longer allowed.
3209 * Add support for ARIA
3213 * s_client will now send the Server Name Indication (SNI) extension by
3214 default unless the new "-noservername" option is used. The server name is
3215 based on the host provided to the "-connect" option unless overridden by
3216 using "-servername".
3220 * Add support for SipHash
3224 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3225 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3226 prevent issues where no progress is being made and the peer continually
3227 sends unrecognised record types, using up resources processing them.
3231 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3232 using the algorithm defined in
3233 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3237 * Heartbeat support has been removed; the ABI is changed for now.
3239 *Richard Levitte, Rich Salz*
3241 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3245 * The RSA "null" method, which was partially supported to avoid patent
3246 issues, has been replaced to always returns NULL.
3253 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3255 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3256 used even when parsing explicit parameters, when loading a encoded key
3257 or calling `EC_GROUP_new_from_ecpkparameters()`/
3258 `EC_GROUP_new_from_ecparameters()`.
3259 This prevents bypass of security hardening and performance gains,
3260 especially for curves with specialized EC_METHODs.
3261 By default, if a key encoded with explicit parameters is loaded and later
3262 encoded, the output is still encoded with explicit parameters, even if
3263 internally a "named" EC_GROUP is used for computation.
3267 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3268 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3269 NULL. After this change, only the cofactor parameter can be NULL. It also
3270 does some minimal sanity checks on the passed order.
3275 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3276 An attack is simple, if the first CMS_recipientInfo is valid but the
3277 second CMS_recipientInfo is chosen ciphertext. If the second
3278 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3279 encryption key will be replaced by garbage, and the message cannot be
3280 decoded, but if the RSA decryption fails, the correct encryption key is
3281 used and the recipient will not notice the attack.
3282 As a work around for this potential attack the length of the decrypted
3283 key must be equal to the cipher default key length, in case the
3284 certificate is not given and all recipientInfo are tried out.
3285 The old behaviour can be re-enabled in the CMS code by setting the
3286 CMS_DEBUG_DECRYPT flag.
3291 * Use Windows installation paths in the mingw builds
3293 Mingw isn't a POSIX environment per se, which means that Windows
3294 paths should be used for installation.
3299 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3301 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3302 This changes the size when using the `genpkey` command when no size is given.
3303 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3304 generation commands to use 2048 bits by default.
3308 * Prevent over long nonces in ChaCha20-Poly1305.
3310 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3311 for every encryption operation. RFC 7539 specifies that the nonce value
3312 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3313 and front pads the nonce with 0 bytes if it is less than 12
3314 bytes. However it also incorrectly allows a nonce to be set of up to 16
3315 bytes. In this case only the last 12 bytes are significant and any
3316 additional leading bytes are ignored.
3318 It is a requirement of using this cipher that nonce values are
3319 unique. Messages encrypted using a reused nonce value are susceptible to
3320 serious confidentiality and integrity attacks. If an application changes
3321 the default nonce length to be longer than 12 bytes and then makes a
3322 change to the leading bytes of the nonce expecting the new value to be a
3323 new unique nonce then such an application could inadvertently encrypt
3324 messages with a reused nonce.
3326 Additionally the ignored bytes in a long nonce are not covered by the
3327 integrity guarantee of this cipher. Any application that relies on the
3328 integrity of these ignored leading bytes of a long nonce may be further
3329 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3330 is safe because no such use sets such a long nonce value. However user
3331 applications that use this cipher directly and set a non-default nonce
3332 length to be longer than 12 bytes may be vulnerable.
3334 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3340 * Added SCA hardening for modular field inversion in EC_GROUP through
3341 a new dedicated field_inv() pointer in EC_METHOD.
3342 This also addresses a leakage affecting conversions from projective
3343 to affine coordinates.
3345 *Billy Bob Brumley, Nicola Tuveri*
3347 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3348 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3352 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3356 * Remove the 'dist' target and add a tarball building script. The
3357 'dist' target has fallen out of use, and it shouldn't be
3358 necessary to configure just to create a source distribution.
3362 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3364 * Timing vulnerability in DSA signature generation
3366 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3367 timing side channel attack. An attacker could use variations in the signing
3368 algorithm to recover the private key.
3370 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3375 * Timing vulnerability in ECDSA signature generation
3377 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3378 timing side channel attack. An attacker could use variations in the signing
3379 algorithm to recover the private key.
3381 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3386 * Add coordinate blinding for EC_POINT and implement projective
3387 coordinate blinding for generic prime curves as a countermeasure to
3388 chosen point SCA attacks.
3390 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3392 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3394 * Client DoS due to large DH parameter
3396 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3397 malicious server can send a very large prime value to the client. This will
3398 cause the client to spend an unreasonably long period of time generating a
3399 key for this prime resulting in a hang until the client has finished. This
3400 could be exploited in a Denial Of Service attack.
3402 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3407 * Cache timing vulnerability in RSA Key Generation
3409 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3410 a cache timing side channel attack. An attacker with sufficient access to
3411 mount cache timing attacks during the RSA key generation process could
3412 recover the private key.
3414 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3415 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3420 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3421 parameter is no longer accepted, as it leads to a corrupt table. NULL
3422 pem_str is reserved for alias entries only.
3426 * Revert blinding in ECDSA sign and instead make problematic addition
3427 length-invariant. Switch even to fixed-length Montgomery multiplication.
3431 * Change generating and checking of primes so that the error rate of not
3432 being prime depends on the intended use based on the size of the input.
3433 For larger primes this will result in more rounds of Miller-Rabin.
3434 The maximal error rate for primes with more than 1080 bits is lowered
3437 *Kurt Roeckx, Annie Yousar*
3439 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3443 * Add blinding to ECDSA and DSA signatures to protect against side channel
3444 attacks discovered by Keegan Ryan (NCC Group).
3448 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3449 now allow empty (zero character) pass phrases.
3453 * Certificate time validation (X509_cmp_time) enforces stricter
3454 compliance with RFC 5280. Fractional seconds and timezone offsets
3455 are no longer allowed.
3459 * Fixed a text canonicalisation bug in CMS
3461 Where a CMS detached signature is used with text content the text goes
3462 through a canonicalisation process first prior to signing or verifying a
3463 signature. This process strips trailing space at the end of lines, converts
3464 line terminators to CRLF and removes additional trailing line terminators
3465 at the end of a file. A bug in the canonicalisation process meant that
3466 some characters, such as form-feed, were incorrectly treated as whitespace
3467 and removed. This is contrary to the specification (RFC5485). This fix
3468 could mean that detached text data signed with an earlier version of
3469 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3470 signed with a fixed OpenSSL may fail to verify with an earlier version of
3471 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3472 and use the "-binary" flag (for the "cms" command line application) or set
3473 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3477 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3479 * Constructed ASN.1 types with a recursive definition could exceed the stack
3481 Constructed ASN.1 types with a recursive definition (such as can be found
3482 in PKCS7) could eventually exceed the stack given malicious input with
3483 excessive recursion. This could result in a Denial Of Service attack. There
3484 are no such structures used within SSL/TLS that come from untrusted sources
3485 so this is considered safe.
3487 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3493 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3495 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3496 effectively reduced to only comparing the least significant bit of each
3497 byte. This allows an attacker to forge messages that would be considered as
3498 authenticated in an amount of tries lower than that guaranteed by the
3499 security claims of the scheme. The module can only be compiled by the
3500 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3502 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3508 * Add a build target 'build_all_generated', to build all generated files
3509 and only that. This can be used to prepare everything that requires
3510 things like perl for a system that lacks perl and then move everything
3511 to that system and do the rest of the build there.
3515 * Backport SSL_OP_NO_RENGOTIATION
3517 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3518 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3519 changes this is no longer possible in 1.1.0. Therefore, the new
3520 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3521 1.1.0 to provide equivalent functionality.
3523 Note that if an application built against 1.1.0h headers (or above) is run
3524 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3525 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3529 * Removed the OS390-Unix config target. It relied on a script that doesn't
3534 * rsaz_1024_mul_avx2 overflow bug on x86_64
3536 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3537 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3538 Analysis suggests that attacks against RSA and DSA as a result of this
3539 defect would be very difficult to perform and are not believed likely.
3540 Attacks against DH1024 are considered just feasible, because most of the
3541 work necessary to deduce information about a private key may be performed
3542 offline. The amount of resources required for such an attack would be
3543 significant. However, for an attack on TLS to be meaningful, the server
3544 would have to share the DH1024 private key among multiple clients, which is
3545 no longer an option since CVE-2016-0701.
3547 This only affects processors that support the AVX2 but not ADX extensions
3548 like Intel Haswell (4th generation).
3550 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3551 was originally found via the OSS-Fuzz project.
3556 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3558 * bn_sqrx8x_internal carry bug on x86_64
3560 There is a carry propagating bug in the x86_64 Montgomery squaring
3561 procedure. No EC algorithms are affected. Analysis suggests that attacks
3562 against RSA and DSA as a result of this defect would be very difficult to
3563 perform and are not believed likely. Attacks against DH are considered just
3564 feasible (although very difficult) because most of the work necessary to
3565 deduce information about a private key may be performed offline. The amount
3566 of resources required for such an attack would be very significant and
3567 likely only accessible to a limited number of attackers. An attacker would
3568 additionally need online access to an unpatched system using the target
3569 private key in a scenario with persistent DH parameters and a private
3570 key that is shared between multiple clients.
3572 This only affects processors that support the BMI1, BMI2 and ADX extensions
3573 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3575 This issue was reported to OpenSSL by the OSS-Fuzz project.
3580 * Malformed X.509 IPAddressFamily could cause OOB read
3582 If an X.509 certificate has a malformed IPAddressFamily extension,
3583 OpenSSL could do a one-byte buffer overread. The most likely result
3584 would be an erroneous display of the certificate in text format.
3586 This issue was reported to OpenSSL by the OSS-Fuzz project.
3591 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3593 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3594 platform rather than 'mingw'.
3598 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3599 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3600 which is the minimum version we support.
3604 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3606 * Encrypt-Then-Mac renegotiation crash
3608 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3609 negotiated where it was not in the original handshake (or vice-versa) then
3610 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3611 and servers are affected.
3613 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3618 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3620 * Truncated packet could crash via OOB read
3622 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3623 cipher is being used, then a truncated packet can cause that host to
3624 perform an out-of-bounds read, usually resulting in a crash.
3626 This issue was reported to OpenSSL by Robert Święcki of Google.
3631 * Bad (EC)DHE parameters cause a client crash
3633 If a malicious server supplies bad parameters for a DHE or ECDHE key
3634 exchange then this can result in the client attempting to dereference a
3635 NULL pointer leading to a client crash. This could be exploited in a Denial
3638 This issue was reported to OpenSSL by Guido Vranken.
3643 * BN_mod_exp may produce incorrect results on x86_64
3645 There is a carry propagating bug in the x86_64 Montgomery squaring
3646 procedure. No EC algorithms are affected. Analysis suggests that attacks
3647 against RSA and DSA as a result of this defect would be very difficult to
3648 perform and are not believed likely. Attacks against DH are considered just
3649 feasible (although very difficult) because most of the work necessary to
3650 deduce information about a private key may be performed offline. The amount
3651 of resources required for such an attack would be very significant and
3652 likely only accessible to a limited number of attackers. An attacker would
3653 additionally need online access to an unpatched system using the target
3654 private key in a scenario with persistent DH parameters and a private
3655 key that is shared between multiple clients. For example this can occur by
3656 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3657 similar to CVE-2015-3193 but must be treated as a separate problem.
3659 This issue was reported to OpenSSL by the OSS-Fuzz project.
3664 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3666 * ChaCha20/Poly1305 heap-buffer-overflow
3668 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3669 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3670 crash. This issue is not considered to be exploitable beyond a DoS.
3672 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3677 * CMS Null dereference
3679 Applications parsing invalid CMS structures can crash with a NULL pointer
3680 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3681 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3682 structure callback if an attempt is made to free certain invalid encodings.
3683 Only CHOICE structures using a callback which do not handle NULL value are
3686 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3691 * Montgomery multiplication may produce incorrect results
3693 There is a carry propagating bug in the Broadwell-specific Montgomery
3694 multiplication procedure that handles input lengths divisible by, but
3695 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3696 and DH private keys are impossible. This is because the subroutine in
3697 question is not used in operations with the private key itself and an input
3698 of the attacker's direct choice. Otherwise the bug can manifest itself as
3699 transient authentication and key negotiation failures or reproducible
3700 erroneous outcome of public-key operations with specially crafted input.
3701 Among EC algorithms only Brainpool P-512 curves are affected and one
3702 presumably can attack ECDH key negotiation. Impact was not analyzed in
3703 detail, because pre-requisites for attack are considered unlikely. Namely
3704 multiple clients have to choose the curve in question and the server has to
3705 share the private key among them, neither of which is default behaviour.
3706 Even then only clients that chose the curve will be affected.
3708 This issue was publicly reported as transient failures and was not
3709 initially recognized as a security issue. Thanks to Richard Morgan for
3710 providing reproducible case.
3715 * Removed automatic addition of RPATH in shared libraries and executables,
3716 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3720 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3722 * Fix Use After Free for large message sizes
3724 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3725 message larger than approx 16k is received then the underlying buffer to
3726 store the incoming message is reallocated and moved. Unfortunately a
3727 dangling pointer to the old location is left which results in an attempt to
3728 write to the previously freed location. This is likely to result in a
3729 crash, however it could potentially lead to execution of arbitrary code.
3731 This issue only affects OpenSSL 1.1.0a.
3733 This issue was reported to OpenSSL by Robert Święcki.
3738 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3740 * OCSP Status Request extension unbounded memory growth
3742 A malicious client can send an excessively large OCSP Status Request
3743 extension. If that client continually requests renegotiation, sending a
3744 large OCSP Status Request extension each time, then there will be unbounded
3745 memory growth on the server. This will eventually lead to a Denial Of
3746 Service attack through memory exhaustion. Servers with a default
3747 configuration are vulnerable even if they do not support OCSP. Builds using
3748 the "no-ocsp" build time option are not affected.
3750 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3755 * SSL_peek() hang on empty record
3757 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3758 sends an empty record. This could be exploited by a malicious peer in a
3759 Denial Of Service attack.
3761 This issue was reported to OpenSSL by Alex Gaynor.
3766 * Excessive allocation of memory in tls_get_message_header() and
3767 dtls1_preprocess_fragment()
3769 A (D)TLS message includes 3 bytes for its length in the header for the
3770 message. This would allow for messages up to 16Mb in length. Messages of
3771 this length are excessive and OpenSSL includes a check to ensure that a
3772 peer is sending reasonably sized messages in order to avoid too much memory
3773 being consumed to service a connection. A flaw in the logic of version
3774 1.1.0 means that memory for the message is allocated too early, prior to
3775 the excessive message length check. Due to way memory is allocated in
3776 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3777 to service a connection. This could lead to a Denial of Service through
3778 memory exhaustion. However, the excessive message length check still takes
3779 place, and this would cause the connection to immediately fail. Assuming
3780 that the application calls SSL_free() on the failed connection in a timely
3781 manner then the 21Mb of allocated memory will then be immediately freed
3782 again. Therefore, the excessive memory allocation will be transitory in
3783 nature. This then means that there is only a security impact if:
3785 1) The application does not call SSL_free() in a timely manner in the event
3786 that the connection fails
3788 2) The application is working in a constrained environment where there is
3789 very little free memory
3791 3) The attacker initiates multiple connection attempts such that there are
3792 multiple connections in a state where memory has been allocated for the
3793 connection; SSL_free() has not yet been called; and there is insufficient
3794 memory to service the multiple requests.
3796 Except in the instance of (1) above any Denial Of Service is likely to be
3797 transitory because as soon as the connection fails the memory is
3798 subsequently freed again in the SSL_free() call. However there is an
3799 increased risk during this period of application crashes due to the lack of
3800 memory - which would then mean a more serious Denial of Service.
3802 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3803 (CVE-2016-6307 and CVE-2016-6308)
3807 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3808 had to be removed. Primary reason is that vendor assembler can't
3809 assemble our modules with -KPIC flag. As result it, assembly
3810 support, was not even available as option. But its lack means
3811 lack of side-channel resistant code, which is incompatible with
3812 security by todays standards. Fortunately gcc is readily available
3813 prepackaged option, which we firmly point at...
3817 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3819 * Windows command-line tool supports UTF-8 opt-in option for arguments
3820 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3821 (to any value) allows Windows user to access PKCS#12 file generated
3822 with Windows CryptoAPI and protected with non-ASCII password, as well
3823 as files generated under UTF-8 locale on Linux also protected with
3828 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3829 have been disabled by default and removed from DEFAULT, just like RC4.
3830 See the RC4 item below to re-enable both.
3834 * The method for finding the storage location for the Windows RAND seed file
3835 has changed. First we check %RANDFILE%. If that is not set then we check
3836 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3837 all else fails we fall back to C:\.
3841 * The EVP_EncryptUpdate() function has had its return type changed from void
3842 to int. A return of 0 indicates and error while a return of 1 indicates
3847 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3848 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3849 off the constant time implementation for RSA, DSA and DH have been made
3850 no-ops and deprecated.
3854 * Windows RAND implementation was simplified to only get entropy by
3855 calling CryptGenRandom(). Various other RAND-related tickets
3858 *Joseph Wylie Yandle, Rich Salz*
3860 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3861 and `OPENSSL_LH_`, respectively. The old names are available
3862 with API compatibility. They new names are now completely documented.
3866 * Unify TYPE_up_ref(obj) methods signature.
3867 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3868 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3869 int (instead of void) like all others TYPE_up_ref() methods.
3870 So now these methods also check the return value of CRYPTO_atomic_add(),
3871 and the validity of object reference counter.
3873 *fdasilvayy@gmail.com*
3875 * With Windows Visual Studio builds, the .pdb files are installed
3876 alongside the installed libraries and executables. For a static
3877 library installation, ossl_static.pdb is the associate compiler
3878 generated .pdb file to be used when linking programs.
3882 * Remove openssl.spec. Packaging files belong with the packagers.
3886 * Automatic Darwin/OSX configuration has had a refresh, it will now
3887 recognise x86_64 architectures automatically. You can still decide
3888 to build for a different bitness with the environment variable
3889 KERNEL_BITS (can be 32 or 64), for example:
3891 KERNEL_BITS=32 ./config
3895 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3896 256 bit AES and HMAC with SHA256.
3900 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3904 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3908 * To enable users to have their own config files and build file templates,
3909 Configure looks in the directory indicated by the environment variable
3910 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3911 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3912 name and is used as is.
3916 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3917 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3918 X509_CERT_FILE_CTX was removed.
3922 * "shared" builds are now the default. To create only static libraries use
3923 the "no-shared" Configure option.
3927 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3928 All of these option have not worked for some while and are fundamental
3933 * Make various cleanup routines no-ops and mark them as deprecated. Most
3934 global cleanup functions are no longer required because they are handled
3935 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3936 Explicitly de-initing can cause problems (e.g. where a library that uses
3937 OpenSSL de-inits, but an application is still using it). The affected
3938 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3939 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3940 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3941 COMP_zlib_cleanup().
3945 * --strict-warnings no longer enables runtime debugging options
3946 such as REF_DEBUG. Instead, debug options are automatically
3947 enabled with '--debug' builds.
3949 *Andy Polyakov, Emilia Käsper*
3951 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3952 have been moved out of the public header files. New functions for managing
3953 these have been added.
3957 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3958 objects have been moved out of the public header files. New
3959 functions for managing these have been added.
3963 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3964 have been moved out of the public header files. New functions for managing
3965 these have been added.
3969 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3970 moved out of the public header files. New functions for managing these
3975 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3979 * Removed the mk1mf build scripts.
3983 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3984 it is always safe to #include a header now.
3988 * Removed the aged BC-32 config and all its supporting scripts
3992 * Removed support for Ultrix, Netware, and OS/2.
3996 * Add support for HKDF.
3998 *Alessandro Ghedini*
4000 * Add support for blake2b and blake2s
4004 * Added support for "pipelining". Ciphers that have the
4005 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4006 encryptions/decryptions simultaneously. There are currently no built-in
4007 ciphers with this property but the expectation is that engines will be able
4008 to offer it to significantly improve throughput. Support has been extended
4009 into libssl so that multiple records for a single connection can be
4010 processed in one go (for >=TLS 1.1).
4014 * Added the AFALG engine. This is an async capable engine which is able to
4015 offload work to the Linux kernel. In this initial version it only supports
4016 AES128-CBC. The kernel must be version 4.1.0 or greater.
4020 * OpenSSL now uses a new threading API. It is no longer necessary to
4021 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4022 are two supported threading models: pthreads and windows threads. It is
4023 also possible to configure OpenSSL at compile time for "no-threads". The
4024 old threading API should no longer be used. The functions have been
4025 replaced with "no-op" compatibility macros.
4027 *Alessandro Ghedini, Matt Caswell*
4029 * Modify behavior of ALPN to invoke callback after SNI/servername
4030 callback, such that updates to the SSL_CTX affect ALPN.
4034 * Add SSL_CIPHER queries for authentication and key-exchange.
4038 * Changes to the DEFAULT cipherlist:
4039 - Prefer (EC)DHE handshakes over plain RSA.
4040 - Prefer AEAD ciphers over legacy ciphers.
4041 - Prefer ECDSA over RSA when both certificates are available.
4042 - Prefer TLSv1.2 ciphers/PRF.
4043 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4048 * Change the ECC default curve list to be this, in order: x25519,
4049 secp256r1, secp521r1, secp384r1.
4053 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4054 disabled by default. They can be re-enabled using the
4055 enable-weak-ssl-ciphers option to Configure.
4059 * If the server has ALPN configured, but supports no protocols that the
4060 client advertises, send a fatal "no_application_protocol" alert.
4061 This behaviour is SHALL in RFC 7301, though it isn't universally
4062 implemented by other servers.
4066 * Add X25519 support.
4067 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4068 for public and private key encoding using the format documented in
4069 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4070 key generation and key derivation.
4072 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4077 * Deprecate SRP_VBASE_get_by_user.
4078 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4079 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4080 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4081 seed, even if the seed is configured.
4083 Users should use SRP_VBASE_get1_by_user instead. Note that in
4084 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4085 also that even though configuring the SRP seed attempts to hide
4086 invalid usernames by continuing the handshake with fake
4087 credentials, this behaviour is not constant time and no strong
4088 guarantees are made that the handshake is indistinguishable from
4089 that of a valid user.
4093 * Configuration change; it's now possible to build dynamic engines
4094 without having to build shared libraries and vice versa. This
4095 only applies to the engines in `engines/`, those in `crypto/engine/`
4096 will always be built into libcrypto (i.e. "static").
4098 Building dynamic engines is enabled by default; to disable, use
4099 the configuration option "disable-dynamic-engine".
4101 The only requirements for building dynamic engines are the
4102 presence of the DSO module and building with position independent
4103 code, so they will also automatically be disabled if configuring
4104 with "disable-dso" or "disable-pic".
4106 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4107 are also taken away from openssl/opensslconf.h, as they are
4112 * Configuration change; if there is a known flag to compile
4113 position independent code, it will always be applied on the
4114 libcrypto and libssl object files, and never on the application
4115 object files. This means other libraries that use routines from
4116 libcrypto / libssl can be made into shared libraries regardless
4117 of how OpenSSL was configured.
4119 If this isn't desirable, the configuration options "disable-pic"
4120 or "no-pic" can be used to disable the use of PIC. This will
4121 also disable building shared libraries and dynamic engines.
4125 * Removed JPAKE code. It was experimental and has no wide use.
4129 * The INSTALL_PREFIX Makefile variable has been renamed to
4130 DESTDIR. That makes for less confusion on what this variable
4131 is for. Also, the configuration option --install_prefix is
4136 * Heartbeat for TLS has been removed and is disabled by default
4137 for DTLS; configure with enable-heartbeats. Code that uses the
4138 old #define's might need to be updated.
4140 *Emilia Käsper, Rich Salz*
4142 * Rename REF_CHECK to REF_DEBUG.
4146 * New "unified" build system
4148 The "unified" build system is aimed to be a common system for all
4149 platforms we support. With it comes new support for VMS.
4151 This system builds supports building in a different directory tree
4152 than the source tree. It produces one Makefile (for unix family
4153 or lookalikes), or one descrip.mms (for VMS).
4155 The source of information to make the Makefile / descrip.mms is
4156 small files called 'build.info', holding the necessary
4157 information for each directory with source to compile, and a
4158 template in Configurations, like unix-Makefile.tmpl or
4161 With this change, the library names were also renamed on Windows
4162 and on VMS. They now have names that are closer to the standard
4163 on Unix, and include the major version number, and in certain
4164 cases, the architecture they are built for. See "Notes on shared
4165 libraries" in INSTALL.
4167 We rely heavily on the perl module Text::Template.
4171 * Added support for auto-initialisation and de-initialisation of the library.
4172 OpenSSL no longer requires explicit init or deinit routines to be called,
4173 except in certain circumstances. See the OPENSSL_init_crypto() and
4174 OPENSSL_init_ssl() man pages for further information.
4178 * The arguments to the DTLSv1_listen function have changed. Specifically the
4179 "peer" argument is now expected to be a BIO_ADDR object.
4181 * Rewrite of BIO networking library. The BIO library lacked consistent
4182 support of IPv6, and adding it required some more extensive
4183 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4184 which hold all types of addresses and chains of address information.
4185 It also introduces a new API, with functions like BIO_socket,
4186 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4187 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4188 have been adapted accordingly.
4192 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4197 * CRIME protection: disable compression by default, even if OpenSSL is
4198 compiled with zlib enabled. Applications can still enable compression
4199 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4200 using the SSL_CONF library to configure compression.
4204 * The signature of the session callback configured with
4205 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4206 was explicitly marked as `const unsigned char*` instead of
4211 * Always DPURIFY. Remove the use of uninitialized memory in the
4212 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4216 * Removed many obsolete configuration items, including
4217 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4218 MD2_CHAR, MD2_INT, MD2_LONG
4220 IDEA_SHORT, IDEA_LONG
4221 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4223 *Rich Salz, with advice from Andy Polyakov*
4225 * Many BN internals have been moved to an internal header file.
4227 *Rich Salz with help from Andy Polyakov*
4229 * Configuration and writing out the results from it has changed.
4230 Files such as Makefile include/openssl/opensslconf.h and are now
4231 produced through general templates, such as Makefile.in and
4232 crypto/opensslconf.h.in and some help from the perl module
4235 Also, the center of configuration information is no longer
4236 Makefile. Instead, Configure produces a perl module in
4237 configdata.pm which holds most of the config data (in the hash
4238 table %config), the target data that comes from the target
4239 configuration in one of the `Configurations/*.conf` files (in
4244 * To clarify their intended purposes, the Configure options
4245 --prefix and --openssldir change their semantics, and become more
4246 straightforward and less interdependent.
4248 --prefix shall be used exclusively to give the location INSTALLTOP
4249 where programs, scripts, libraries, include files and manuals are
4250 going to be installed. The default is now /usr/local.
4252 --openssldir shall be used exclusively to give the default
4253 location OPENSSLDIR where certificates, private keys, CRLs are
4254 managed. This is also where the default openssl.cnf gets
4256 If the directory given with this option is a relative path, the
4257 values of both the --prefix value and the --openssldir value will
4258 be combined to become OPENSSLDIR.
4259 The default for --openssldir is INSTALLTOP/ssl.
4261 Anyone who uses --openssldir to specify where OpenSSL is to be
4262 installed MUST change to use --prefix instead.
4266 * The GOST engine was out of date and therefore it has been removed. An up
4267 to date GOST engine is now being maintained in an external repository.
4268 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4269 support for GOST ciphersuites (these are only activated if a GOST engine
4274 * EGD is no longer supported by default; use enable-egd when
4277 *Ben Kaduk and Rich Salz*
4279 * The distribution now has Makefile.in files, which are used to
4280 create Makefile's when Configure is run. *Configure must be run
4281 before trying to build now.*
4285 * The return value for SSL_CIPHER_description() for error conditions
4290 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4292 Obtaining and performing DNSSEC validation of TLSA records is
4293 the application's responsibility. The application provides
4294 the TLSA records of its choice to OpenSSL, and these are then
4295 used to authenticate the peer.
4297 The TLSA records need not even come from DNS. They can, for
4298 example, be used to implement local end-entity certificate or
4299 trust-anchor "pinning", where the "pin" data takes the form
4300 of TLSA records, which can augment or replace verification
4301 based on the usual WebPKI public certification authorities.
4305 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4306 continues to support deprecated interfaces in default builds.
4307 However, applications are strongly advised to compile their
4308 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4309 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4310 or the 1.1.0 releases.
4312 In environments in which all applications have been ported to
4313 not use any deprecated interfaces OpenSSL's Configure script
4314 should be used with the --api=1.1.0 option to entirely remove
4315 support for the deprecated features from the library and
4316 unconditionally disable them in the installed headers.
4317 Essentially the same effect can be achieved with the "no-deprecated"
4318 argument to Configure, except that this will always restrict
4319 the build to just the latest API, rather than a fixed API
4322 As applications are ported to future revisions of the API,
4323 they should update their compile-time OPENSSL_API_COMPAT define
4324 accordingly, but in most cases should be able to continue to
4325 compile with later releases.
4327 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4328 0x10000000L and 0x00908000L, respectively. However those
4329 versions did not support the OPENSSL_API_COMPAT feature, and
4330 so applications are not typically tested for explicit support
4331 of just the undeprecated features of either release.
4335 * Add support for setting the minimum and maximum supported protocol.
4336 It can bet set via the SSL_set_min_proto_version() and
4337 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4338 MaxProtocol. It's recommended to use the new APIs to disable
4339 protocols instead of disabling individual protocols using
4340 SSL_set_options() or SSL_CONF's Protocol. This change also
4341 removes support for disabling TLS 1.2 in the OpenSSL TLS
4342 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4346 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4350 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4351 and integrates ECDSA and ECDH functionality into EC. Implementations can
4352 now redirect key generation and no longer need to convert to or from
4355 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4356 include the ec.h header file instead.
4360 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4361 ciphers who are no longer supported and drops support the ephemeral RSA key
4362 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4366 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4367 opaque. For HMAC_CTX, the following constructors and destructors
4370 HMAC_CTX *HMAC_CTX_new(void);
4371 void HMAC_CTX_free(HMAC_CTX *ctx);
4373 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4374 destroy such methods has been added. See EVP_MD_meth_new(3) and
4375 EVP_CIPHER_meth_new(3) for documentation.
4378 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4379 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4380 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4381 an already created structure.
4382 2) For consistency with the majority of our object creators and
4383 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4384 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4385 for deprecated builds.
4389 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4390 cryptographic operations to be performed asynchronously as long as an
4391 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4392 further details. Libssl has also had this capability integrated with the
4393 introduction of the new mode SSL_MODE_ASYNC and associated error
4394 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4395 pages. This work was developed in partnership with Intel Corp.
4399 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4400 always enabled now. If you want to disable the support you should
4401 exclude it using the list of supported ciphers. This also means that the
4402 "-no_ecdhe" option has been removed from s_server.
4406 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4407 SSL_{CTX_}set1_curves() which can set a list.
4411 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4412 curve you want to support using SSL_{CTX_}set1_curves().
4416 * State machine rewrite. The state machine code has been significantly
4417 refactored in order to remove much duplication of code and solve issues
4418 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4419 further details). This change does have some associated API changes.
4420 Notably the SSL_state() function has been removed and replaced by
4421 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4422 SSL_set_state() has been removed altogether. The previous handshake states
4423 defined in ssl.h and ssl3.h have also been removed.
4427 * All instances of the string "ssleay" in the public API were replaced
4428 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4429 Some error codes related to internal RSA_eay API's were renamed.
4433 * The demo files in crypto/threads were moved to demo/threads.
4437 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4440 *Matt Caswell, Rich Salz*
4442 * New ASN.1 embed macro.
4444 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4445 structure is not allocated: it is part of the parent. That is instead of
4453 This reduces memory fragmentation and make it impossible to accidentally
4454 set a mandatory field to NULL.
4456 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4457 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4458 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4463 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4467 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4468 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4469 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4470 DES and RC4 ciphersuites.
4474 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4475 This changes the decoding behaviour for some invalid messages,
4476 though the change is mostly in the more lenient direction, and
4477 legacy behaviour is preserved as much as possible.
4481 * Fix no-stdio build.
4482 *David Woodhouse <David.Woodhouse@intel.com> and also*
4483 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4485 * New testing framework
4486 The testing framework has been largely rewritten and is now using
4487 perl and the perl modules Test::Harness and an extended variant of
4488 Test::More called OpenSSL::Test to do its work. All test scripts in
4489 test/ have been rewritten into test recipes, and all direct calls to
4490 executables in test/Makefile have become individual recipes using the
4491 simplified testing OpenSSL::Test::Simple.
4493 For documentation on our testing modules, do:
4495 perldoc test/testlib/OpenSSL/Test/Simple.pm
4496 perldoc test/testlib/OpenSSL/Test.pm
4500 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4501 are used; the latter aborts on memory leaks (usually checked on exit).
4502 Some undocumented "set malloc, etc., hooks" functions were removed
4503 and others were changed. All are now documented.
4507 * In DSA_generate_parameters_ex, if the provided seed is too short,
4510 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4512 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4513 from RFC4279, RFC4785, RFC5487, RFC5489.
4515 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4516 original RSA_PSK patch.
4520 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4521 era flag was never set throughout the codebase (only read). Also removed
4522 SSL3_FLAGS_POP_BUFFER which was only used if
4523 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4527 * Changed the default name options in the "ca", "crl", "req" and "x509"
4528 to be "oneline" instead of "compat".
4532 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4533 not aware of clients that still exhibit this bug, and the workaround
4534 hasn't been working properly for a while.
4538 * The return type of BIO_number_read() and BIO_number_written() as well as
4539 the corresponding num_read and num_write members in the BIO structure has
4540 changed from unsigned long to uint64_t. On platforms where an unsigned
4541 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4546 * Given the pervasive nature of TLS extensions it is inadvisable to run
4547 OpenSSL without support for them. It also means that maintaining
4548 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4549 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4553 * Removed support for the two export grade static DH ciphersuites
4554 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4555 were newly added (along with a number of other static DH ciphersuites) to
4556 1.0.2. However the two export ones have *never* worked since they were
4557 introduced. It seems strange in any case to be adding new export
4558 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4562 * Version negotiation has been rewritten. In particular SSLv23_method(),
4563 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4564 and turned into macros which simply call the new preferred function names
4565 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4566 should use the new names instead. Also as part of this change the ssl23.h
4567 header file has been removed.
4571 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4572 code and the associated standard is no longer considered fit-for-purpose.
4576 * RT2547 was closed. When generating a private key, try to make the
4577 output file readable only by the owner. This behavior change might
4578 be noticeable when interacting with other software.
4580 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4585 * Added HTTP GET support to the ocsp command.
4589 * Changed default digest for the dgst and enc commands from MD5 to
4594 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4598 * Added support for TLS extended master secret from
4599 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4600 initial patch which was a great help during development.
4604 * All libssl internal structures have been removed from the public header
4605 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4606 now redundant). Users should not attempt to access internal structures
4607 directly. Instead they should use the provided API functions.
4611 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4612 Access to deprecated functions can be re-enabled by running config with
4613 "enable-deprecated". In addition applications wishing to use deprecated
4614 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4615 will, by default, disable some transitive includes that previously existed
4616 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4620 * Added support for OCB mode. OpenSSL has been granted a patent license
4621 compatible with the OpenSSL license for use of OCB. Details are available
4622 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4623 for OCB can be removed by calling config with no-ocb.
4627 * SSLv2 support has been removed. It still supports receiving an SSLv2
4628 compatible client hello.
4632 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4633 done while fixing the error code for the key-too-small case.
4635 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4637 * CA.sh has been removed; use CA.pl instead.
4641 * Removed old DES API.
4645 * Remove various unsupported platforms:
4651 Sinix/ReliantUNIX RM400
4656 16-bit platforms such as WIN16
4660 * Clean up OPENSSL_NO_xxx #define's
4661 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4662 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4663 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4664 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4665 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4666 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4667 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4668 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4669 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4670 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4674 * Cleaned up dead code
4675 Remove all but one '#ifdef undef' which is to be looked at.
4679 * Clean up calling of xxx_free routines.
4680 Just like free(), fix most of the xxx_free routines to accept
4681 NULL. Remove the non-null checks from callers. Save much code.
4685 * Add secure heap for storage of private keys (when possible).
4686 Add BIO_s_secmem(), CBIGNUM, etc.
4687 Contributed by Akamai Technologies under our Corporate CLA.
4691 * Experimental support for a new, fast, unbiased prime candidate generator,
4692 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4694 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4696 * New output format NSS in the sess_id command line tool. This allows
4697 exporting the session id and the master key in NSS keylog format.
4699 *Martin Kaiser <martin@kaiser.cx>*
4701 * Harmonize version and its documentation. -f flag is used to display
4704 *mancha <mancha1@zoho.com>*
4706 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4707 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4709 *mancha <mancha1@zoho.com>*
4711 * Fix some double frees. These are not thought to be exploitable.
4713 *mancha <mancha1@zoho.com>*
4715 * A missing bounds check in the handling of the TLS heartbeat extension
4716 can be used to reveal up to 64k of memory to a connected client or
4719 Thanks for Neel Mehta of Google Security for discovering this bug and to
4720 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4721 preparing the fix ([CVE-2014-0160])
4723 *Adam Langley, Bodo Moeller*
4725 * Fix for the attack described in the paper "Recovering OpenSSL
4726 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4727 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4728 <http://eprint.iacr.org/2014/140>
4730 Thanks to Yuval Yarom and Naomi Benger for discovering this
4731 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4733 *Yuval Yarom and Naomi Benger*
4735 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4736 this fixes a limitation in previous versions of OpenSSL.
4740 * Experimental encrypt-then-mac support.
4742 Experimental support for encrypt then mac from
4743 draft-gutmann-tls-encrypt-then-mac-02.txt
4745 To enable it set the appropriate extension number (0x42 for the test
4746 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4748 For non-compliant peers (i.e. just about everything) this should have no
4751 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4755 * Add EVP support for key wrapping algorithms, to avoid problems with
4756 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4757 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4758 algorithms and include tests cases.
4762 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4767 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4768 MGF1 digest and OAEP label.
4772 * Make openssl verify return errors.
4774 *Chris Palmer <palmer@google.com> and Ben Laurie*
4776 * New function ASN1_TIME_diff to calculate the difference between two
4777 ASN1_TIME structures or one structure and the current time.
4781 * Update fips_test_suite to support multiple command line options. New
4782 test to induce all self test errors in sequence and check expected
4787 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4788 sign or verify all in one operation.
4792 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4793 test programs and fips_test_suite. Includes functionality to parse
4794 the minimal script output of fipsalgest.pl directly.
4798 * Add authorisation parameter to FIPS_module_mode_set().
4802 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4806 * Use separate DRBG fields for internal and external flags. New function
4807 FIPS_drbg_health_check() to perform on demand health checking. Add
4808 generation tests to fips_test_suite with reduced health check interval to
4809 demonstrate periodic health checking. Add "nodh" option to
4810 fips_test_suite to skip very slow DH test.
4814 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4819 * More extensive health check for DRBG checking many more failure modes.
4820 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4821 combination: call this in fips_test_suite.
4825 * Add support for canonical generation of DSA parameter 'g'. See
4828 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4829 POST to handle HMAC cases.
4833 * Add functions FIPS_module_version() and FIPS_module_version_text()
4834 to return numerical and string versions of the FIPS module number.
4838 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4839 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4840 outside the validated module in the FIPS capable OpenSSL.
4844 * Minor change to DRBG entropy callback semantics. In some cases
4845 there is no multiple of the block length between min_len and
4846 max_len. Allow the callback to return more than max_len bytes
4847 of entropy but discard any extra: it is the callback's responsibility
4848 to ensure that the extra data discarded does not impact the
4849 requested amount of entropy.
4853 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4854 information in FIPS186-3, SP800-57 and SP800-131A.
4858 * CCM support via EVP. Interface is very similar to GCM case except we
4859 must supply all data in one chunk (i.e. no update, final) and the
4860 message length must be supplied if AAD is used. Add algorithm test
4865 * Initial version of POST overhaul. Add POST callback to allow the status
4866 of POST to be monitored and/or failures induced. Modify fips_test_suite
4867 to use callback. Always run all selftests even if one fails.
4871 * XTS support including algorithm test driver in the fips_gcmtest program.
4872 Note: this does increase the maximum key length from 32 to 64 bytes but
4873 there should be no binary compatibility issues as existing applications
4874 will never use XTS mode.
4878 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4879 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4880 performs algorithm blocking for unapproved PRNG types. Also do not
4881 set PRNG type in FIPS_mode_set(): leave this to the application.
4882 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4883 the standard OpenSSL PRNG: set additional data to a date time vector.
4887 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4888 This shouldn't present any incompatibility problems because applications
4889 shouldn't be using these directly and any that are will need to rethink
4890 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4894 * Extensive self tests and health checking required by SP800-90 DRBG.
4895 Remove strength parameter from FIPS_drbg_instantiate and always
4896 instantiate at maximum supported strength.
4900 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4904 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4908 * New function DH_compute_key_padded() to compute a DH key and pad with
4909 leading zeroes if needed: this complies with SP800-56A et al.
4913 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4914 anything, incomplete, subject to change and largely untested at present.
4918 * Modify fipscanisteronly build option to only build the necessary object
4919 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4923 * Add experimental option FIPSSYMS to give all symbols in
4924 fipscanister.o and FIPS or fips prefix. This will avoid
4925 conflicts with future versions of OpenSSL. Add perl script
4926 util/fipsas.pl to preprocess assembly language source files
4927 and rename any affected symbols.
4931 * Add selftest checks and algorithm block of non-fips algorithms in
4932 FIPS mode. Remove DES2 from selftests.
4936 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4937 return internal method without any ENGINE dependencies. Add new
4938 tiny fips sign and verify functions.
4942 * New build option no-ec2m to disable characteristic 2 code.
4946 * New build option "fipscanisteronly". This only builds fipscanister.o
4947 and (currently) associated fips utilities. Uses the file Makefile.fips
4948 instead of Makefile.org as the prototype.
4952 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4953 Update fips_gcmtest to use IV generator.
4957 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4958 setting output buffer to NULL. The `*Final` function must be
4959 called although it will not retrieve any additional data. The tag
4960 can be set or retrieved with a ctrl. The IV length is by default 12
4961 bytes (96 bits) but can be set to an alternative value. If the IV
4962 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4967 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4968 underlying do_cipher function handles all cipher semantics itself
4969 including padding and finalisation. This is useful if (for example)
4970 an ENGINE cipher handles block padding itself. The behaviour of
4971 do_cipher is subtly changed if this flag is set: the return value
4972 is the number of characters written to the output buffer (zero is
4973 no longer an error code) or a negative error code. Also if the
4974 input buffer is NULL and length 0 finalisation should be performed.
4978 * If a candidate issuer certificate is already part of the constructed
4979 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4983 * Improve forward-security support: add functions
4985 void SSL_CTX_set_not_resumable_session_callback(
4986 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4987 void SSL_set_not_resumable_session_callback(
4988 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4990 for use by SSL/TLS servers; the callback function will be called whenever a
4991 new session is created, and gets to decide whether the session may be
4992 cached to make it resumable (return 0) or not (return 1). (As by the
4993 SSL/TLS protocol specifications, the session_id sent by the server will be
4994 empty to indicate that the session is not resumable; also, the server will
4995 not generate RFC 4507 (RFC 5077) session tickets.)
4997 A simple reasonable callback implementation is to return is_forward_secure.
4998 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4999 by the SSL/TLS server library, indicating whether it can provide forward
5002 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5004 * New -verify_name option in command line utilities to set verification
5009 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5010 Add CMAC pkey methods.
5014 * Experimental renegotiation in s_server -www mode. If the client
5015 browses /reneg connection is renegotiated. If /renegcert it is
5016 renegotiated requesting a certificate.
5020 * Add an "external" session cache for debugging purposes to s_server. This
5021 should help trace issues which normally are only apparent in deployed
5022 multi-process servers.
5026 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5027 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5028 BIO_set_cipher() and some obscure PEM functions were changed so they
5029 can now return an error. The RAND changes required a change to the
5030 RAND_METHOD structure.
5034 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5035 a gcc attribute to warn if the result of a function is ignored. This
5036 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5037 whose return value is often ignored.
5041 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5042 These allow SCTs (signed certificate timestamps) to be requested and
5043 validated when establishing a connection.
5045 *Rob Percival <robpercival@google.com>*
5050 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5052 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5053 used even when parsing explicit parameters, when loading a encoded key
5054 or calling `EC_GROUP_new_from_ecpkparameters()`/
5055 `EC_GROUP_new_from_ecparameters()`.
5056 This prevents bypass of security hardening and performance gains,
5057 especially for curves with specialized EC_METHODs.
5058 By default, if a key encoded with explicit parameters is loaded and later
5059 encoded, the output is still encoded with explicit parameters, even if
5060 internally a "named" EC_GROUP is used for computation.
5064 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5065 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5066 NULL. After this change, only the cofactor parameter can be NULL. It also
5067 does some minimal sanity checks on the passed order.
5072 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5073 An attack is simple, if the first CMS_recipientInfo is valid but the
5074 second CMS_recipientInfo is chosen ciphertext. If the second
5075 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5076 encryption key will be replaced by garbage, and the message cannot be
5077 decoded, but if the RSA decryption fails, the correct encryption key is
5078 used and the recipient will not notice the attack.
5079 As a work around for this potential attack the length of the decrypted
5080 key must be equal to the cipher default key length, in case the
5081 certificate is not given and all recipientInfo are tried out.
5082 The old behaviour can be re-enabled in the CMS code by setting the
5083 CMS_DEBUG_DECRYPT flag.
5088 * Document issue with installation paths in diverse Windows builds
5090 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5091 binaries and run-time config file.
5096 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5098 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5099 This changes the size when using the `genpkey` command when no size is given.
5100 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5101 generation commands to use 2048 bits by default.
5105 * Add FIPS support for Android Arm 64-bit
5107 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5108 Module in Version 2.0.10. For some reason, the corresponding target
5109 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5110 built with FIPS support on Android Arm 64-bit. This omission has been
5113 *Matthias St. Pierre*
5115 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5117 * 0-byte record padding oracle
5119 If an application encounters a fatal protocol error and then calls
5120 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5121 then OpenSSL can respond differently to the calling application if a 0 byte
5122 record is received with invalid padding compared to if a 0 byte record is
5123 received with an invalid MAC. If the application then behaves differently
5124 based on that in a way that is detectable to the remote peer, then this
5125 amounts to a padding oracle that could be used to decrypt data.
5127 In order for this to be exploitable "non-stitched" ciphersuites must be in
5128 use. Stitched ciphersuites are optimised implementations of certain
5129 commonly used ciphersuites. Also the application must call SSL_shutdown()
5130 twice even if a protocol error has occurred (applications should not do
5131 this but some do anyway).
5133 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5134 Aviram, with additional investigation by Steven Collison and Andrew
5135 Hourselt. It was reported to OpenSSL on 10th December 2018.
5140 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5144 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5146 * Microarchitecture timing vulnerability in ECC scalar multiplication
5148 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5149 shown to be vulnerable to a microarchitecture timing side channel attack.
5150 An attacker with sufficient access to mount local timing attacks during
5151 ECDSA signature generation could recover the private key.
5153 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5154 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5160 * Timing vulnerability in DSA signature generation
5162 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5163 timing side channel attack. An attacker could use variations in the signing
5164 algorithm to recover the private key.
5166 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5171 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5172 Module, accidentally introduced while backporting security fixes from the
5173 development branch and hindering the use of ECC in FIPS mode.
5177 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5179 * Client DoS due to large DH parameter
5181 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5182 malicious server can send a very large prime value to the client. This will
5183 cause the client to spend an unreasonably long period of time generating a
5184 key for this prime resulting in a hang until the client has finished. This
5185 could be exploited in a Denial Of Service attack.
5187 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5192 * Cache timing vulnerability in RSA Key Generation
5194 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5195 a cache timing side channel attack. An attacker with sufficient access to
5196 mount cache timing attacks during the RSA key generation process could
5197 recover the private key.
5199 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5200 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5205 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5206 parameter is no longer accepted, as it leads to a corrupt table. NULL
5207 pem_str is reserved for alias entries only.
5211 * Revert blinding in ECDSA sign and instead make problematic addition
5212 length-invariant. Switch even to fixed-length Montgomery multiplication.
5216 * Change generating and checking of primes so that the error rate of not
5217 being prime depends on the intended use based on the size of the input.
5218 For larger primes this will result in more rounds of Miller-Rabin.
5219 The maximal error rate for primes with more than 1080 bits is lowered
5222 *Kurt Roeckx, Annie Yousar*
5224 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5228 * Add blinding to ECDSA and DSA signatures to protect against side channel
5229 attacks discovered by Keegan Ryan (NCC Group).
5233 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5234 now allow empty (zero character) pass phrases.
5238 * Certificate time validation (X509_cmp_time) enforces stricter
5239 compliance with RFC 5280. Fractional seconds and timezone offsets
5240 are no longer allowed.
5244 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5246 * Constructed ASN.1 types with a recursive definition could exceed the stack
5248 Constructed ASN.1 types with a recursive definition (such as can be found
5249 in PKCS7) could eventually exceed the stack given malicious input with
5250 excessive recursion. This could result in a Denial Of Service attack. There
5251 are no such structures used within SSL/TLS that come from untrusted sources
5252 so this is considered safe.
5254 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5260 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5262 * Read/write after SSL object in error state
5264 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5265 mechanism. The intent was that if a fatal error occurred during a handshake
5266 then OpenSSL would move into the error state and would immediately fail if
5267 you attempted to continue the handshake. This works as designed for the
5268 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5269 SSL_connect()), however due to a bug it does not work correctly if
5270 SSL_read() or SSL_write() is called directly. In that scenario, if the
5271 handshake fails then a fatal error will be returned in the initial function
5272 call. If SSL_read()/SSL_write() is subsequently called by the application
5273 for the same SSL object then it will succeed and the data is passed without
5274 being decrypted/encrypted directly from the SSL/TLS record layer.
5276 In order to exploit this issue an application bug would have to be present
5277 that resulted in a call to SSL_read()/SSL_write() being issued after having
5278 already received a fatal error.
5280 This issue was reported to OpenSSL by David Benjamin (Google).
5285 * rsaz_1024_mul_avx2 overflow bug on x86_64
5287 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5288 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5289 Analysis suggests that attacks against RSA and DSA as a result of this
5290 defect would be very difficult to perform and are not believed likely.
5291 Attacks against DH1024 are considered just feasible, because most of the
5292 work necessary to deduce information about a private key may be performed
5293 offline. The amount of resources required for such an attack would be
5294 significant. However, for an attack on TLS to be meaningful, the server
5295 would have to share the DH1024 private key among multiple clients, which is
5296 no longer an option since CVE-2016-0701.
5298 This only affects processors that support the AVX2 but not ADX extensions
5299 like Intel Haswell (4th generation).
5301 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5302 was originally found via the OSS-Fuzz project.
5307 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5309 * bn_sqrx8x_internal carry bug on x86_64
5311 There is a carry propagating bug in the x86_64 Montgomery squaring
5312 procedure. No EC algorithms are affected. Analysis suggests that attacks
5313 against RSA and DSA as a result of this defect would be very difficult to
5314 perform and are not believed likely. Attacks against DH are considered just
5315 feasible (although very difficult) because most of the work necessary to
5316 deduce information about a private key may be performed offline. The amount
5317 of resources required for such an attack would be very significant and
5318 likely only accessible to a limited number of attackers. An attacker would
5319 additionally need online access to an unpatched system using the target
5320 private key in a scenario with persistent DH parameters and a private
5321 key that is shared between multiple clients.
5323 This only affects processors that support the BMI1, BMI2 and ADX extensions
5324 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5326 This issue was reported to OpenSSL by the OSS-Fuzz project.
5331 * Malformed X.509 IPAddressFamily could cause OOB read
5333 If an X.509 certificate has a malformed IPAddressFamily extension,
5334 OpenSSL could do a one-byte buffer overread. The most likely result
5335 would be an erroneous display of the certificate in text format.
5337 This issue was reported to OpenSSL by the OSS-Fuzz project.
5341 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5343 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5344 platform rather than 'mingw'.
5348 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5350 * Truncated packet could crash via OOB read
5352 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5353 cipher is being used, then a truncated packet can cause that host to
5354 perform an out-of-bounds read, usually resulting in a crash.
5356 This issue was reported to OpenSSL by Robert Święcki of Google.
5361 * BN_mod_exp may produce incorrect results on x86_64
5363 There is a carry propagating bug in the x86_64 Montgomery squaring
5364 procedure. No EC algorithms are affected. Analysis suggests that attacks
5365 against RSA and DSA as a result of this defect would be very difficult to
5366 perform and are not believed likely. Attacks against DH are considered just
5367 feasible (although very difficult) because most of the work necessary to
5368 deduce information about a private key may be performed offline. The amount
5369 of resources required for such an attack would be very significant and
5370 likely only accessible to a limited number of attackers. An attacker would
5371 additionally need online access to an unpatched system using the target
5372 private key in a scenario with persistent DH parameters and a private
5373 key that is shared between multiple clients. For example this can occur by
5374 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5375 similar to CVE-2015-3193 but must be treated as a separate problem.
5377 This issue was reported to OpenSSL by the OSS-Fuzz project.
5382 * Montgomery multiplication may produce incorrect results
5384 There is a carry propagating bug in the Broadwell-specific Montgomery
5385 multiplication procedure that handles input lengths divisible by, but
5386 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5387 and DH private keys are impossible. This is because the subroutine in
5388 question is not used in operations with the private key itself and an input
5389 of the attacker's direct choice. Otherwise the bug can manifest itself as
5390 transient authentication and key negotiation failures or reproducible
5391 erroneous outcome of public-key operations with specially crafted input.
5392 Among EC algorithms only Brainpool P-512 curves are affected and one
5393 presumably can attack ECDH key negotiation. Impact was not analyzed in
5394 detail, because pre-requisites for attack are considered unlikely. Namely
5395 multiple clients have to choose the curve in question and the server has to
5396 share the private key among them, neither of which is default behaviour.
5397 Even then only clients that chose the curve will be affected.
5399 This issue was publicly reported as transient failures and was not
5400 initially recognized as a security issue. Thanks to Richard Morgan for
5401 providing reproducible case.
5406 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5407 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5408 prevent issues where no progress is being made and the peer continually
5409 sends unrecognised record types, using up resources processing them.
5413 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5415 * Missing CRL sanity check
5417 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5418 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5419 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5421 This issue only affects the OpenSSL 1.0.2i
5426 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5428 * OCSP Status Request extension unbounded memory growth
5430 A malicious client can send an excessively large OCSP Status Request
5431 extension. If that client continually requests renegotiation, sending a
5432 large OCSP Status Request extension each time, then there will be unbounded
5433 memory growth on the server. This will eventually lead to a Denial Of
5434 Service attack through memory exhaustion. Servers with a default
5435 configuration are vulnerable even if they do not support OCSP. Builds using
5436 the "no-ocsp" build time option are not affected.
5438 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5443 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5446 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5452 * OOB write in MDC2_Update()
5454 An overflow can occur in MDC2_Update() either if called directly or
5455 through the EVP_DigestUpdate() function using MDC2. If an attacker
5456 is able to supply very large amounts of input data after a previous
5457 call to EVP_EncryptUpdate() with a partial block then a length check
5458 can overflow resulting in a heap corruption.
5460 The amount of data needed is comparable to SIZE_MAX which is impractical
5463 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5468 * Malformed SHA512 ticket DoS
5470 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5471 DoS attack where a malformed ticket will result in an OOB read which will
5474 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5475 a custom server callback and ticket lookup mechanism.
5477 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5482 * OOB write in BN_bn2dec()
5484 The function BN_bn2dec() does not check the return value of BN_div_word().
5485 This can cause an OOB write if an application uses this function with an
5486 overly large BIGNUM. This could be a problem if an overly large certificate
5487 or CRL is printed out from an untrusted source. TLS is not affected because
5488 record limits will reject an oversized certificate before it is parsed.
5490 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5495 * OOB read in TS_OBJ_print_bio()
5497 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5498 the total length the OID text representation would use and not the amount
5499 of data written. This will result in OOB reads when large OIDs are
5502 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5507 * Pointer arithmetic undefined behaviour
5509 Avoid some undefined pointer arithmetic
5511 A common idiom in the codebase is to check limits in the following manner:
5514 Where "p" points to some malloc'd data of SIZE bytes and
5517 "len" here could be from some externally supplied data (e.g. from a TLS
5520 The rules of C pointer arithmetic are such that "p + len" is only well
5521 defined where len <= SIZE. Therefore the above idiom is actually
5522 undefined behaviour.
5524 For example this could cause problems if some malloc implementation
5525 provides an address for "p" such that "p + len" actually overflows for
5526 values of len that are too big and therefore p + len < limit.
5528 This issue was reported to OpenSSL by Guido Vranken
5533 * Constant time flag not preserved in DSA signing
5535 Operations in the DSA signing algorithm should run in constant time in
5536 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5537 implementation means that a non-constant time codepath is followed for
5538 certain operations. This has been demonstrated through a cache-timing
5539 attack to be sufficient for an attacker to recover the private DSA key.
5541 This issue was reported by César Pereida (Aalto University), Billy Brumley
5542 (Tampere University of Technology), and Yuval Yarom (The University of
5543 Adelaide and NICTA).
5548 * DTLS buffered message DoS
5550 In a DTLS connection where handshake messages are delivered out-of-order
5551 those messages that OpenSSL is not yet ready to process will be buffered
5552 for later use. Under certain circumstances, a flaw in the logic means that
5553 those messages do not get removed from the buffer even though the handshake
5554 has been completed. An attacker could force up to approx. 15 messages to
5555 remain in the buffer when they are no longer required. These messages will
5556 be cleared when the DTLS connection is closed. The default maximum size for
5557 a message is 100k. Therefore, the attacker could force an additional 1500k
5558 to be consumed per connection. By opening many simultaneous connections an
5559 attacker could cause a DoS attack through memory exhaustion.
5561 This issue was reported to OpenSSL by Quan Luo.
5566 * DTLS replay protection DoS
5568 A flaw in the DTLS replay attack protection mechanism means that records
5569 that arrive for future epochs update the replay protection "window" before
5570 the MAC for the record has been validated. This could be exploited by an
5571 attacker by sending a record for the next epoch (which does not have to
5572 decrypt or have a valid MAC), with a very large sequence number. This means
5573 that all subsequent legitimate packets are dropped causing a denial of
5574 service for a specific DTLS connection.
5576 This issue was reported to OpenSSL by the OCAP audit team.
5581 * Certificate message OOB reads
5583 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5584 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5585 theoretical DoS risk but this has not been observed in practice on common
5588 The messages affected are client certificate, client certificate request
5589 and server certificate. As a result the attack can only be performed
5590 against a client or a server which enables client authentication.
5592 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5597 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5599 * Prevent padding oracle in AES-NI CBC MAC check
5601 A MITM attacker can use a padding oracle attack to decrypt traffic
5602 when the connection uses an AES CBC cipher and the server support
5605 This issue was introduced as part of the fix for Lucky 13 padding
5606 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5607 constant time by making sure that always the same bytes are read and
5608 compared against either the MAC or padding bytes. But it no longer
5609 checked that there was enough data to have both the MAC and padding
5612 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5616 * Fix EVP_EncodeUpdate overflow
5618 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5619 Base64 encoding of binary data. If an attacker is able to supply very large
5620 amounts of input data then a length check can overflow resulting in a heap
5623 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5624 the `PEM_write_bio*` family of functions. These are mainly used within the
5625 OpenSSL command line applications, so any application which processes data
5626 from an untrusted source and outputs it as a PEM file should be considered
5627 vulnerable to this issue. User applications that call these APIs directly
5628 with large amounts of untrusted data may also be vulnerable.
5630 This issue was reported by Guido Vranken.
5635 * Fix EVP_EncryptUpdate overflow
5637 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5638 is able to supply very large amounts of input data after a previous call to
5639 EVP_EncryptUpdate() with a partial block then a length check can overflow
5640 resulting in a heap corruption. Following an analysis of all OpenSSL
5641 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5642 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5643 the first called function after an EVP_EncryptInit(), and therefore that
5644 specific call must be safe. The second form is where the length passed to
5645 EVP_EncryptUpdate() can be seen from the code to be some small value and
5646 therefore there is no possibility of an overflow. Since all instances are
5647 one of these two forms, it is believed that there can be no overflows in
5648 internal code due to this problem. It should be noted that
5649 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5650 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5651 of these calls have also been analysed too and it is believed there are no
5652 instances in internal usage where an overflow could occur.
5654 This issue was reported by Guido Vranken.
5659 * Prevent ASN.1 BIO excessive memory allocation
5661 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5662 a short invalid encoding can cause allocation of large amounts of memory
5663 potentially consuming excessive resources or exhausting memory.
5665 Any application parsing untrusted data through d2i BIO functions is
5666 affected. The memory based functions such as d2i_X509() are *not* affected.
5667 Since the memory based functions are used by the TLS library, TLS
5668 applications are not affected.
5670 This issue was reported by Brian Carpenter.
5677 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5678 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5679 in arbitrary stack data being returned in the buffer.
5681 This issue was reported by Guido Vranken.
5686 * Modify behavior of ALPN to invoke callback after SNI/servername
5687 callback, such that updates to the SSL_CTX affect ALPN.
5691 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5696 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5697 methods are enabled and ssl2 is disabled the methods return NULL.
5701 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5703 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5704 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5705 provide any "EXPORT" or "LOW" strength ciphers.
5709 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5710 is by default disabled at build-time. Builds that are not configured with
5711 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5712 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5713 will need to explicitly call either of:
5715 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5717 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5719 as appropriate. Even if either of those is used, or the application
5720 explicitly uses the version-specific SSLv2_method() or its client and
5721 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5722 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5723 ciphers, and SSLv2 56-bit DES are no longer available.
5728 * Fix a double-free in DSA code
5730 A double free bug was discovered when OpenSSL parses malformed DSA private
5731 keys and could lead to a DoS attack or memory corruption for applications
5732 that receive DSA private keys from untrusted sources. This scenario is
5735 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5741 * Disable SRP fake user seed to address a server memory leak.
5743 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5745 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5746 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5747 was changed to ignore the "fake user" SRP seed, even if the seed
5750 Users should use SRP_VBASE_get1_by_user instead. Note that in
5751 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5752 also that even though configuring the SRP seed attempts to hide
5753 invalid usernames by continuing the handshake with fake
5754 credentials, this behaviour is not constant time and no strong
5755 guarantees are made that the handshake is indistinguishable from
5756 that of a valid user.
5761 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5763 In the BN_hex2bn function the number of hex digits is calculated using an
5764 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5765 large values of `i` this can result in `bn_expand` not allocating any
5766 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5767 field as NULL leading to a subsequent NULL ptr deref. For very large values
5768 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5769 In this case memory is allocated to the internal BIGNUM data field, but it
5770 is insufficiently sized leading to heap corruption. A similar issue exists
5771 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5772 is ever called by user applications with very large untrusted hex/dec data.
5773 This is anticipated to be a rare occurrence.
5775 All OpenSSL internal usage of these functions use data that is not expected
5776 to be untrusted, e.g. config file data or application command line
5777 arguments. If user developed applications generate config file data based
5778 on untrusted data then it is possible that this could also lead to security
5779 consequences. This is also anticipated to be rare.
5781 This issue was reported to OpenSSL by Guido Vranken.
5786 * Fix memory issues in `BIO_*printf` functions
5788 The internal `fmtstr` function used in processing a "%s" format string in
5789 the `BIO_*printf` functions could overflow while calculating the length of a
5790 string and cause an OOB read when printing very long strings.
5792 Additionally the internal `doapr_outch` function can attempt to write to an
5793 OOB memory location (at an offset from the NULL pointer) in the event of a
5794 memory allocation failure. In 1.0.2 and below this could be caused where
5795 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5796 could be in processing a very long "%s" format string. Memory leaks can
5799 The first issue may mask the second issue dependent on compiler behaviour.
5800 These problems could enable attacks where large amounts of untrusted data
5801 is passed to the `BIO_*printf` functions. If applications use these functions
5802 in this way then they could be vulnerable. OpenSSL itself uses these
5803 functions when printing out human-readable dumps of ASN.1 data. Therefore
5804 applications that print this data could be vulnerable if the data is from
5805 untrusted sources. OpenSSL command line applications could also be
5806 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5807 as command line arguments.
5809 Libssl is not considered directly vulnerable. Additionally certificates etc
5810 received via remote connections via libssl are also unlikely to be able to
5811 trigger these issues because of message size limits enforced within libssl.
5813 This issue was reported to OpenSSL Guido Vranken.
5818 * Side channel attack on modular exponentiation
5820 A side-channel attack was found which makes use of cache-bank conflicts on
5821 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5822 of RSA keys. The ability to exploit this issue is limited as it relies on
5823 an attacker who has control of code in a thread running on the same
5824 hyper-threaded core as the victim thread which is performing decryptions.
5826 This issue was reported to OpenSSL by Yuval Yarom, The University of
5827 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5828 Nadia Heninger, University of Pennsylvania with more information at
5829 <http://cachebleed.info>.
5834 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5835 if no keysize is specified with default_bits. This fixes an
5836 omission in an earlier change that changed all RSA/DSA key generation
5837 commands to use 2048 bits by default.
5841 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5843 * DH small subgroups
5845 Historically OpenSSL only ever generated DH parameters based on "safe"
5846 primes. More recently (in version 1.0.2) support was provided for
5847 generating X9.42 style parameter files such as those required for RFC 5114
5848 support. The primes used in such files may not be "safe". Where an
5849 application is using DH configured with parameters based on primes that are
5850 not "safe" then an attacker could use this fact to find a peer's private
5851 DH exponent. This attack requires that the attacker complete multiple
5852 handshakes in which the peer uses the same private DH exponent. For example
5853 this could be used to discover a TLS server's private DH exponent if it's
5854 reusing the private DH exponent or it's using a static DH ciphersuite.
5856 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5857 TLS. It is not on by default. If the option is not set then the server
5858 reuses the same private DH exponent for the life of the server process and
5859 would be vulnerable to this attack. It is believed that many popular
5860 applications do set this option and would therefore not be at risk.
5862 The fix for this issue adds an additional check where a "q" parameter is
5863 available (as is the case in X9.42 based parameters). This detects the
5864 only known attack, and is the only possible defense for static DH
5865 ciphersuites. This could have some performance impact.
5867 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5868 default and cannot be disabled. This could have some performance impact.
5870 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5875 * SSLv2 doesn't block disabled ciphers
5877 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5878 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5879 been disabled, provided that the SSLv2 protocol was not also disabled via
5882 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5883 and Sebastian Schinzel.
5888 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5890 * BN_mod_exp may produce incorrect results on x86_64
5892 There is a carry propagating bug in the x86_64 Montgomery squaring
5893 procedure. No EC algorithms are affected. Analysis suggests that attacks
5894 against RSA and DSA as a result of this defect would be very difficult to
5895 perform and are not believed likely. Attacks against DH are considered just
5896 feasible (although very difficult) because most of the work necessary to
5897 deduce information about a private key may be performed offline. The amount
5898 of resources required for such an attack would be very significant and
5899 likely only accessible to a limited number of attackers. An attacker would
5900 additionally need online access to an unpatched system using the target
5901 private key in a scenario with persistent DH parameters and a private
5902 key that is shared between multiple clients. For example this can occur by
5903 default in OpenSSL DHE based SSL/TLS ciphersuites.
5905 This issue was reported to OpenSSL by Hanno Böck.
5910 * Certificate verify crash with missing PSS parameter
5912 The signature verification routines will crash with a NULL pointer
5913 dereference if presented with an ASN.1 signature using the RSA PSS
5914 algorithm and absent mask generation function parameter. Since these
5915 routines are used to verify certificate signature algorithms this can be
5916 used to crash any certificate verification operation and exploited in a
5917 DoS attack. Any application which performs certificate verification is
5918 vulnerable including OpenSSL clients and servers which enable client
5921 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5926 * X509_ATTRIBUTE memory leak
5928 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5929 memory. This structure is used by the PKCS#7 and CMS routines so any
5930 application which reads PKCS#7 or CMS data from untrusted sources is
5931 affected. SSL/TLS is not affected.
5933 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5939 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5940 This changes the decoding behaviour for some invalid messages,
5941 though the change is mostly in the more lenient direction, and
5942 legacy behaviour is preserved as much as possible.
5946 * In DSA_generate_parameters_ex, if the provided seed is too short,
5949 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5951 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5953 * Alternate chains certificate forgery
5955 During certificate verification, OpenSSL will attempt to find an
5956 alternative certificate chain if the first attempt to build such a chain
5957 fails. An error in the implementation of this logic can mean that an
5958 attacker could cause certain checks on untrusted certificates to be
5959 bypassed, such as the CA flag, enabling them to use a valid leaf
5960 certificate to act as a CA and "issue" an invalid certificate.
5962 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5967 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5969 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5970 incompatibility in the handling of HMAC. The previous ABI has now been
5975 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5977 * Malformed ECParameters causes infinite loop
5979 When processing an ECParameters structure OpenSSL enters an infinite loop
5980 if the curve specified is over a specially malformed binary polynomial
5983 This can be used to perform denial of service against any
5984 system which processes public keys, certificate requests or
5985 certificates. This includes TLS clients and TLS servers with
5986 client authentication enabled.
5988 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5993 * Exploitable out-of-bounds read in X509_cmp_time
5995 X509_cmp_time does not properly check the length of the ASN1_TIME
5996 string and can read a few bytes out of bounds. In addition,
5997 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6000 An attacker can use this to craft malformed certificates and CRLs of
6001 various sizes and potentially cause a segmentation fault, resulting in
6002 a DoS on applications that verify certificates or CRLs. TLS clients
6003 that verify CRLs are affected. TLS clients and servers with client
6004 authentication enabled may be affected if they use custom verification
6007 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6008 independently by Hanno Böck.
6013 * PKCS7 crash with missing EnvelopedContent
6015 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6016 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6017 with missing content and trigger a NULL pointer dereference on parsing.
6019 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6020 structures from untrusted sources are affected. OpenSSL clients and
6021 servers are not affected.
6023 This issue was reported to OpenSSL by Michal Zalewski (Google).
6028 * CMS verify infinite loop with unknown hash function
6030 When verifying a signedData message the CMS code can enter an infinite loop
6031 if presented with an unknown hash function OID. This can be used to perform
6032 denial of service against any system which verifies signedData messages using
6034 This issue was reported to OpenSSL by Johannes Bauer.
6039 * Race condition handling NewSessionTicket
6041 If a NewSessionTicket is received by a multi-threaded client when attempting to
6042 reuse a previous ticket then a race condition can occur potentially leading to
6043 a double free of the ticket data.
6048 * Only support 256-bit or stronger elliptic curves with the
6049 'ecdh_auto' setting (server) or by default (client). Of supported
6050 curves, prefer P-256 (both).
6054 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6056 * ClientHello sigalgs DoS fix
6058 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6059 invalid signature algorithms extension a NULL pointer dereference will
6060 occur. This can be exploited in a DoS attack against the server.
6062 This issue was was reported to OpenSSL by David Ramos of Stanford
6066 *Stephen Henson and Matt Caswell*
6068 * Multiblock corrupted pointer fix
6070 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6071 feature only applies on 64 bit x86 architecture platforms that support AES
6072 NI instructions. A defect in the implementation of "multiblock" can cause
6073 OpenSSL's internal write buffer to become incorrectly set to NULL when
6074 using non-blocking IO. Typically, when the user application is using a
6075 socket BIO for writing, this will only result in a failed connection.
6076 However if some other BIO is used then it is likely that a segmentation
6077 fault will be triggered, thus enabling a potential DoS attack.
6079 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6084 * Segmentation fault in DTLSv1_listen fix
6086 The DTLSv1_listen function is intended to be stateless and processes the
6087 initial ClientHello from many peers. It is common for user code to loop
6088 over the call to DTLSv1_listen until a valid ClientHello is received with
6089 an associated cookie. A defect in the implementation of DTLSv1_listen means
6090 that state is preserved in the SSL object from one invocation to the next
6091 that can lead to a segmentation fault. Errors processing the initial
6092 ClientHello can trigger this scenario. An example of such an error could be
6093 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6096 This issue was reported to OpenSSL by Per Allansson.
6101 * Segmentation fault in ASN1_TYPE_cmp fix
6103 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6104 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6105 certificate signature algorithm consistency this can be used to crash any
6106 certificate verification operation and exploited in a DoS attack. Any
6107 application which performs certificate verification is vulnerable including
6108 OpenSSL clients and servers which enable client authentication.
6113 * Segmentation fault for invalid PSS parameters fix
6115 The signature verification routines will crash with a NULL pointer
6116 dereference if presented with an ASN.1 signature using the RSA PSS
6117 algorithm and invalid parameters. Since these routines are used to verify
6118 certificate signature algorithms this can be used to crash any
6119 certificate verification operation and exploited in a DoS attack. Any
6120 application which performs certificate verification is vulnerable including
6121 OpenSSL clients and servers which enable client authentication.
6123 This issue was was reported to OpenSSL by Brian Carpenter.
6128 * ASN.1 structure reuse memory corruption fix
6130 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6131 memory corruption via an invalid write. Such reuse is and has been
6132 strongly discouraged and is believed to be rare.
6134 Applications that parse structures containing CHOICE or ANY DEFINED BY
6135 components may be affected. Certificate parsing (d2i_X509 and related
6136 functions) are however not affected. OpenSSL clients and servers are
6142 * PKCS7 NULL pointer dereferences fix
6144 The PKCS#7 parsing code does not handle missing outer ContentInfo
6145 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6146 missing content and trigger a NULL pointer dereference on parsing.
6148 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6149 otherwise parse PKCS#7 structures from untrusted sources are
6150 affected. OpenSSL clients and servers are not affected.
6152 This issue was reported to OpenSSL by Michal Zalewski (Google).
6157 * DoS via reachable assert in SSLv2 servers fix
6159 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6160 servers that both support SSLv2 and enable export cipher suites by sending
6161 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6163 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6164 (OpenSSL development team).
6169 * Empty CKE with client auth and DHE fix
6171 If client auth is used then a server can seg fault in the event of a DHE
6172 ciphersuite being selected and a zero length ClientKeyExchange message
6173 being sent by the client. This could be exploited in a DoS attack.
6178 * Handshake with unseeded PRNG fix
6180 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6181 with an unseeded PRNG. The conditions are:
6182 - The client is on a platform where the PRNG has not been seeded
6183 automatically, and the user has not seeded manually
6184 - A protocol specific client method version has been used (i.e. not
6185 SSL_client_methodv23)
6186 - A ciphersuite is used that does not require additional random data from
6187 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6189 If the handshake succeeds then the client random that has been used will
6190 have been generated from a PRNG with insufficient entropy and therefore the
6191 output may be predictable.
6193 For example using the following command with an unseeded openssl will
6194 succeed on an unpatched platform:
6196 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6201 * Use After Free following d2i_ECPrivatekey error fix
6203 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6204 could cause a use after free condition. This, in turn, could cause a double
6205 free in several private key parsing functions (such as d2i_PrivateKey
6206 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6207 for applications that receive EC private keys from untrusted
6208 sources. This scenario is considered rare.
6210 This issue was discovered by the BoringSSL project and fixed in their
6216 * X509_to_X509_REQ NULL pointer deref fix
6218 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6219 the certificate key is invalid. This function is rarely used in practice.
6221 This issue was discovered by Brian Carpenter.
6226 * Removed the export ciphers from the DEFAULT ciphers
6230 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6232 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6233 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6234 So far those who have to target multiple platforms would compromise
6235 and argue that binary targeting say ARMv5 would still execute on
6236 ARMv8. "Universal" build resolves this compromise by providing
6237 near-optimal performance even on newer platforms.
6241 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6242 (other platforms pending).
6244 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6246 * Add support for the SignedCertificateTimestampList certificate and
6247 OCSP response extensions from RFC6962.
6251 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6252 for corner cases. (Certain input points at infinity could lead to
6253 bogus results, with non-infinity inputs mapped to infinity too.)
6257 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6258 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6259 common cases are optimized and there still is room for further
6260 improvements. Vector Permutation AES for Altivec is also added.
6264 * Add support for little-endian ppc64 Linux target.
6266 *Marcelo Cerri (IBM)*
6268 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6269 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6270 are optimized and there still is room for further improvements.
6271 Both 32- and 64-bit modes are supported.
6273 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6275 * Improved ARMv7 NEON support.
6279 * Support for SPARC Architecture 2011 crypto extensions, first
6280 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6281 SHA256/512, MD5, GHASH and modular exponentiation.
6283 *Andy Polyakov, David Miller*
6285 * Accelerated modular exponentiation for Intel processors, a.k.a.
6288 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6290 * Support for new and upcoming Intel processors, including AVX2,
6291 BMI and SHA ISA extensions. This includes additional "stitched"
6292 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6295 This work was sponsored by Intel Corp.
6299 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6300 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6301 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6305 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6306 this fixes a limitation in previous versions of OpenSSL.
6310 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6311 MGF1 digest and OAEP label.
6315 * Add EVP support for key wrapping algorithms, to avoid problems with
6316 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6317 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6318 algorithms and include tests cases.
6322 * Add functions to allocate and set the fields of an ECDSA_METHOD
6325 *Douglas E. Engert, Steve Henson*
6327 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6328 difference in days and seconds between two tm or ASN1_TIME structures.
6332 * Add -rev test option to s_server to just reverse order of characters
6333 received by client and send back to server. Also prints an abbreviated
6334 summary of the connection parameters.
6338 * New option -brief for s_client and s_server to print out a brief summary
6339 of connection parameters.
6343 * Add callbacks for arbitrary TLS extensions.
6345 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6347 * New option -crl_download in several openssl utilities to download CRLs
6348 from CRLDP extension in certificates.
6352 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6356 * New function X509_CRL_diff to generate a delta CRL from the difference
6357 of two full CRLs. Add support to "crl" utility.
6361 * New functions to set lookup_crls function and to retrieve
6362 X509_STORE from X509_STORE_CTX.
6366 * Print out deprecated issuer and subject unique ID fields in
6371 * Extend OCSP I/O functions so they can be used for simple general purpose
6372 HTTP as well as OCSP. New wrapper function which can be used to download
6373 CRLs using the OCSP API.
6377 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6381 * `SSL_CONF*` functions. These provide a common framework for application
6382 configuration using configuration files or command lines.
6386 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6387 message callback and prints the results. Needs compile time option
6388 "enable-ssl-trace". New options to s_client and s_server to enable
6393 * New ctrl and macro to retrieve supported points extensions.
6394 Print out extension in s_server and s_client.
6398 * New functions to retrieve certificate signature and signature
6403 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6408 * New Suite B modes for TLS code. These use and enforce the requirements
6409 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6410 only use Suite B curves. The Suite B modes can be set by using the
6411 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6415 * New chain verification flags for Suite B levels of security. Check
6416 algorithms are acceptable when flags are set in X509_verify_cert.
6420 * Make tls1_check_chain return a set of flags indicating checks passed
6421 by a certificate chain. Add additional tests to handle client
6422 certificates: checks for matching certificate type and issuer name
6427 * If an attempt is made to use a signature algorithm not in the peer
6428 preference list abort the handshake. If client has no suitable
6429 signature algorithms in response to a certificate request do not
6430 use the certificate.
6434 * If server EC tmp key is not in client preference list abort handshake.
6438 * Add support for certificate stores in CERT structure. This makes it
6439 possible to have different stores per SSL structure or one store in
6440 the parent SSL_CTX. Include distinct stores for certificate chain
6441 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6442 to build and store a certificate chain in CERT structure: returning
6443 an error if the chain cannot be built: this will allow applications
6444 to test if a chain is correctly configured.
6446 Note: if the CERT based stores are not set then the parent SSL_CTX
6447 store is used to retain compatibility with existing behaviour.
6451 * New function ssl_set_client_disabled to set a ciphersuite disabled
6452 mask based on the current session, check mask when sending client
6453 hello and checking the requested ciphersuite.
6457 * New ctrls to retrieve and set certificate types in a certificate
6458 request message. Print out received values in s_client. If certificate
6459 types is not set with custom values set sensible values based on
6460 supported signature algorithms.
6464 * Support for distinct client and server supported signature algorithms.
6468 * Add certificate callback. If set this is called whenever a certificate
6469 is required by client or server. An application can decide which
6470 certificate chain to present based on arbitrary criteria: for example
6471 supported signature algorithms. Add very simple example to s_server.
6472 This fixes many of the problems and restrictions of the existing client
6473 certificate callback: for example you can now clear an existing
6474 certificate and specify the whole chain.
6478 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6479 the certificate can be used for (if anything). Set valid_flags field
6480 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6481 to have similar checks in it.
6483 Add new "cert_flags" field to CERT structure and include a "strict mode".
6484 This enforces some TLS certificate requirements (such as only permitting
6485 certificate signature algorithms contained in the supported algorithms
6486 extension) which some implementations ignore: this option should be used
6487 with caution as it could cause interoperability issues.
6491 * Update and tidy signature algorithm extension processing. Work out
6492 shared signature algorithms based on preferences and peer algorithms
6493 and print them out in s_client and s_server. Abort handshake if no
6494 shared signature algorithms.
6498 * Add new functions to allow customised supported signature algorithms
6499 for SSL and SSL_CTX structures. Add options to s_client and s_server
6504 * New function SSL_certs_clear() to delete all references to certificates
6505 from an SSL structure. Before this once a certificate had been added
6506 it couldn't be removed.
6510 * Integrate hostname, email address and IP address checking with certificate
6511 verification. New verify options supporting checking in openssl utility.
6515 * Fixes and wildcard matching support to hostname and email checking
6516 functions. Add manual page.
6518 *Florian Weimer (Red Hat Product Security Team)*
6520 * New functions to check a hostname email or IP address against a
6521 certificate. Add options x509 utility to print results of checks against
6526 * Fix OCSP checking.
6528 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6530 * Initial experimental support for explicitly trusted non-root CAs.
6531 OpenSSL still tries to build a complete chain to a root but if an
6532 intermediate CA has a trust setting included that is used. The first
6533 setting is used: whether to trust (e.g., -addtrust option to the x509
6538 * Add -trusted_first option which attempts to find certificates in the
6539 trusted store even if an untrusted chain is also supplied.
6543 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6544 platform support for Linux and Android.
6548 * Support for linux-x32, ILP32 environment in x86_64 framework.
6552 * Experimental multi-implementation support for FIPS capable OpenSSL.
6553 When in FIPS mode the approved implementations are used as normal,
6554 when not in FIPS mode the internal unapproved versions are used instead.
6555 This means that the FIPS capable OpenSSL isn't forced to use the
6556 (often lower performance) FIPS implementations outside FIPS mode.
6560 * Transparently support X9.42 DH parameters when calling
6561 PEM_read_bio_DHparameters. This means existing applications can handle
6562 the new parameter format automatically.
6566 * Initial experimental support for X9.42 DH parameter format: mainly
6567 to support use of 'q' parameter for RFC5114 parameters.
6571 * Add DH parameters from RFC5114 including test data to dhtest.
6575 * Support for automatic EC temporary key parameter selection. If enabled
6576 the most preferred EC parameters are automatically used instead of
6577 hardcoded fixed parameters. Now a server just has to call:
6578 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6579 support ECDH and use the most appropriate parameters.
6583 * Enhance and tidy EC curve and point format TLS extension code. Use
6584 static structures instead of allocation if default values are used.
6585 New ctrls to set curves we wish to support and to retrieve shared curves.
6586 Print out shared curves in s_server. New options to s_server and s_client
6587 to set list of supported curves.
6591 * New ctrls to retrieve supported signature algorithms and
6592 supported curve values as an array of NIDs. Extend openssl utility
6593 to print out received values.
6597 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6598 between NIDs and the more common NIST names such as "P-256". Enhance
6599 ecparam utility and ECC method to recognise the NIST names for curves.
6603 * Enhance SSL/TLS certificate chain handling to support different
6604 chains for each certificate instead of one chain in the parent SSL_CTX.
6608 * Support for fixed DH ciphersuite client authentication: where both
6609 server and client use DH certificates with common parameters.
6613 * Support for fixed DH ciphersuites: those requiring DH server
6618 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6620 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6621 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6622 X509_CINF_get_signature were reverted post internal team review.
6627 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6629 * OCSP Status Request extension unbounded memory growth
6631 A malicious client can send an excessively large OCSP Status Request
6632 extension. If that client continually requests renegotiation, sending a
6633 large OCSP Status Request extension each time, then there will be unbounded
6634 memory growth on the server. This will eventually lead to a Denial Of
6635 Service attack through memory exhaustion. Servers with a default
6636 configuration are vulnerable even if they do not support OCSP. Builds using
6637 the "no-ocsp" build time option are not affected.
6639 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6644 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6647 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6653 * OOB write in MDC2_Update()
6655 An overflow can occur in MDC2_Update() either if called directly or
6656 through the EVP_DigestUpdate() function using MDC2. If an attacker
6657 is able to supply very large amounts of input data after a previous
6658 call to EVP_EncryptUpdate() with a partial block then a length check
6659 can overflow resulting in a heap corruption.
6661 The amount of data needed is comparable to SIZE_MAX which is impractical
6664 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6669 * Malformed SHA512 ticket DoS
6671 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6672 DoS attack where a malformed ticket will result in an OOB read which will
6675 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6676 a custom server callback and ticket lookup mechanism.
6678 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6683 * OOB write in BN_bn2dec()
6685 The function BN_bn2dec() does not check the return value of BN_div_word().
6686 This can cause an OOB write if an application uses this function with an
6687 overly large BIGNUM. This could be a problem if an overly large certificate
6688 or CRL is printed out from an untrusted source. TLS is not affected because
6689 record limits will reject an oversized certificate before it is parsed.
6691 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6696 * OOB read in TS_OBJ_print_bio()
6698 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6699 the total length the OID text representation would use and not the amount
6700 of data written. This will result in OOB reads when large OIDs are
6703 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6708 * Pointer arithmetic undefined behaviour
6710 Avoid some undefined pointer arithmetic
6712 A common idiom in the codebase is to check limits in the following manner:
6715 Where "p" points to some malloc'd data of SIZE bytes and
6718 "len" here could be from some externally supplied data (e.g. from a TLS
6721 The rules of C pointer arithmetic are such that "p + len" is only well
6722 defined where len <= SIZE. Therefore, the above idiom is actually
6723 undefined behaviour.
6725 For example this could cause problems if some malloc implementation
6726 provides an address for "p" such that "p + len" actually overflows for
6727 values of len that are too big and therefore p + len < limit.
6729 This issue was reported to OpenSSL by Guido Vranken
6734 * Constant time flag not preserved in DSA signing
6736 Operations in the DSA signing algorithm should run in constant time in
6737 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6738 implementation means that a non-constant time codepath is followed for
6739 certain operations. This has been demonstrated through a cache-timing
6740 attack to be sufficient for an attacker to recover the private DSA key.
6742 This issue was reported by César Pereida (Aalto University), Billy Brumley
6743 (Tampere University of Technology), and Yuval Yarom (The University of
6744 Adelaide and NICTA).
6749 * DTLS buffered message DoS
6751 In a DTLS connection where handshake messages are delivered out-of-order
6752 those messages that OpenSSL is not yet ready to process will be buffered
6753 for later use. Under certain circumstances, a flaw in the logic means that
6754 those messages do not get removed from the buffer even though the handshake
6755 has been completed. An attacker could force up to approx. 15 messages to
6756 remain in the buffer when they are no longer required. These messages will
6757 be cleared when the DTLS connection is closed. The default maximum size for
6758 a message is 100k. Therefore, the attacker could force an additional 1500k
6759 to be consumed per connection. By opening many simultaneous connections an
6760 attacker could cause a DoS attack through memory exhaustion.
6762 This issue was reported to OpenSSL by Quan Luo.
6767 * DTLS replay protection DoS
6769 A flaw in the DTLS replay attack protection mechanism means that records
6770 that arrive for future epochs update the replay protection "window" before
6771 the MAC for the record has been validated. This could be exploited by an
6772 attacker by sending a record for the next epoch (which does not have to
6773 decrypt or have a valid MAC), with a very large sequence number. This means
6774 that all subsequent legitimate packets are dropped causing a denial of
6775 service for a specific DTLS connection.
6777 This issue was reported to OpenSSL by the OCAP audit team.
6782 * Certificate message OOB reads
6784 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6785 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6786 theoretical DoS risk but this has not been observed in practice on common
6789 The messages affected are client certificate, client certificate request
6790 and server certificate. As a result the attack can only be performed
6791 against a client or a server which enables client authentication.
6793 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6798 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6800 * Prevent padding oracle in AES-NI CBC MAC check
6802 A MITM attacker can use a padding oracle attack to decrypt traffic
6803 when the connection uses an AES CBC cipher and the server support
6806 This issue was introduced as part of the fix for Lucky 13 padding
6807 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6808 constant time by making sure that always the same bytes are read and
6809 compared against either the MAC or padding bytes. But it no longer
6810 checked that there was enough data to have both the MAC and padding
6813 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6818 * Fix EVP_EncodeUpdate overflow
6820 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6821 Base64 encoding of binary data. If an attacker is able to supply very large
6822 amounts of input data then a length check can overflow resulting in a heap
6825 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6826 the `PEM_write_bio*` family of functions. These are mainly used within the
6827 OpenSSL command line applications, so any application which processes data
6828 from an untrusted source and outputs it as a PEM file should be considered
6829 vulnerable to this issue. User applications that call these APIs directly
6830 with large amounts of untrusted data may also be vulnerable.
6832 This issue was reported by Guido Vranken.
6837 * Fix EVP_EncryptUpdate overflow
6839 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6840 is able to supply very large amounts of input data after a previous call to
6841 EVP_EncryptUpdate() with a partial block then a length check can overflow
6842 resulting in a heap corruption. Following an analysis of all OpenSSL
6843 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6844 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6845 the first called function after an EVP_EncryptInit(), and therefore that
6846 specific call must be safe. The second form is where the length passed to
6847 EVP_EncryptUpdate() can be seen from the code to be some small value and
6848 therefore there is no possibility of an overflow. Since all instances are
6849 one of these two forms, it is believed that there can be no overflows in
6850 internal code due to this problem. It should be noted that
6851 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6852 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6853 of these calls have also been analysed too and it is believed there are no
6854 instances in internal usage where an overflow could occur.
6856 This issue was reported by Guido Vranken.
6861 * Prevent ASN.1 BIO excessive memory allocation
6863 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6864 a short invalid encoding can casuse allocation of large amounts of memory
6865 potentially consuming excessive resources or exhausting memory.
6867 Any application parsing untrusted data through d2i BIO functions is
6868 affected. The memory based functions such as d2i_X509() are *not* affected.
6869 Since the memory based functions are used by the TLS library, TLS
6870 applications are not affected.
6872 This issue was reported by Brian Carpenter.
6879 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6880 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6881 in arbitrary stack data being returned in the buffer.
6883 This issue was reported by Guido Vranken.
6888 * Modify behavior of ALPN to invoke callback after SNI/servername
6889 callback, such that updates to the SSL_CTX affect ALPN.
6893 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6898 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6899 methods are enabled and ssl2 is disabled the methods return NULL.
6903 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6905 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6906 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6907 provide any "EXPORT" or "LOW" strength ciphers.
6911 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6912 is by default disabled at build-time. Builds that are not configured with
6913 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6914 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6915 will need to explicitly call either of:
6917 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6919 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6921 as appropriate. Even if either of those is used, or the application
6922 explicitly uses the version-specific SSLv2_method() or its client and
6923 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6924 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6925 ciphers, and SSLv2 56-bit DES are no longer available.
6930 * Fix a double-free in DSA code
6932 A double free bug was discovered when OpenSSL parses malformed DSA private
6933 keys and could lead to a DoS attack or memory corruption for applications
6934 that receive DSA private keys from untrusted sources. This scenario is
6937 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6943 * Disable SRP fake user seed to address a server memory leak.
6945 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6947 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6948 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6949 was changed to ignore the "fake user" SRP seed, even if the seed
6952 Users should use SRP_VBASE_get1_by_user instead. Note that in
6953 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6954 also that even though configuring the SRP seed attempts to hide
6955 invalid usernames by continuing the handshake with fake
6956 credentials, this behaviour is not constant time and no strong
6957 guarantees are made that the handshake is indistinguishable from
6958 that of a valid user.
6963 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6965 In the BN_hex2bn function the number of hex digits is calculated using an
6966 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6967 large values of `i` this can result in `bn_expand` not allocating any
6968 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6969 field as NULL leading to a subsequent NULL ptr deref. For very large values
6970 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6971 In this case memory is allocated to the internal BIGNUM data field, but it
6972 is insufficiently sized leading to heap corruption. A similar issue exists
6973 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6974 is ever called by user applications with very large untrusted hex/dec data.
6975 This is anticipated to be a rare occurrence.
6977 All OpenSSL internal usage of these functions use data that is not expected
6978 to be untrusted, e.g. config file data or application command line
6979 arguments. If user developed applications generate config file data based
6980 on untrusted data then it is possible that this could also lead to security
6981 consequences. This is also anticipated to be rare.
6983 This issue was reported to OpenSSL by Guido Vranken.
6988 * Fix memory issues in `BIO_*printf` functions
6990 The internal `fmtstr` function used in processing a "%s" format string in
6991 the `BIO_*printf` functions could overflow while calculating the length of a
6992 string and cause an OOB read when printing very long strings.
6994 Additionally the internal `doapr_outch` function can attempt to write to an
6995 OOB memory location (at an offset from the NULL pointer) in the event of a
6996 memory allocation failure. In 1.0.2 and below this could be caused where
6997 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6998 could be in processing a very long "%s" format string. Memory leaks can
7001 The first issue may mask the second issue dependent on compiler behaviour.
7002 These problems could enable attacks where large amounts of untrusted data
7003 is passed to the `BIO_*printf` functions. If applications use these functions
7004 in this way then they could be vulnerable. OpenSSL itself uses these
7005 functions when printing out human-readable dumps of ASN.1 data. Therefore
7006 applications that print this data could be vulnerable if the data is from
7007 untrusted sources. OpenSSL command line applications could also be
7008 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7009 as command line arguments.
7011 Libssl is not considered directly vulnerable. Additionally certificates etc
7012 received via remote connections via libssl are also unlikely to be able to
7013 trigger these issues because of message size limits enforced within libssl.
7015 This issue was reported to OpenSSL Guido Vranken.
7020 * Side channel attack on modular exponentiation
7022 A side-channel attack was found which makes use of cache-bank conflicts on
7023 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7024 of RSA keys. The ability to exploit this issue is limited as it relies on
7025 an attacker who has control of code in a thread running on the same
7026 hyper-threaded core as the victim thread which is performing decryptions.
7028 This issue was reported to OpenSSL by Yuval Yarom, The University of
7029 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7030 Nadia Heninger, University of Pennsylvania with more information at
7031 <http://cachebleed.info>.
7036 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7037 if no keysize is specified with default_bits. This fixes an
7038 omission in an earlier change that changed all RSA/DSA key generation
7039 commands to use 2048 bits by default.
7043 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7045 * Protection for DH small subgroup attacks
7047 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7048 switched on by default and cannot be disabled. This could have some
7053 * SSLv2 doesn't block disabled ciphers
7055 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7056 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7057 been disabled, provided that the SSLv2 protocol was not also disabled via
7060 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7061 and Sebastian Schinzel.
7066 * Reject DH handshakes with parameters shorter than 1024 bits.
7070 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7072 * Certificate verify crash with missing PSS parameter
7074 The signature verification routines will crash with a NULL pointer
7075 dereference if presented with an ASN.1 signature using the RSA PSS
7076 algorithm and absent mask generation function parameter. Since these
7077 routines are used to verify certificate signature algorithms this can be
7078 used to crash any certificate verification operation and exploited in a
7079 DoS attack. Any application which performs certificate verification is
7080 vulnerable including OpenSSL clients and servers which enable client
7083 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7088 * X509_ATTRIBUTE memory leak
7090 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7091 memory. This structure is used by the PKCS#7 and CMS routines so any
7092 application which reads PKCS#7 or CMS data from untrusted sources is
7093 affected. SSL/TLS is not affected.
7095 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7101 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7102 This changes the decoding behaviour for some invalid messages,
7103 though the change is mostly in the more lenient direction, and
7104 legacy behaviour is preserved as much as possible.
7108 * In DSA_generate_parameters_ex, if the provided seed is too short,
7109 use a random seed, as already documented.
7111 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7113 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7115 * Alternate chains certificate forgery
7117 During certificate verfification, OpenSSL will attempt to find an
7118 alternative certificate chain if the first attempt to build such a chain
7119 fails. An error in the implementation of this logic can mean that an
7120 attacker could cause certain checks on untrusted certificates to be
7121 bypassed, such as the CA flag, enabling them to use a valid leaf
7122 certificate to act as a CA and "issue" an invalid certificate.
7124 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7130 * Race condition handling PSK identify hint
7132 If PSK identity hints are received by a multi-threaded client then
7133 the values are wrongly updated in the parent SSL_CTX structure. This can
7134 result in a race condition potentially leading to a double free of the
7140 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7142 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7143 incompatibility in the handling of HMAC. The previous ABI has now been
7146 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7148 * Malformed ECParameters causes infinite loop
7150 When processing an ECParameters structure OpenSSL enters an infinite loop
7151 if the curve specified is over a specially malformed binary polynomial
7154 This can be used to perform denial of service against any
7155 system which processes public keys, certificate requests or
7156 certificates. This includes TLS clients and TLS servers with
7157 client authentication enabled.
7159 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7164 * Exploitable out-of-bounds read in X509_cmp_time
7166 X509_cmp_time does not properly check the length of the ASN1_TIME
7167 string and can read a few bytes out of bounds. In addition,
7168 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7171 An attacker can use this to craft malformed certificates and CRLs of
7172 various sizes and potentially cause a segmentation fault, resulting in
7173 a DoS on applications that verify certificates or CRLs. TLS clients
7174 that verify CRLs are affected. TLS clients and servers with client
7175 authentication enabled may be affected if they use custom verification
7178 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7179 independently by Hanno Böck.
7184 * PKCS7 crash with missing EnvelopedContent
7186 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7187 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7188 with missing content and trigger a NULL pointer dereference on parsing.
7190 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7191 structures from untrusted sources are affected. OpenSSL clients and
7192 servers are not affected.
7194 This issue was reported to OpenSSL by Michal Zalewski (Google).
7199 * CMS verify infinite loop with unknown hash function
7201 When verifying a signedData message the CMS code can enter an infinite loop
7202 if presented with an unknown hash function OID. This can be used to perform
7203 denial of service against any system which verifies signedData messages using
7205 This issue was reported to OpenSSL by Johannes Bauer.
7210 * Race condition handling NewSessionTicket
7212 If a NewSessionTicket is received by a multi-threaded client when attempting to
7213 reuse a previous ticket then a race condition can occur potentially leading to
7214 a double free of the ticket data.
7219 * Reject DH handshakes with parameters shorter than 768 bits.
7221 *Kurt Roeckx and Emilia Kasper*
7223 * dhparam: generate 2048-bit parameters by default.
7225 *Kurt Roeckx and Emilia Kasper*
7227 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7229 * Segmentation fault in ASN1_TYPE_cmp fix
7231 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7232 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7233 certificate signature algorithm consistency this can be used to crash any
7234 certificate verification operation and exploited in a DoS attack. Any
7235 application which performs certificate verification is vulnerable including
7236 OpenSSL clients and servers which enable client authentication.
7241 * ASN.1 structure reuse memory corruption fix
7243 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7244 memory corruption via an invalid write. Such reuse is and has been
7245 strongly discouraged and is believed to be rare.
7247 Applications that parse structures containing CHOICE or ANY DEFINED BY
7248 components may be affected. Certificate parsing (d2i_X509 and related
7249 functions) are however not affected. OpenSSL clients and servers are
7255 * PKCS7 NULL pointer dereferences fix
7257 The PKCS#7 parsing code does not handle missing outer ContentInfo
7258 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7259 missing content and trigger a NULL pointer dereference on parsing.
7261 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7262 otherwise parse PKCS#7 structures from untrusted sources are
7263 affected. OpenSSL clients and servers are not affected.
7265 This issue was reported to OpenSSL by Michal Zalewski (Google).
7270 * DoS via reachable assert in SSLv2 servers fix
7272 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7273 servers that both support SSLv2 and enable export cipher suites by sending
7274 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7276 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7277 (OpenSSL development team).
7282 * Use After Free following d2i_ECPrivatekey error fix
7284 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7285 could cause a use after free condition. This, in turn, could cause a double
7286 free in several private key parsing functions (such as d2i_PrivateKey
7287 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7288 for applications that receive EC private keys from untrusted
7289 sources. This scenario is considered rare.
7291 This issue was discovered by the BoringSSL project and fixed in their
7297 * X509_to_X509_REQ NULL pointer deref fix
7299 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7300 the certificate key is invalid. This function is rarely used in practice.
7302 This issue was discovered by Brian Carpenter.
7307 * Removed the export ciphers from the DEFAULT ciphers
7311 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7313 * Build fixes for the Windows and OpenVMS platforms
7315 *Matt Caswell and Richard Levitte*
7317 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7319 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7320 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7321 dereference. This could lead to a Denial Of Service attack. Thanks to
7322 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7327 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7328 dtls1_buffer_record function under certain conditions. In particular this
7329 could occur if an attacker sent repeated DTLS records with the same
7330 sequence number but for the next epoch. The memory leak could be exploited
7331 by an attacker in a Denial of Service attack through memory exhaustion.
7332 Thanks to Chris Mueller for reporting this issue.
7337 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7338 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7339 method would be set to NULL which could later result in a NULL pointer
7340 dereference. Thanks to Frank Schmirler for reporting this issue.
7345 * Abort handshake if server key exchange message is omitted for ephemeral
7348 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7349 reporting this issue.
7354 * Remove non-export ephemeral RSA code on client and server. This code
7355 violated the TLS standard by allowing the use of temporary RSA keys in
7356 non-export ciphersuites and could be used by a server to effectively
7357 downgrade the RSA key length used to a value smaller than the server
7358 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7359 INRIA or reporting this issue.
7364 * Fixed issue where DH client certificates are accepted without verification.
7365 An OpenSSL server will accept a DH certificate for client authentication
7366 without the certificate verify message. This effectively allows a client to
7367 authenticate without the use of a private key. This only affects servers
7368 which trust a client certificate authority which issues certificates
7369 containing DH keys: these are extremely rare and hardly ever encountered.
7370 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7376 * Ensure that the session ID context of an SSL is updated when its
7377 SSL_CTX is updated via SSL_set_SSL_CTX.
7379 The session ID context is typically set from the parent SSL_CTX,
7380 and can vary with the CTX.
7384 * Fix various certificate fingerprint issues.
7386 By using non-DER or invalid encodings outside the signed portion of a
7387 certificate the fingerprint can be changed without breaking the signature.
7388 Although no details of the signed portion of the certificate can be changed
7389 this can cause problems with some applications: e.g. those using the
7390 certificate fingerprint for blacklists.
7392 1. Reject signatures with non zero unused bits.
7394 If the BIT STRING containing the signature has non zero unused bits reject
7395 the signature. All current signature algorithms require zero unused bits.
7397 2. Check certificate algorithm consistency.
7399 Check the AlgorithmIdentifier inside TBS matches the one in the
7400 certificate signature. NB: this will result in signature failure
7401 errors for some broken certificates.
7403 Thanks to Konrad Kraszewski from Google for reporting this issue.
7405 3. Check DSA/ECDSA signatures use DER.
7407 Re-encode DSA/ECDSA signatures and compare with the original received
7408 signature. Return an error if there is a mismatch.
7410 This will reject various cases including garbage after signature
7411 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7412 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7413 (negative or with leading zeroes).
7415 Further analysis was conducted and fixes were developed by Stephen Henson
7416 of the OpenSSL core team.
7422 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7423 results on some platforms, including x86_64. This bug occurs at random
7424 with a very low probability, and is not known to be exploitable in any
7425 way, though its exact impact is difficult to determine. Thanks to Pieter
7426 Wuille (Blockstream) who reported this issue and also suggested an initial
7427 fix. Further analysis was conducted by the OpenSSL development team and
7428 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7429 the OpenSSL core team.
7434 * Do not resume sessions on the server if the negotiated protocol
7435 version does not match the session's version. Resuming with a different
7436 version, while not strictly forbidden by the RFC, is of questionable
7437 sanity and breaks all known clients.
7439 *David Benjamin, Emilia Käsper*
7441 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7442 early CCS messages during renegotiation. (Note that because
7443 renegotiation is encrypted, this early CCS was not exploitable.)
7447 * Tighten client-side session ticket handling during renegotiation:
7448 ensure that the client only accepts a session ticket if the server sends
7449 the extension anew in the ServerHello. Previously, a TLS client would
7450 reuse the old extension state and thus accept a session ticket if one was
7451 announced in the initial ServerHello.
7453 Similarly, ensure that the client requires a session ticket if one
7454 was advertised in the ServerHello. Previously, a TLS client would
7455 ignore a missing NewSessionTicket message.
7459 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7463 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7464 sends a carefully crafted handshake message, to cause OpenSSL to fail
7465 to free up to 64k of memory causing a memory leak. This could be
7466 exploited in a Denial Of Service attack. This issue affects OpenSSL
7467 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7468 whether SRTP is used or configured. Implementations of OpenSSL that
7469 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7471 The fix was developed by the OpenSSL team.
7476 * Session Ticket Memory Leak.
7478 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7479 integrity of that ticket is first verified. In the event of a session
7480 ticket integrity check failing, OpenSSL will fail to free memory
7481 causing a memory leak. By sending a large number of invalid session
7482 tickets an attacker could exploit this issue in a Denial Of Service
7488 * Build option no-ssl3 is incomplete.
7490 When OpenSSL is configured with "no-ssl3" as a build option, servers
7491 could accept and complete an SSL 3.0 handshake, and clients could be
7492 configured to send them.
7495 *Akamai and the OpenSSL team*
7497 * Add support for TLS_FALLBACK_SCSV.
7498 Client applications doing fallback retries should call
7499 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7502 *Adam Langley, Bodo Moeller*
7504 * Add additional DigestInfo checks.
7506 Re-encode DigestInto in DER and check against the original when
7507 verifying RSA signature: this will reject any improperly encoded
7508 DigestInfo structures.
7510 Note: this is a precautionary measure and no attacks are currently known.
7514 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7516 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7517 SRP code can be overrun an internal buffer. Add sanity check that
7518 g, A, B < N to SRP code.
7520 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7521 Group for discovering this issue.
7526 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7527 TLS 1.0 instead of higher protocol versions when the ClientHello message
7528 is badly fragmented. This allows a man-in-the-middle attacker to force a
7529 downgrade to TLS 1.0 even if both the server and the client support a
7530 higher protocol version, by modifying the client's TLS records.
7532 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7533 researching this issue.
7538 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7539 to a denial of service attack. A malicious server can crash the client
7540 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7541 ciphersuite and sending carefully crafted handshake messages.
7543 Thanks to Felix Gröbert (Google) for discovering and researching this
7549 * By sending carefully crafted DTLS packets an attacker could cause openssl
7550 to leak memory. This can be exploited through a Denial of Service attack.
7551 Thanks to Adam Langley for discovering and researching this issue.
7556 * An attacker can force openssl to consume large amounts of memory whilst
7557 processing DTLS handshake messages. This can be exploited through a
7558 Denial of Service attack.
7559 Thanks to Adam Langley for discovering and researching this issue.
7564 * An attacker can force an error condition which causes openssl to crash
7565 whilst processing DTLS packets due to memory being freed twice. This
7566 can be exploited through a Denial of Service attack.
7567 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7573 * If a multithreaded client connects to a malicious server using a resumed
7574 session and the server sends an ec point format extension it could write
7575 up to 255 bytes to freed memory.
7577 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7583 * A malicious server can crash an OpenSSL client with a null pointer
7584 dereference (read) by specifying an SRP ciphersuite even though it was not
7585 properly negotiated with the client. This can be exploited through a
7586 Denial of Service attack.
7588 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7589 discovering and researching this issue.
7594 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7595 X509_name_oneline, X509_name_print_ex et al. to leak some information
7596 from the stack. Applications may be affected if they echo pretty printing
7597 output to the attacker.
7599 Thanks to Ivan Fratric (Google) for discovering this issue.
7602 *Emilia Käsper, and Steve Henson*
7604 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7605 for corner cases. (Certain input points at infinity could lead to
7606 bogus results, with non-infinity inputs mapped to infinity too.)
7610 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7612 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7613 handshake can force the use of weak keying material in OpenSSL
7614 SSL/TLS clients and servers.
7616 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7617 researching this issue. ([CVE-2014-0224])
7619 *KIKUCHI Masashi, Steve Henson*
7621 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7622 OpenSSL DTLS client the code can be made to recurse eventually crashing
7625 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7628 *Imre Rad, Steve Henson*
7630 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7631 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7632 client or server. This is potentially exploitable to run arbitrary
7633 code on a vulnerable client or server.
7635 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7637 *Jüri Aedla, Steve Henson*
7639 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7640 are subject to a denial of service attack.
7642 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7643 this issue. ([CVE-2014-3470])
7645 *Felix Gröbert, Ivan Fratric, Steve Henson*
7647 * Harmonize version and its documentation. -f flag is used to display
7650 *mancha <mancha1@zoho.com>*
7652 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7653 in i2d_ECPrivateKey.
7655 *mancha <mancha1@zoho.com>*
7657 * Fix some double frees. These are not thought to be exploitable.
7659 *mancha <mancha1@zoho.com>*
7661 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7663 * A missing bounds check in the handling of the TLS heartbeat extension
7664 can be used to reveal up to 64k of memory to a connected client or
7667 Thanks for Neel Mehta of Google Security for discovering this bug and to
7668 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7669 preparing the fix ([CVE-2014-0160])
7671 *Adam Langley, Bodo Moeller*
7673 * Fix for the attack described in the paper "Recovering OpenSSL
7674 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7675 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7676 <http://eprint.iacr.org/2014/140>
7678 Thanks to Yuval Yarom and Naomi Benger for discovering this
7679 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7681 *Yuval Yarom and Naomi Benger*
7683 * TLS pad extension: draft-agl-tls-padding-03
7685 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7686 TLS client Hello record length value would otherwise be > 255 and
7687 less that 512 pad with a dummy extension containing zeroes so it
7688 is at least 512 bytes long.
7690 *Adam Langley, Steve Henson*
7692 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7694 * Fix for TLS record tampering bug. A carefully crafted invalid
7695 handshake could crash OpenSSL with a NULL pointer exception.
7696 Thanks to Anton Johansson for reporting this issues.
7699 * Keep original DTLS digest and encryption contexts in retransmission
7700 structures so we can use the previous session parameters if they need
7701 to be resent. ([CVE-2013-6450])
7705 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7706 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7707 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7708 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7709 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7710 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7712 *Rob Stradling, Adam Langley*
7714 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7716 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7717 supporting platforms or when small records were transferred.
7719 *Andy Polyakov, Steve Henson*
7721 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7723 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7725 This addresses the flaw in CBC record processing discovered by
7726 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7727 at: <http://www.isg.rhul.ac.uk/tls/>
7729 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7730 Security Group at Royal Holloway, University of London
7731 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7732 Emilia Käsper for the initial patch.
7735 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7737 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7738 ciphersuites which can be exploited in a denial of service attack.
7739 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7740 and detecting this bug and to Wolfgang Ettlinger
7741 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7746 * Return an error when checking OCSP signatures when key is NULL.
7747 This fixes a DoS attack. ([CVE-2013-0166])
7751 * Make openssl verify return errors.
7753 *Chris Palmer <palmer@google.com> and Ben Laurie*
7755 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7756 the right response is stapled. Also change SSL_get_certificate()
7757 so it returns the certificate actually sent.
7758 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7760 *Rob Stradling <rob.stradling@comodo.com>*
7762 * Fix possible deadlock when decoding public keys.
7766 * Don't use TLS 1.0 record version number in initial client hello
7771 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7773 * Sanity check record length before skipping explicit IV in TLS
7774 1.2, 1.1 and DTLS to fix DoS attack.
7776 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7777 fuzzing as a service testing platform.
7782 * Initialise tkeylen properly when encrypting CMS messages.
7783 Thanks to Solar Designer of Openwall for reporting this issue.
7787 * In FIPS mode don't try to use composite ciphers as they are not
7792 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7794 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7795 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7796 mean any application compiled against OpenSSL 1.0.0 headers setting
7797 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7798 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7799 0x10000000L Any application which was previously compiled against
7800 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7801 will need to be recompiled as a result. Letting be results in
7802 inability to disable specifically TLS 1.1 and in client context,
7803 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7807 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7808 disable just protocol X, but all protocols above X *if* there are
7809 protocols *below* X still enabled. In more practical terms it means
7810 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7811 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7812 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7817 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7819 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7820 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7821 in CRYPTO_realloc_clean.
7823 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7824 issue and to Adam Langley <agl@chromium.org> for fixing it.
7827 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7829 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7833 * Workarounds for some broken servers that "hang" if a client hello
7834 record length exceeds 255 bytes.
7836 1. Do not use record version number > TLS 1.0 in initial client
7837 hello: some (but not all) hanging servers will now work.
7838 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7839 the number of ciphers sent in the client hello. This should be
7840 set to an even number, such as 50, for example by passing:
7841 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7842 Most broken servers should now work.
7843 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7844 TLS 1.2 client support entirely.
7848 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7852 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7854 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7855 STRING form instead of a DigestInfo.
7859 * The format used for MDC2 RSA signatures is inconsistent between EVP
7860 and the RSA_sign/RSA_verify functions. This was made more apparent when
7861 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7862 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7863 the correct format in RSA_verify so both forms transparently work.
7867 * Some servers which support TLS 1.0 can choke if we initially indicate
7868 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7869 encrypted premaster secret. As a workaround use the maximum permitted
7870 client version in client hello, this should keep such servers happy
7871 and still work with previous versions of OpenSSL.
7875 * Add support for TLS/DTLS heartbeats.
7877 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7879 * Add support for SCTP.
7881 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7883 * Improved PRNG seeding for VOS.
7885 *Paul Green <Paul.Green@stratus.com>*
7887 * Extensive assembler packs updates, most notably:
7889 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7890 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7891 - x86_64: bit-sliced AES implementation;
7892 - ARM: NEON support, contemporary platforms optimizations;
7893 - s390x: z196 support;
7894 - `*`: GHASH and GF(2^m) multiplication implementations;
7898 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7899 (removal of unnecessary code)
7901 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7903 * Add TLS key material exporter from RFC 5705.
7907 * Add DTLS-SRTP negotiation from RFC 5764.
7911 * Add Next Protocol Negotiation,
7912 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7913 disabled with a no-npn flag to config or Configure. Code donated
7916 *Adam Langley <agl@google.com> and Ben Laurie*
7918 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7919 NIST-P256, NIST-P521, with constant-time single point multiplication on
7920 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7921 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7922 Code made available under Apache License version 2.0.
7924 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7925 line to include this in your build of OpenSSL, and run "make depend" (or
7926 "make update"). This enables the following EC_METHODs:
7928 EC_GFp_nistp224_method()
7929 EC_GFp_nistp256_method()
7930 EC_GFp_nistp521_method()
7932 EC_GROUP_new_by_curve_name() will automatically use these (while
7933 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7936 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7938 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7939 all platforms. Move ssize_t definition from e_os.h to the public
7940 header file e_os2.h as it now appears in public header file cms.h
7944 * New -sigopt option to the ca, req and x509 utilities. Additional
7945 signature parameters can be passed using this option and in
7950 * Add RSA PSS signing function. This will generate and set the
7951 appropriate AlgorithmIdentifiers for PSS based on those in the
7952 corresponding EVP_MD_CTX structure. No application support yet.
7956 * Support for companion algorithm specific ASN1 signing routines.
7957 New function ASN1_item_sign_ctx() signs a pre-initialised
7958 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7959 the appropriate parameters.
7963 * Add new algorithm specific ASN1 verification initialisation function
7964 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7965 handling will be the same no matter what EVP_PKEY_METHOD is used.
7966 Add a PSS handler to support verification of PSS signatures: checked
7967 against a number of sample certificates.
7971 * Add signature printing for PSS. Add PSS OIDs.
7973 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7975 * Add algorithm specific signature printing. An individual ASN1 method
7976 can now print out signatures instead of the standard hex dump.
7978 More complex signatures (e.g. PSS) can print out more meaningful
7979 information. Include DSA version that prints out the signature
7984 * Password based recipient info support for CMS library: implementing
7989 * Split password based encryption into PBES2 and PBKDF2 functions. This
7990 neatly separates the code into cipher and PBE sections and is required
7991 for some algorithms that split PBES2 into separate pieces (such as
7992 password based CMS).
7996 * Session-handling fixes:
7997 - Fix handling of connections that are resuming with a session ID,
7998 but also support Session Tickets.
7999 - Fix a bug that suppressed issuing of a new ticket if the client
8000 presented a ticket with an expired session.
8001 - Try to set the ticket lifetime hint to something reasonable.
8002 - Make tickets shorter by excluding irrelevant information.
8003 - On the client side, don't ignore renewed tickets.
8005 *Adam Langley, Bodo Moeller (Google)*
8007 * Fix PSK session representation.
8011 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8013 This work was sponsored by Intel.
8017 * Add GCM support to TLS library. Some custom code is needed to split
8018 the IV between the fixed (from PRF) and explicit (from TLS record)
8019 portions. This adds all GCM ciphersuites supported by RFC5288 and
8020 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8021 add a special AESGCM string for GCM only.
8025 * Expand range of ctrls for AES GCM. Permit setting invocation
8026 field on decrypt and retrieval of invocation field only on encrypt.
8030 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8031 As required by RFC5289 these ciphersuites cannot be used if for
8032 versions of TLS earlier than 1.2.
8036 * For FIPS capable OpenSSL interpret a NULL default public key method
8037 as unset and return the appropriate default but do *not* set the default.
8038 This means we can return the appropriate method in applications that
8039 switch between FIPS and non-FIPS modes.
8043 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8044 ENGINE is used then we cannot handle that in the FIPS module so we
8045 keep original code iff non-FIPS operations are allowed.
8049 * Add -attime option to openssl utilities.
8051 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8053 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8057 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8058 FIPS EC methods unconditionally for now.
8062 * New build option no-ec2m to disable characteristic 2 code.
8066 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8067 all cases can be covered as some introduce binary incompatibilities.
8071 * Redirect RSA operations to FIPS module including keygen,
8072 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8076 * Add similar low-level API blocking to ciphers.
8080 * low-level digest APIs are not approved in FIPS mode: any attempt
8081 to use these will cause a fatal error. Applications that *really* want
8082 to use them can use the `private_*` version instead.
8086 * Redirect cipher operations to FIPS module for FIPS builds.
8090 * Redirect digest operations to FIPS module for FIPS builds.
8094 * Update build system to add "fips" flag which will link in fipscanister.o
8095 for static and shared library builds embedding a signature if needed.
8099 * Output TLS supported curves in preference order instead of numerical
8100 order. This is currently hardcoded for the highest order curves first.
8101 This should be configurable so applications can judge speed vs strength.
8105 * Add TLS v1.2 server support for client authentication.
8109 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8114 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8115 FIPS modules versions.
8119 * Add TLS v1.2 client side support for client authentication. Keep cache
8120 of handshake records longer as we don't know the hash algorithm to use
8121 until after the certificate request message is received.
8125 * Initial TLS v1.2 client support. Add a default signature algorithms
8126 extension including all the algorithms we support. Parse new signature
8127 format in client key exchange. Relax some ECC signing restrictions for
8128 TLS v1.2 as indicated in RFC5246.
8132 * Add server support for TLS v1.2 signature algorithms extension. Switch
8133 to new signature format when needed using client digest preference.
8134 All server ciphersuites should now work correctly in TLS v1.2. No client
8135 support yet and no support for client certificates.
8139 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8140 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8141 ciphersuites. At present only RSA key exchange ciphersuites work with
8142 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8143 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8144 and version checking.
8148 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8149 with this defined it will not be affected by any changes to ssl internal
8150 structures. Add several utility functions to allow openssl application
8151 to work with OPENSSL_NO_SSL_INTERN defined.
8155 * A long standing patch to add support for SRP from EdelWeb (Peter
8156 Sylvester and Christophe Renou) was integrated.
8157 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8158 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8161 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8165 * Permit abbreviated handshakes when renegotiating using the function
8166 SSL_renegotiate_abbreviated().
8168 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8170 * Add call to ENGINE_register_all_complete() to
8171 ENGINE_load_builtin_engines(), so some implementations get used
8172 automatically instead of needing explicit application support.
8176 * Add support for TLS key exporter as described in RFC5705.
8178 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8180 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8181 a few changes are required:
8183 Add SSL_OP_NO_TLSv1_1 flag.
8184 Add TLSv1_1 methods.
8185 Update version checking logic to handle version 1.1.
8186 Add explicit IV handling (ported from DTLS code).
8187 Add command line options to s_client/s_server.
8194 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8196 * X509_ATTRIBUTE memory leak
8198 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8199 memory. This structure is used by the PKCS#7 and CMS routines so any
8200 application which reads PKCS#7 or CMS data from untrusted sources is
8201 affected. SSL/TLS is not affected.
8203 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8209 * Race condition handling PSK identify hint
8211 If PSK identity hints are received by a multi-threaded client then
8212 the values are wrongly updated in the parent SSL_CTX structure. This can
8213 result in a race condition potentially leading to a double free of the
8219 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8221 * Malformed ECParameters causes infinite loop
8223 When processing an ECParameters structure OpenSSL enters an infinite loop
8224 if the curve specified is over a specially malformed binary polynomial
8227 This can be used to perform denial of service against any
8228 system which processes public keys, certificate requests or
8229 certificates. This includes TLS clients and TLS servers with
8230 client authentication enabled.
8232 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8237 * Exploitable out-of-bounds read in X509_cmp_time
8239 X509_cmp_time does not properly check the length of the ASN1_TIME
8240 string and can read a few bytes out of bounds. In addition,
8241 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8244 An attacker can use this to craft malformed certificates and CRLs of
8245 various sizes and potentially cause a segmentation fault, resulting in
8246 a DoS on applications that verify certificates or CRLs. TLS clients
8247 that verify CRLs are affected. TLS clients and servers with client
8248 authentication enabled may be affected if they use custom verification
8251 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8252 independently by Hanno Böck.
8257 * PKCS7 crash with missing EnvelopedContent
8259 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8260 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8261 with missing content and trigger a NULL pointer dereference on parsing.
8263 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8264 structures from untrusted sources are affected. OpenSSL clients and
8265 servers are not affected.
8267 This issue was reported to OpenSSL by Michal Zalewski (Google).
8272 * CMS verify infinite loop with unknown hash function
8274 When verifying a signedData message the CMS code can enter an infinite loop
8275 if presented with an unknown hash function OID. This can be used to perform
8276 denial of service against any system which verifies signedData messages using
8278 This issue was reported to OpenSSL by Johannes Bauer.
8283 * Race condition handling NewSessionTicket
8285 If a NewSessionTicket is received by a multi-threaded client when attempting to
8286 reuse a previous ticket then a race condition can occur potentially leading to
8287 a double free of the ticket data.
8292 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8294 * Segmentation fault in ASN1_TYPE_cmp fix
8296 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8297 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8298 certificate signature algorithm consistency this can be used to crash any
8299 certificate verification operation and exploited in a DoS attack. Any
8300 application which performs certificate verification is vulnerable including
8301 OpenSSL clients and servers which enable client authentication.
8306 * ASN.1 structure reuse memory corruption fix
8308 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8309 memory corruption via an invalid write. Such reuse is and has been
8310 strongly discouraged and is believed to be rare.
8312 Applications that parse structures containing CHOICE or ANY DEFINED BY
8313 components may be affected. Certificate parsing (d2i_X509 and related
8314 functions) are however not affected. OpenSSL clients and servers are
8320 * PKCS7 NULL pointer dereferences fix
8322 The PKCS#7 parsing code does not handle missing outer ContentInfo
8323 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8324 missing content and trigger a NULL pointer dereference on parsing.
8326 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8327 otherwise parse PKCS#7 structures from untrusted sources are
8328 affected. OpenSSL clients and servers are not affected.
8330 This issue was reported to OpenSSL by Michal Zalewski (Google).
8335 * DoS via reachable assert in SSLv2 servers fix
8337 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8338 servers that both support SSLv2 and enable export cipher suites by sending
8339 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8341 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8342 (OpenSSL development team).
8347 * Use After Free following d2i_ECPrivatekey error fix
8349 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8350 could cause a use after free condition. This, in turn, could cause a double
8351 free in several private key parsing functions (such as d2i_PrivateKey
8352 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8353 for applications that receive EC private keys from untrusted
8354 sources. This scenario is considered rare.
8356 This issue was discovered by the BoringSSL project and fixed in their
8362 * X509_to_X509_REQ NULL pointer deref fix
8364 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8365 the certificate key is invalid. This function is rarely used in practice.
8367 This issue was discovered by Brian Carpenter.
8372 * Removed the export ciphers from the DEFAULT ciphers
8376 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8378 * Build fixes for the Windows and OpenVMS platforms
8380 *Matt Caswell and Richard Levitte*
8382 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8384 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8385 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8386 dereference. This could lead to a Denial Of Service attack. Thanks to
8387 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8392 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8393 dtls1_buffer_record function under certain conditions. In particular this
8394 could occur if an attacker sent repeated DTLS records with the same
8395 sequence number but for the next epoch. The memory leak could be exploited
8396 by an attacker in a Denial of Service attack through memory exhaustion.
8397 Thanks to Chris Mueller for reporting this issue.
8402 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8403 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8404 method would be set to NULL which could later result in a NULL pointer
8405 dereference. Thanks to Frank Schmirler for reporting this issue.
8410 * Abort handshake if server key exchange message is omitted for ephemeral
8413 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8414 reporting this issue.
8419 * Remove non-export ephemeral RSA code on client and server. This code
8420 violated the TLS standard by allowing the use of temporary RSA keys in
8421 non-export ciphersuites and could be used by a server to effectively
8422 downgrade the RSA key length used to a value smaller than the server
8423 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8424 INRIA or reporting this issue.
8429 * Fixed issue where DH client certificates are accepted without verification.
8430 An OpenSSL server will accept a DH certificate for client authentication
8431 without the certificate verify message. This effectively allows a client to
8432 authenticate without the use of a private key. This only affects servers
8433 which trust a client certificate authority which issues certificates
8434 containing DH keys: these are extremely rare and hardly ever encountered.
8435 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8441 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8442 results on some platforms, including x86_64. This bug occurs at random
8443 with a very low probability, and is not known to be exploitable in any
8444 way, though its exact impact is difficult to determine. Thanks to Pieter
8445 Wuille (Blockstream) who reported this issue and also suggested an initial
8446 fix. Further analysis was conducted by the OpenSSL development team and
8447 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8448 the OpenSSL core team.
8453 * Fix various certificate fingerprint issues.
8455 By using non-DER or invalid encodings outside the signed portion of a
8456 certificate the fingerprint can be changed without breaking the signature.
8457 Although no details of the signed portion of the certificate can be changed
8458 this can cause problems with some applications: e.g. those using the
8459 certificate fingerprint for blacklists.
8461 1. Reject signatures with non zero unused bits.
8463 If the BIT STRING containing the signature has non zero unused bits reject
8464 the signature. All current signature algorithms require zero unused bits.
8466 2. Check certificate algorithm consistency.
8468 Check the AlgorithmIdentifier inside TBS matches the one in the
8469 certificate signature. NB: this will result in signature failure
8470 errors for some broken certificates.
8472 Thanks to Konrad Kraszewski from Google for reporting this issue.
8474 3. Check DSA/ECDSA signatures use DER.
8476 Reencode DSA/ECDSA signatures and compare with the original received
8477 signature. Return an error if there is a mismatch.
8479 This will reject various cases including garbage after signature
8480 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8481 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8482 (negative or with leading zeroes).
8484 Further analysis was conducted and fixes were developed by Stephen Henson
8485 of the OpenSSL core team.
8491 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8493 * Session Ticket Memory Leak.
8495 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8496 integrity of that ticket is first verified. In the event of a session
8497 ticket integrity check failing, OpenSSL will fail to free memory
8498 causing a memory leak. By sending a large number of invalid session
8499 tickets an attacker could exploit this issue in a Denial Of Service
8505 * Build option no-ssl3 is incomplete.
8507 When OpenSSL is configured with "no-ssl3" as a build option, servers
8508 could accept and complete an SSL 3.0 handshake, and clients could be
8509 configured to send them.
8512 *Akamai and the OpenSSL team*
8514 * Add support for TLS_FALLBACK_SCSV.
8515 Client applications doing fallback retries should call
8516 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8519 *Adam Langley, Bodo Moeller*
8521 * Add additional DigestInfo checks.
8523 Reencode DigestInto in DER and check against the original when
8524 verifying RSA signature: this will reject any improperly encoded
8525 DigestInfo structures.
8527 Note: this is a precautionary measure and no attacks are currently known.
8531 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8533 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8534 to a denial of service attack. A malicious server can crash the client
8535 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8536 ciphersuite and sending carefully crafted handshake messages.
8538 Thanks to Felix Gröbert (Google) for discovering and researching this
8544 * By sending carefully crafted DTLS packets an attacker could cause openssl
8545 to leak memory. This can be exploited through a Denial of Service attack.
8546 Thanks to Adam Langley for discovering and researching this issue.
8551 * An attacker can force openssl to consume large amounts of memory whilst
8552 processing DTLS handshake messages. This can be exploited through a
8553 Denial of Service attack.
8554 Thanks to Adam Langley for discovering and researching this issue.
8559 * An attacker can force an error condition which causes openssl to crash
8560 whilst processing DTLS packets due to memory being freed twice. This
8561 can be exploited through a Denial of Service attack.
8562 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8568 * If a multithreaded client connects to a malicious server using a resumed
8569 session and the server sends an ec point format extension it could write
8570 up to 255 bytes to freed memory.
8572 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8578 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8579 X509_name_oneline, X509_name_print_ex et al. to leak some information
8580 from the stack. Applications may be affected if they echo pretty printing
8581 output to the attacker.
8583 Thanks to Ivan Fratric (Google) for discovering this issue.
8586 *Emilia Käsper, and Steve Henson*
8588 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8589 for corner cases. (Certain input points at infinity could lead to
8590 bogus results, with non-infinity inputs mapped to infinity too.)
8594 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8596 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8597 handshake can force the use of weak keying material in OpenSSL
8598 SSL/TLS clients and servers.
8600 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8601 researching this issue. ([CVE-2014-0224])
8603 *KIKUCHI Masashi, Steve Henson*
8605 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8606 OpenSSL DTLS client the code can be made to recurse eventually crashing
8609 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8612 *Imre Rad, Steve Henson*
8614 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8615 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8616 client or server. This is potentially exploitable to run arbitrary
8617 code on a vulnerable client or server.
8619 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8621 *Jüri Aedla, Steve Henson*
8623 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8624 are subject to a denial of service attack.
8626 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8627 this issue. ([CVE-2014-3470])
8629 *Felix Gröbert, Ivan Fratric, Steve Henson*
8631 * Harmonize version and its documentation. -f flag is used to display
8634 *mancha <mancha1@zoho.com>*
8636 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8637 in i2d_ECPrivateKey.
8639 *mancha <mancha1@zoho.com>*
8641 * Fix some double frees. These are not thought to be exploitable.
8643 *mancha <mancha1@zoho.com>*
8645 * Fix for the attack described in the paper "Recovering OpenSSL
8646 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8647 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8648 <http://eprint.iacr.org/2014/140>
8650 Thanks to Yuval Yarom and Naomi Benger for discovering this
8651 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8653 *Yuval Yarom and Naomi Benger*
8655 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8657 * Keep original DTLS digest and encryption contexts in retransmission
8658 structures so we can use the previous session parameters if they need
8659 to be resent. ([CVE-2013-6450])
8663 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8664 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8665 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8666 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8667 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8668 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8670 *Rob Stradling, Adam Langley*
8672 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8674 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8676 This addresses the flaw in CBC record processing discovered by
8677 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8678 at: <http://www.isg.rhul.ac.uk/tls/>
8680 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8681 Security Group at Royal Holloway, University of London
8682 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8683 Emilia Käsper for the initial patch.
8686 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8688 * Return an error when checking OCSP signatures when key is NULL.
8689 This fixes a DoS attack. ([CVE-2013-0166])
8693 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8694 the right response is stapled. Also change SSL_get_certificate()
8695 so it returns the certificate actually sent.
8696 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8697 (This is a backport)
8699 *Rob Stradling <rob.stradling@comodo.com>*
8701 * Fix possible deadlock when decoding public keys.
8705 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8707 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8710 * Sanity check record length before skipping explicit IV in DTLS
8713 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8714 fuzzing as a service testing platform.
8719 * Initialise tkeylen properly when encrypting CMS messages.
8720 Thanks to Solar Designer of Openwall for reporting this issue.
8724 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8726 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8727 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8728 in CRYPTO_realloc_clean.
8730 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8731 issue and to Adam Langley <agl@chromium.org> for fixing it.
8734 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8736 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8738 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8739 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8740 content decryption and always return the same error. Note: this attack
8741 needs on average 2^20 messages so it only affects automated senders. The
8742 old behaviour can be re-enabled in the CMS code by setting the
8743 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8744 an MMA defence is not necessary.
8745 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8746 this issue. ([CVE-2012-0884])
8750 * Fix CVE-2011-4619: make sure we really are receiving a
8751 client hello before rejecting multiple SGC restarts. Thanks to
8752 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8756 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8758 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8759 Thanks to Antonio Martin, Enterprise Secure Access Research and
8760 Development, Cisco Systems, Inc. for discovering this bug and
8761 preparing a fix. ([CVE-2012-0050])
8765 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8767 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8768 of the Vaudenay padding oracle attack on CBC mode encryption
8769 which enables an efficient plaintext recovery attack against
8770 the OpenSSL implementation of DTLS. Their attack exploits timing
8771 differences arising during decryption processing. A research
8772 paper describing this attack can be found at:
8773 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8774 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8775 Security Group at Royal Holloway, University of London
8776 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8777 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8778 for preparing the fix. ([CVE-2011-4108])
8780 *Robin Seggelmann, Michael Tuexen*
8782 * Clear bytes used for block padding of SSL 3.0 records.
8785 *Adam Langley (Google)*
8787 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8788 Kadianakis <desnacked@gmail.com> for discovering this issue and
8789 Adam Langley for preparing the fix. ([CVE-2011-4619])
8791 *Adam Langley (Google)*
8793 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8795 *Andrey Kulikov <amdeich@gmail.com>*
8797 * Prevent malformed RFC3779 data triggering an assertion failure.
8798 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8799 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8801 *Rob Austein <sra@hactrn.net>*
8803 * Improved PRNG seeding for VOS.
8805 *Paul Green <Paul.Green@stratus.com>*
8807 * Fix ssl_ciph.c set-up race.
8809 *Adam Langley (Google)*
8811 * Fix spurious failures in ecdsatest.c.
8813 *Emilia Käsper (Google)*
8815 * Fix the BIO_f_buffer() implementation (which was mixing different
8816 interpretations of the `..._len` fields).
8818 *Adam Langley (Google)*
8820 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8821 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8822 threads won't reuse the same blinding coefficients.
8824 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8825 lock to call BN_BLINDING_invert_ex, and avoids one use of
8826 BN_BLINDING_update for each BN_BLINDING structure (previously,
8827 the last update always remained unused).
8829 *Emilia Käsper (Google)*
8831 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8833 *Bob Buckholz (Google)*
8835 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8837 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8838 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8840 *Kaspar Brand <ossl@velox.ch>*
8842 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8843 for multi-threaded use of ECDH. ([CVE-2011-3210])
8845 *Adam Langley (Google)*
8847 * Fix x509_name_ex_d2i memory leak on bad inputs.
8851 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8852 signature public key algorithm by using OID xref utilities instead.
8853 Before this you could only use some ECC ciphersuites with SHA1 only.
8857 * Add protection against ECDSA timing attacks as mentioned in the paper
8858 by Billy Bob Brumley and Nicola Tuveri, see:
8859 <http://eprint.iacr.org/2011/232.pdf>
8861 *Billy Bob Brumley and Nicola Tuveri*
8863 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8865 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8867 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8869 * Fix bug in string printing code: if *any* escaping is enabled we must
8870 escape the escape character (backslash) or the resulting string is
8875 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8877 * Disable code workaround for ancient and obsolete Netscape browsers
8878 and servers: an attacker can use it in a ciphersuite downgrade attack.
8879 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8883 * Fixed J-PAKE implementation error, originally discovered by
8884 Sebastien Martini, further info and confirmation from Stefan
8885 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8889 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8891 * Fix extension code to avoid race conditions which can result in a buffer
8892 overrun vulnerability: resumed sessions must not be modified as they can
8893 be shared by multiple threads. CVE-2010-3864
8897 * Fix WIN32 build system to correctly link an ENGINE directory into
8902 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8904 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8907 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8909 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8911 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8912 context. The operation can be customised via the ctrl mechanism in
8913 case ENGINEs want to include additional functionality.
8917 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8921 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8922 output hashes compatible with older versions of OpenSSL.
8924 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8926 * Fix compression algorithm handling: if resuming a session use the
8927 compression algorithm of the resumed session instead of determining
8928 it from client hello again. Don't allow server to change algorithm.
8932 * Add load_crls() function to commands tidying load_certs() too. Add option
8933 to verify utility to allow additional CRLs to be included.
8937 * Update OCSP request code to permit adding custom headers to the request:
8938 some responders need this.
8942 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8945 *Julia Lawall <julia@diku.dk>*
8947 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8948 needlessly dereferenced structures, used obsolete functions and
8949 didn't handle all updated verify codes correctly.
8953 * Disable MD2 in the default configuration.
8957 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8958 indicate the initial BIO being pushed or popped. This makes it possible
8959 to determine whether the BIO is the one explicitly called or as a result
8960 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8961 it handles reference counts correctly and doesn't zero out the I/O bio
8962 when it is not being explicitly popped. WARNING: applications which
8963 included workarounds for the old buggy behaviour will need to be modified
8964 or they could free up already freed BIOs.
8968 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8969 renaming to all platforms (within the 0.9.8 branch, this was
8970 done conditionally on Netware platforms to avoid a name clash).
8972 *Guenter <lists@gknw.net>*
8974 * Add ECDHE and PSK support to DTLS.
8976 *Michael Tuexen <tuexen@fh-muenster.de>*
8978 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8983 * Add "missing" function EVP_MD_flags() (without this the only way to
8984 retrieve a digest flags is by accessing the structure directly. Update
8985 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8986 or cipher is registered as in the "from" argument. Print out all
8987 registered digests in the dgst usage message instead of manually
8988 attempting to work them out.
8992 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8993 this allows the use of compression and extensions. Change default cipher
8994 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8995 by default unless an application cipher string requests it.
8999 * Alter match criteria in PKCS12_parse(). It used to try to use local
9000 key ids to find matching certificates and keys but some PKCS#12 files
9001 don't follow the (somewhat unwritten) rules and this strategy fails.
9002 Now just gather all certificates together and the first private key
9003 then look for the first certificate that matches the key.
9007 * Support use of registered digest and cipher names for dgst and cipher
9008 commands instead of having to add each one as a special case. So now
9015 openssl dgst -sha256 foo
9017 and this works for ENGINE based algorithms too.
9021 * Update Gost ENGINE to support parameter files.
9023 *Victor B. Wagner <vitus@cryptocom.ru>*
9025 * Support GeneralizedTime in ca utility.
9027 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9029 * Enhance the hash format used for certificate directory links. The new
9030 form uses the canonical encoding (meaning equivalent names will work
9031 even if they aren't identical) and uses SHA1 instead of MD5. This form
9032 is incompatible with the older format and as a result c_rehash should
9033 be used to rebuild symbolic links.
9037 * Make PKCS#8 the default write format for private keys, replacing the
9038 traditional format. This form is standardised, more secure and doesn't
9039 include an implicit MD5 dependency.
9043 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9044 committed to OpenSSL should pass this lot as a minimum.
9048 * Add session ticket override functionality for use by EAP-FAST.
9050 *Jouni Malinen <j@w1.fi>*
9052 * Modify HMAC functions to return a value. Since these can be implemented
9053 in an ENGINE errors can occur.
9057 * Type-checked OBJ_bsearch_ex.
9061 * Type-checked OBJ_bsearch. Also some constification necessitated
9062 by type-checking. Still to come: TXT_DB, bsearch(?),
9063 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9068 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9069 seconds to a tm structure directly, instead of going through OS
9070 specific date routines. This avoids any issues with OS routines such
9071 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9072 and X509_time_adj_ex() to cover the extended range. The existing
9073 X509_time_adj() is still usable and will no longer have any date issues.
9077 * Delta CRL support. New use deltas option which will attempt to locate
9078 and search any appropriate delta CRLs available.
9080 This work was sponsored by Google.
9084 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9085 code and add additional score elements. Validate alternate CRL paths
9086 as part of the CRL checking and indicate a new error "CRL path validation
9087 error" in this case. Applications wanting additional details can use
9088 the verify callback and check the new "parent" field. If this is not
9089 NULL CRL path validation is taking place. Existing applications won't
9090 see this because it requires extended CRL support which is off by
9093 This work was sponsored by Google.
9097 * Support for freshest CRL extension.
9099 This work was sponsored by Google.
9103 * Initial indirect CRL support. Currently only supported in the CRLs
9104 passed directly and not via lookup. Process certificate issuer
9105 CRL entry extension and lookup CRL entries by bother issuer name
9106 and serial number. Check and process CRL issuer entry in IDP extension.
9108 This work was sponsored by Google.
9112 * Add support for distinct certificate and CRL paths. The CRL issuer
9113 certificate is validated separately in this case. Only enabled if
9114 an extended CRL support flag is set: this flag will enable additional
9115 CRL functionality in future.
9117 This work was sponsored by Google.
9121 * Add support for policy mappings extension.
9123 This work was sponsored by Google.
9127 * Fixes to pathlength constraint, self issued certificate handling,
9128 policy processing to align with RFC3280 and PKITS tests.
9130 This work was sponsored by Google.
9134 * Support for name constraints certificate extension. DN, email, DNS
9135 and URI types are currently supported.
9137 This work was sponsored by Google.
9141 * To cater for systems that provide a pointer-based thread ID rather
9142 than numeric, deprecate the current numeric thread ID mechanism and
9143 replace it with a structure and associated callback type. This
9144 mechanism allows a numeric "hash" to be extracted from a thread ID in
9145 either case, and on platforms where pointers are larger than 'long',
9146 mixing is done to help ensure the numeric 'hash' is usable even if it
9147 can't be guaranteed unique. The default mechanism is to use "&errno"
9148 as a pointer-based thread ID to distinguish between threads.
9150 Applications that want to provide their own thread IDs should now use
9151 CRYPTO_THREADID_set_callback() to register a callback that will call
9152 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9154 Note that ERR_remove_state() is now deprecated, because it is tied
9155 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9156 to free the current thread's error state should be replaced by
9157 ERR_remove_thread_state(NULL).
9159 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9160 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9161 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9162 application was previously providing a numeric thread callback that
9163 was inappropriate for distinguishing threads, then uniqueness might
9164 have been obtained with &errno that happened immediately in the
9165 intermediate development versions of OpenSSL; this is no longer the
9166 case, the numeric thread callback will now override the automatic use
9169 *Geoff Thorpe, with help from Bodo Moeller*
9171 * Initial support for different CRL issuing certificates. This covers a
9172 simple case where the self issued certificates in the chain exist and
9173 the real CRL issuer is higher in the existing chain.
9175 This work was sponsored by Google.
9179 * Removed effectively defunct crypto/store from the build.
9183 * Revamp of STACK to provide stronger type-checking. Still to come:
9184 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9185 ASN1_STRING, CONF_VALUE.
9189 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9190 RAM on SSL connections. This option can save about 34k per idle SSL.
9194 * Revamp of LHASH to provide stronger type-checking. Still to come:
9195 STACK, TXT_DB, bsearch, qsort.
9199 * Initial support for Cryptographic Message Syntax (aka CMS) based
9200 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9201 support for data, signedData, compressedData, digestedData and
9202 encryptedData, envelopedData types included. Scripts to check against
9203 RFC4134 examples draft and interop and consistency checks of many
9204 content types and variants.
9208 * Add options to enc utility to support use of zlib compression BIO.
9212 * Extend mk1mf to support importing of options and assembly language
9213 files from Configure script, currently only included in VC-WIN32.
9214 The assembly language rules can now optionally generate the source
9215 files from the associated perl scripts.
9219 * Implement remaining functionality needed to support GOST ciphersuites.
9220 Interop testing has been performed using CryptoPro implementations.
9222 *Victor B. Wagner <vitus@cryptocom.ru>*
9224 * s390x assembler pack.
9228 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9233 * Implement Opaque PRF Input TLS extension as specified in
9234 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9235 official specification yet and no extension type assignment by
9236 IANA exists, this extension (for now) will have to be explicitly
9237 enabled when building OpenSSL by providing the extension number
9238 to use. For example, specify an option
9240 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9242 to the "config" or "Configure" script to enable the extension,
9243 assuming extension number 0x9527 (which is a completely arbitrary
9244 and unofficial assignment based on the MD5 hash of the Internet
9245 Draft). Note that by doing so, you potentially lose
9246 interoperability with other TLS implementations since these might
9247 be using the same extension number for other purposes.
9249 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9250 opaque PRF input value to use in the handshake. This will create
9251 an internal copy of the length-'len' string at 'src', and will
9252 return non-zero for success.
9254 To get more control and flexibility, provide a callback function
9257 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9258 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9262 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9265 Callback function 'cb' will be called in handshakes, and is
9266 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9267 Argument 'arg' is for application purposes (the value as given to
9268 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9269 be provided to the callback function). The callback function
9270 has to return non-zero to report success: usually 1 to use opaque
9271 PRF input just if possible, or 2 to enforce use of the opaque PRF
9272 input. In the latter case, the library will abort the handshake
9273 if opaque PRF input is not successfully negotiated.
9275 Arguments 'peerinput' and 'len' given to the callback function
9276 will always be NULL and 0 in the case of a client. A server will
9277 see the client's opaque PRF input through these variables if
9278 available (NULL and 0 otherwise). Note that if the server
9279 provides an opaque PRF input, the length must be the same as the
9280 length of the client's opaque PRF input.
9282 Note that the callback function will only be called when creating
9283 a new session (session resumption can resume whatever was
9284 previously negotiated), and will not be called in SSL 2.0
9285 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9286 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9287 for applications that need to enforce opaque PRF input.
9291 * Update ssl code to support digests other than SHA1+MD5 for handshake
9294 *Victor B. Wagner <vitus@cryptocom.ru>*
9296 * Add RFC4507 support to OpenSSL. This includes the corrections in
9297 RFC4507bis. The encrypted ticket format is an encrypted encoded
9298 SSL_SESSION structure, that way new session features are automatically
9301 If a client application caches session in an SSL_SESSION structure
9302 support is transparent because tickets are now stored in the encoded
9305 The SSL_CTX structure automatically generates keys for ticket
9306 protection in servers so again support should be possible
9307 with no application modification.
9309 If a client or server wishes to disable RFC4507 support then the option
9310 SSL_OP_NO_TICKET can be set.
9312 Add a TLS extension debugging callback to allow the contents of any client
9313 or server extensions to be examined.
9315 This work was sponsored by Google.
9319 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9320 OpenSSL should now compile cleanly on gcc 4.2
9322 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9324 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9325 support including streaming MAC support: this is required for GOST
9326 ciphersuite support.
9328 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9330 * Add option -stream to use PKCS#7 streaming in smime utility. New
9331 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9332 to output in BER and PEM format.
9336 * Experimental support for use of HMAC via EVP_PKEY interface. This
9337 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9338 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9339 ENGINE support for HMAC keys which are unextractable. New -mac and
9340 -macopt options to dgst utility.
9344 * New option -sigopt to dgst utility. Update dgst to use
9345 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9346 alternative signing parameters such as X9.31 or PSS in the dgst
9351 * Change ssl_cipher_apply_rule(), the internal function that does
9352 the work each time a ciphersuite string requests enabling
9353 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9354 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9355 the order of disabled ciphersuites such that those ciphersuites
9356 that most recently went from enabled to disabled not only stay
9357 in order with respect to each other, but also have higher priority
9358 than other disabled ciphersuites the next time ciphersuites are
9361 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9362 the same ciphersuites as with "HIGH" alone, but in a specific
9363 order where the PSK ciphersuites come first (since they are the
9364 most recently disabled ciphersuites when "HIGH" is parsed).
9366 Also, change ssl_create_cipher_list() (using this new
9367 functionality) such that between otherwise identical
9368 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9373 * Change ssl_create_cipher_list() so that it automatically
9374 arranges the ciphersuites in reasonable order before starting
9375 to process the rule string. Thus, the definition for "DEFAULT"
9376 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9377 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9378 This makes it much easier to arrive at a reasonable default order
9379 in applications for which anonymous ciphers are OK (meaning
9380 that you can't actually use DEFAULT).
9382 *Bodo Moeller; suggested by Victor Duchovni*
9384 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9385 processing) into multiple integers instead of setting
9386 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9387 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9388 (These masks as well as the individual bit definitions are hidden
9389 away into the non-exported interface ssl/ssl_locl.h, so this
9390 change to the definition of the SSL_CIPHER structure shouldn't
9391 affect applications.) This give us more bits for each of these
9392 categories, so there is no longer a need to coagulate AES128 and
9393 AES256 into a single algorithm bit, and to coagulate Camellia128
9394 and Camellia256 into a single algorithm bit, which has led to all
9397 Thus, among other things, the kludge introduced in 0.9.7m and
9398 0.9.8e for masking out AES256 independently of AES128 or masking
9399 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9401 With the change, we also introduce new ciphersuite aliases that
9402 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9407 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9408 Use the leftmost N bytes of the signature input if the input is
9409 larger than the prime q (with N being the size in bytes of q).
9413 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9414 it yet and it is largely untested.
9418 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9422 * Initial incomplete changes to avoid need for function casts in OpenSSL
9423 some compilers (gcc 4.2 and later) reject their use. Safestack is
9424 reimplemented. Update ASN1 to avoid use of legacy functions.
9428 * Win32/64 targets are linked with Winsock2.
9432 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9433 to external functions. This can be used to increase CRL handling
9434 efficiency especially when CRLs are very large by (for example) storing
9435 the CRL revoked certificates in a database.
9439 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9440 new CRLs added to a directory can be used. New command line option
9441 -verify_return_error to s_client and s_server. This causes real errors
9442 to be returned by the verify callback instead of carrying on no matter
9443 what. This reflects the way a "real world" verify callback would behave.
9447 * GOST engine, supporting several GOST algorithms and public key formats.
9448 Kindly donated by Cryptocom.
9452 * Partial support for Issuing Distribution Point CRL extension. CRLs
9453 partitioned by DP are handled but no indirect CRL or reason partitioning
9454 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9455 selected via a scoring technique which handles IDP and AKID in CRLs.
9459 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9460 will ultimately be used for all verify operations: this will remove the
9461 X509_STORE dependency on certificate verification and allow alternative
9462 lookup methods. X509_STORE based implementations of these two callbacks.
9466 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9467 Modify get_crl() to find a valid (unexpired) CRL if possible.
9471 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9472 this would be called X509_CRL_cmp() but that name is already used by
9473 a function that just compares CRL issuer names. Cache several CRL
9474 extensions in X509_CRL structure and cache CRLDP in X509.
9478 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9479 this maps equivalent X509_NAME structures into a consistent structure.
9480 Name comparison can then be performed rapidly using memcmp().
9484 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9489 * Allow digests to supply their own micalg string for S/MIME type using
9490 the ctrl EVP_MD_CTRL_MICALG.
9494 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9495 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9496 ctrl. It can then customise the structure before and/or after signing
9501 * New function OBJ_add_sigid() to allow application defined signature OIDs
9502 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9503 to free up any added signature OIDs.
9507 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9508 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9509 digest and cipher tables. New options added to openssl utility:
9510 list-message-digest-algorithms and list-cipher-algorithms.
9514 * Change the array representation of binary polynomials: the list
9515 of degrees of non-zero coefficients is now terminated with -1.
9516 Previously it was terminated with 0, which was also part of the
9517 value; thus, the array representation was not applicable to
9518 polynomials where t^0 has coefficient zero. This change makes
9519 the array representation useful in a more general context.
9523 * Various modifications and fixes to SSL/TLS cipher string
9524 handling. For ECC, the code now distinguishes between fixed ECDH
9525 with RSA certificates on the one hand and with ECDSA certificates
9526 on the other hand, since these are separate ciphersuites. The
9527 unused code for Fortezza ciphersuites has been removed.
9529 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9530 (not "ECDHE"). For consistency with the code for DH
9531 certificates, use of ECDH certificates is now considered ECDH
9532 authentication, not RSA or ECDSA authentication (the latter is
9533 merely the CA's signing algorithm and not actively used in the
9536 The temporary ciphersuite alias "ECCdraft" is no longer
9537 available, and ECC ciphersuites are no longer excluded from "ALL"
9538 and "DEFAULT". The following aliases now exist for RFC 4492
9539 ciphersuites, most of these by analogy with the DH case:
9541 kECDHr - ECDH cert, signed with RSA
9542 kECDHe - ECDH cert, signed with ECDSA
9543 kECDH - ECDH cert (signed with either RSA or ECDSA)
9544 kEECDH - ephemeral ECDH
9545 ECDH - ECDH cert or ephemeral ECDH
9551 AECDH - anonymous ECDH
9552 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9556 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9557 Use correct micalg parameters depending on digest(s) in signed message.
9561 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9562 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9566 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9567 an engine to register a method. Add ENGINE lookups for methods and
9568 functional reference processing.
9572 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9573 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9578 * New -resign option to smime utility. This adds one or more signers
9579 to an existing PKCS#7 signedData structure. Also -md option to use an
9580 alternative message digest algorithm for signing.
9584 * Tidy up PKCS#7 routines and add new functions to make it easier to
9585 create PKCS7 structures containing multiple signers. Update smime
9586 application to support multiple signers.
9590 * New -macalg option to pkcs12 utility to allow setting of an alternative
9595 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9596 Reorganize PBE internals to lookup from a static table using NIDs,
9597 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9598 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9599 PRF which will be automatically used with PBES2.
9603 * Replace the algorithm specific calls to generate keys in "req" with the
9608 * Update PKCS#7 enveloped data routines to use new API. This is now
9609 supported by any public key method supporting the encrypt operation. A
9610 ctrl is added to allow the public key algorithm to examine or modify
9611 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9616 * Add a ctrl to asn1 method to allow a public key algorithm to express
9617 a default digest type to use. In most cases this will be SHA1 but some
9618 algorithms (such as GOST) need to specify an alternative digest. The
9619 return value indicates how strong the preference is 1 means optional and
9620 2 is mandatory (that is it is the only supported type). Modify
9621 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9622 use the default md. Update openssl utilities to use the default digest
9623 type for signing if it is not explicitly indicated.
9627 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9628 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9629 signing method from the key type. This effectively removes the link
9630 between digests and public key types.
9634 * Add an OID cross reference table and utility functions. Its purpose is to
9635 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9636 rsaEncryption. This will allow some of the algorithm specific hackery
9637 needed to use the correct OID to be removed.
9641 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9642 structures for PKCS7_sign(). They are now set up by the relevant public
9647 * Add provisional EC pkey method with support for ECDSA and ECDH.
9651 * Add support for key derivation (agreement) in the API, DH method and
9656 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9657 public and private key formats. As a side effect these add additional
9658 command line functionality not previously available: DSA signatures can be
9659 generated and verified using pkeyutl and DH key support and generation in
9666 *Oliver Tappe <zooey@hirschkaefer.de>*
9668 * New make target "install_html_docs" installs HTML renditions of the
9671 *Oliver Tappe <zooey@hirschkaefer.de>*
9673 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9674 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9675 support key and parameter generation and add initial key generation
9676 functionality for RSA.
9680 * Add functions for main EVP_PKEY_method operations. The undocumented
9681 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9682 `EVP_PKEY_{encrypt,decrypt}_old`.
9686 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9687 key API, doesn't do much yet.
9691 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9692 public key algorithms. New option to openssl utility:
9693 "list-public-key-algorithms" to print out info.
9697 * Implement the Supported Elliptic Curves Extension for
9698 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9702 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9703 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9707 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9708 utilities such as rsa, dsa, dsaparam etc except they process any key
9713 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9714 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9715 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9720 * Initial support for pluggable public key ASN1.
9721 De-spaghettify the public key ASN1 handling. Move public and private
9722 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9723 algorithm specific handling to a single module within the relevant
9724 algorithm directory. Add functions to allow (near) opaque processing
9725 of public and private key structures.
9729 * Implement the Supported Point Formats Extension for
9730 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9734 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9735 for the psk identity [hint] and the psk callback functions to the
9736 SSL_SESSION, SSL and SSL_CTX structure.
9739 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9743 SSL_CTX_use_psk_identity_hint
9744 SSL_get_psk_identity_hint
9745 SSL_get_psk_identity
9746 SSL_use_psk_identity_hint
9748 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9750 * Add RFC 3161 compliant time stamp request creation, response generation
9751 and response verification functionality.
9753 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9755 * Add initial support for TLS extensions, specifically for the server_name
9756 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9757 have new members for a hostname. The SSL data structure has an
9758 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9759 stored in that context to allow for session resumption, even after the
9760 SSL has been switched to a new SSL_CTX in reaction to a client's
9761 server_name extension.
9763 New functions (subject to change):
9765 SSL_get_servername()
9766 SSL_get_servername_type()
9769 New CTRL codes and macros (subject to change):
9771 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9772 - SSL_CTX_set_tlsext_servername_callback()
9773 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9774 - SSL_CTX_set_tlsext_servername_arg()
9775 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9777 openssl s_client has a new '-servername ...' option.
9779 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9780 '-key2 ...', '-servername_fatal' (subject to change). This allows
9781 testing the HostName extension for a specific single hostname ('-cert'
9782 and '-key' remain fallbacks for handshakes without HostName
9783 negotiation). If the unrecognized_name alert has to be sent, this by
9784 default is a warning; it becomes fatal with the '-servername_fatal'
9787 *Peter Sylvester, Remy Allais, Christophe Renou*
9789 * Whirlpool hash implementation is added.
9793 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9794 bn(64,32). Because of instruction set limitations it doesn't have
9795 any negative impact on performance. This was done mostly in order
9796 to make it possible to share assembler modules, such as bn_mul_mont
9797 implementations, between 32- and 64-bit builds without hassle.
9801 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9802 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9807 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9808 dedicated Montgomery multiplication procedure, is introduced.
9809 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9810 "64-bit" performance on certain 32-bit targets.
9814 * New option SSL_OP_NO_COMP to disable use of compression selectively
9815 in SSL structures. New SSL ctrl to set maximum send fragment size.
9816 Save memory by setting the I/O buffer sizes dynamically instead of
9817 using the maximum available value.
9821 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9822 in addition to the text details.
9826 * Very, very preliminary EXPERIMENTAL support for printing of general
9827 ASN1 structures. This currently produces rather ugly output and doesn't
9828 handle several customised structures at all.
9832 * Integrated support for PVK file format and some related formats such
9833 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9834 these in the 'rsa' and 'dsa' utilities.
9838 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9842 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9843 place for the (very old) "NETSCAPE" format certificates which are now
9844 handled using new ASN1 code equivalents.
9848 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9849 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9850 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9854 * Modify CRL distribution points extension code to print out previously
9855 unsupported fields. Enhance extension setting code to allow setting of
9860 * Add print and set support for Issuing Distribution Point CRL extension.
9864 * Change 'Configure' script to enable Camellia by default.
9871 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9873 * When rejecting SSL/TLS records due to an incorrect version number, never
9874 update s->server with a new major version number. As of
9875 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9876 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9877 the previous behavior could result in a read attempt at NULL when
9878 receiving specific incorrect SSL/TLS records once record payload
9879 protection is active. ([CVE-2010-0740])
9881 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9883 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9884 could be crashed if the relevant tables were not present (e.g. chrooted).
9886 *Tomas Hoger <thoger@redhat.com>*
9888 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9890 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9892 *Martin Olsson, Neel Mehta*
9894 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9895 accommodate for stack sorting, always a write lock!).
9899 * On some versions of WIN32 Heap32Next is very slow. This can cause
9900 excessive delays in the RAND_poll(): over a minute. As a workaround
9901 include a time check in the inner Heap32Next loop too.
9905 * The code that handled flushing of data in SSL/TLS originally used the
9906 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9907 the problem outlined in PR#1949. The fix suggested there however can
9908 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9909 of Apache). So instead simplify the code to flush unconditionally.
9910 This should be fine since flushing with no data to flush is a no op.
9914 * Handle TLS versions 2.0 and later properly and correctly use the
9915 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9916 off ancient servers have a habit of sticking around for a while...
9920 * Modify compression code so it frees up structures without using the
9921 ex_data callbacks. This works around a problem where some applications
9922 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9923 restarting) then use compression (e.g. SSL with compression) later.
9924 This results in significant per-connection memory leaks and
9925 has caused some security issues including CVE-2008-1678 and
9930 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9931 change when encrypting or decrypting.
9935 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9936 connect and renegotiate with servers which do not support RI.
9937 Until RI is more widely deployed this option is enabled by default.
9941 * Add "missing" ssl ctrls to clear options and mode.
9945 * If client attempts to renegotiate and doesn't support RI respond with
9946 a no_renegotiation alert as required by RFC5746. Some renegotiating
9947 TLS clients will continue a connection gracefully when they receive
9948 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9949 waiting for a server hello which it will never receive. Now we treat a
9950 received no_renegotiation alert as a fatal error. This is because
9951 applications requesting a renegotiation might well expect it to succeed
9952 and would have no code in place to handle the server denying it so the
9953 only safe thing to do is to terminate the connection.
9957 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9958 peer supports secure renegotiation and 0 otherwise. Print out peer
9959 renegotiation support in s_client/s_server.
9963 * Replace the highly broken and deprecated SPKAC certification method with
9964 the updated NID creation version. This should correctly handle UTF8.
9968 * Implement RFC5746. Re-enable renegotiation but require the extension
9969 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9970 turns out to be a bad idea. It has been replaced by
9971 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9972 SSL_CTX_set_options(). This is really not recommended unless you
9973 know what you are doing.
9975 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9977 * Fixes to stateless session resumption handling. Use initial_ctx when
9978 issuing and attempting to decrypt tickets in case it has changed during
9979 servername handling. Use a non-zero length session ID when attempting
9980 stateless session resumption: this makes it possible to determine if
9981 a resumption has occurred immediately after receiving server hello
9982 (several places in OpenSSL subtly assume this) instead of later in
9987 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9988 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9989 fixes for a few places where the return code is not checked
9992 *Julia Lawall <julia@diku.dk>*
9994 * Add --strict-warnings option to Configure script to include devteam
9995 warnings in other configurations.
9999 * Add support for --libdir option and LIBDIR variable in makefiles. This
10000 makes it possible to install openssl libraries in locations which
10001 have names other than "lib", for example "/usr/lib64" which some
10004 *Steve Henson, based on patch from Jeremy Utley*
10006 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10007 X690 8.9.12 and can produce some misleading textual output of OIDs.
10009 *Steve Henson, reported by Dan Kaminsky*
10011 * Delete MD2 from algorithm tables. This follows the recommendation in
10012 several standards that it is not used in new applications due to
10013 several cryptographic weaknesses. For binary compatibility reasons
10014 the MD2 API is still compiled in by default.
10018 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10023 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10024 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10027 *Guenter <lists@gknw.net>*
10029 * Fix the server certificate chain building code to use X509_verify_cert(),
10030 it used to have an ad-hoc builder which was unable to cope with anything
10031 other than a simple chain.
10033 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10035 * Don't check self signed certificate signatures in X509_verify_cert()
10036 by default (a flag can override this): it just wastes time without
10037 adding any security. As a useful side effect self signed root CAs
10038 with non-FIPS digests are now usable in FIPS mode.
10042 * In dtls1_process_out_of_seq_message() the check if the current message
10043 is already buffered was missing. For every new message was memory
10044 allocated, allowing an attacker to perform an denial of service attack
10045 with sending out of seq handshake messages until there is no memory
10046 left. Additionally every future message was buffered, even if the
10047 sequence number made no sense and would be part of another handshake.
10048 So only messages with sequence numbers less than 10 in advance will be
10049 buffered. ([CVE-2009-1378])
10051 *Robin Seggelmann, discovered by Daniel Mentz*
10053 * Records are buffered if they arrive with a future epoch to be
10054 processed after finishing the corresponding handshake. There is
10055 currently no limitation to this buffer allowing an attacker to perform
10056 a DOS attack with sending records with future epochs until there is no
10057 memory left. This patch adds the pqueue_size() function to determine
10058 the size of a buffer and limits the record buffer to 100 entries.
10061 *Robin Seggelmann, discovered by Daniel Mentz*
10063 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10064 parent structure is freed. ([CVE-2009-1379])
10068 * Handle non-blocking I/O properly in SSL_shutdown() call.
10070 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10072 * Add `2.5.4.*` OIDs
10074 *Ilya O. <vrghost@gmail.com>*
10076 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10078 * Disable renegotiation completely - this fixes a severe security
10079 problem ([CVE-2009-3555]) at the cost of breaking all
10080 renegotiation. Renegotiation can be re-enabled by setting
10081 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10082 run-time. This is really not recommended unless you know what
10087 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10089 * Don't set val to NULL when freeing up structures, it is freed up by
10090 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10091 zeroing past the valid field. ([CVE-2009-0789])
10093 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10095 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10096 checked correctly. This would allow some invalid signed attributes to
10097 appear to verify correctly. ([CVE-2009-0591])
10099 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10101 * Reject UniversalString and BMPString types with invalid lengths. This
10102 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10103 a legal length. ([CVE-2009-0590])
10107 * Set S/MIME signing as the default purpose rather than setting it
10108 unconditionally. This allows applications to override it at the store
10113 * Permit restricted recursion of ASN1 strings. This is needed in practice
10114 to handle some structures.
10118 * Improve efficiency of mem_gets: don't search whole buffer each time
10121 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10123 * New -hex option for openssl rand.
10127 * Print out UTF8String and NumericString when parsing ASN1.
10131 * Support NumericString type for name components.
10135 * Allow CC in the environment to override the automatically chosen
10136 compiler. Note that nothing is done to ensure flags work with the
10141 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10143 * Properly check EVP_VerifyFinal() and similar return values
10146 *Ben Laurie, Bodo Moeller, Google Security Team*
10148 * Enable TLS extensions by default.
10152 * Allow the CHIL engine to be loaded, whether the application is
10153 multithreaded or not. (This does not release the developer from the
10154 obligation to set up the dynamic locking callbacks.)
10156 *Sander Temme <sander@temme.net>*
10158 * Use correct exit code if there is an error in dgst command.
10160 *Steve Henson; problem pointed out by Roland Dirlewanger*
10162 * Tweak Configure so that you need to say "experimental-jpake" to enable
10163 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10167 * Add experimental JPAKE support, including demo authentication in
10168 s_client and s_server.
10172 * Set the comparison function in v3_addr_canonize().
10174 *Rob Austein <sra@hactrn.net>*
10176 * Add support for XMPP STARTTLS in s_client.
10178 *Philip Paeps <philip@freebsd.org>*
10180 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10181 to ensure that even with this option, only ciphersuites in the
10182 server's preference list will be accepted. (Note that the option
10183 applies only when resuming a session, so the earlier behavior was
10184 just about the algorithm choice for symmetric cryptography.)
10188 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10190 * Fix NULL pointer dereference if a DTLS server received
10191 ChangeCipherSpec as first record ([CVE-2009-1386]).
10195 * Fix a state transition in s3_srvr.c and d1_srvr.c
10196 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10198 *Nagendra Modadugu*
10200 * The fix in 0.9.8c that supposedly got rid of unsafe
10201 double-checked locking was incomplete for RSA blinding,
10202 addressing just one layer of what turns out to have been
10203 doubly unsafe triple-checked locking.
10205 So now fix this for real by retiring the MONT_HELPER macro
10206 in crypto/rsa/rsa_eay.c.
10208 *Bodo Moeller; problem pointed out by Marius Schilder*
10210 * Various precautionary measures:
10212 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10214 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10215 (NB: This would require knowledge of the secret session ticket key
10216 to exploit, in which case you'd be SOL either way.)
10218 - Change bn_nist.c so that it will properly handle input BIGNUMs
10219 outside the expected range.
10221 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10224 *Neel Mehta, Bodo Moeller*
10226 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10227 the load fails. Useful for distros.
10229 *Ben Laurie and the FreeBSD team*
10231 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10235 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10239 * Expand ENGINE to support engine supplied SSL client certificate functions.
10241 This work was sponsored by Logica.
10245 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10246 keystores. Support for SSL/TLS client authentication too.
10247 Not compiled unless enable-capieng specified to Configure.
10249 This work was sponsored by Logica.
10253 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10254 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10255 attribute creation routines such as certificate requests and PKCS#12
10260 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10262 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10263 handshake which could lead to a client crash as found using the
10264 Codenomicon TLS test suite ([CVE-2008-1672])
10266 *Steve Henson, Mark Cox*
10268 * Fix double free in TLS server name extensions which could lead to
10269 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10273 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10275 Clear the error queue to ensure that error entries left from
10276 older function calls do not interfere with the correct operation.
10278 *Lutz Jaenicke, Erik de Castro Lopo*
10280 * Remove root CA certificates of commercial CAs:
10282 The OpenSSL project does not recommend any specific CA and does not
10283 have any policy with respect to including or excluding any CA.
10284 Therefore, it does not make any sense to ship an arbitrary selection
10285 of root CA certificates with the OpenSSL software.
10289 * RSA OAEP patches to fix two separate invalid memory reads.
10290 The first one involves inputs when 'lzero' is greater than
10291 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10292 before the beginning of from). The second one involves inputs where
10293 the 'db' section contains nothing but zeroes (there is a one-byte
10294 invalid read after the end of 'db').
10296 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10298 * Partial backport from 0.9.9-dev:
10300 Introduce bn_mul_mont (dedicated Montgomery multiplication
10301 procedure) as a candidate for BIGNUM assembler implementation.
10302 While 0.9.9-dev uses assembler for various architectures, only
10303 x86_64 is available by default here in the 0.9.8 branch, and
10304 32-bit x86 is available through a compile-time setting.
10306 To try the 32-bit x86 assembler implementation, use Configure
10307 option "enable-montasm" (which exists only for this backport).
10309 As "enable-montasm" for 32-bit x86 disclaims code stability
10310 anyway, in this constellation we activate additional code
10311 backported from 0.9.9-dev for further performance improvements,
10312 namely BN_from_montgomery_word. (To enable this otherwise,
10313 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10315 *Andy Polyakov (backport partially by Bodo Moeller)*
10317 * Add TLS session ticket callback. This allows an application to set
10318 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10319 values. This is useful for key rollover for example where several key
10320 sets may exist with different names.
10324 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10325 This was broken until now in 0.9.8 releases, such that the only way
10326 a registered ENGINE could be used (assuming it initialises
10327 successfully on the host) was to explicitly set it as the default
10328 for the relevant algorithms. This is in contradiction with 0.9.7
10329 behaviour and the documentation. With this fix, when an ENGINE is
10330 registered into a given algorithm's table of implementations, the
10331 'uptodate' flag is reset so that auto-discovery will be used next
10332 time a new context for that algorithm attempts to select an
10335 *Ian Lister (tweaked by Geoff Thorpe)*
10337 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10338 implementation in the following ways:
10340 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10343 Lack of BER streaming support means one pass streaming processing is
10344 only supported if data is detached: setting the streaming flag is
10345 ignored for embedded content.
10347 CMS support is disabled by default and must be explicitly enabled
10348 with the enable-cms configuration option.
10352 * Update the GMP engine glue to do direct copies between BIGNUM and
10353 mpz_t when openssl and GMP use the same limb size. Otherwise the
10354 existing "conversion via a text string export" trick is still used.
10356 *Paul Sheer <paulsheer@gmail.com>*
10358 * Zlib compression BIO. This is a filter BIO which compressed and
10359 uncompresses any data passed through it.
10363 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10364 RFC3394 compatible AES key wrapping.
10368 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10369 sets string data without copying. X509_ALGOR_set0() and
10370 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10371 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10372 from an X509_ATTRIBUTE structure optionally checking it occurs only
10373 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10378 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10379 to get the expected BN_FLG_CONSTTIME behavior.
10381 *Bodo Moeller (Google)*
10385 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10386 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10387 - added some more tests to do_tests.pl
10388 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10389 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10390 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10391 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10392 - various changes to netware.pl to enable gcc-cross builds on Win32
10394 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10395 - various changes to fix missing prototype warnings
10396 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10397 - added AES, WHIRLPOOL and CPUID assembler code to build files
10398 - added missing AES assembler make rules to mk1mf.pl
10399 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10401 *Guenter Knauf <eflash@gmx.net>*
10403 * Implement certificate status request TLS extension defined in RFC3546.
10404 A client can set the appropriate parameters and receive the encoded
10405 OCSP response via a callback. A server can query the supplied parameters
10406 and set the encoded OCSP response in the callback. Add simplified examples
10407 to s_client and s_server.
10411 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10413 * Fix various bugs:
10414 + Binary incompatibility of ssl_ctx_st structure
10415 + DTLS interoperation with non-compliant servers
10416 + Don't call get_session_cb() without proposed session
10417 + Fix ia64 assembler code
10419 *Andy Polyakov, Steve Henson*
10421 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10423 * DTLS Handshake overhaul. There were longstanding issues with
10424 OpenSSL DTLS implementation, which were making it impossible for
10425 RFC 4347 compliant client to communicate with OpenSSL server.
10426 Unfortunately just fixing these incompatibilities would "cut off"
10427 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10428 server keeps tolerating non RFC compliant syntax. The opposite is
10429 not true, 0.9.8f client can not communicate with earlier server.
10430 This update even addresses CVE-2007-4995.
10434 * Changes to avoid need for function casts in OpenSSL: some compilers
10435 (gcc 4.2 and later) reject their use.
10436 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10439 * Add RFC4507 support to OpenSSL. This includes the corrections in
10440 RFC4507bis. The encrypted ticket format is an encrypted encoded
10441 SSL_SESSION structure, that way new session features are automatically
10444 If a client application caches session in an SSL_SESSION structure
10445 support is transparent because tickets are now stored in the encoded
10448 The SSL_CTX structure automatically generates keys for ticket
10449 protection in servers so again support should be possible
10450 with no application modification.
10452 If a client or server wishes to disable RFC4507 support then the option
10453 SSL_OP_NO_TICKET can be set.
10455 Add a TLS extension debugging callback to allow the contents of any client
10456 or server extensions to be examined.
10458 This work was sponsored by Google.
10462 * Add initial support for TLS extensions, specifically for the server_name
10463 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10464 have new members for a hostname. The SSL data structure has an
10465 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10466 stored in that context to allow for session resumption, even after the
10467 SSL has been switched to a new SSL_CTX in reaction to a client's
10468 server_name extension.
10470 New functions (subject to change):
10472 SSL_get_servername()
10473 SSL_get_servername_type()
10476 New CTRL codes and macros (subject to change):
10478 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10479 - SSL_CTX_set_tlsext_servername_callback()
10480 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10481 - SSL_CTX_set_tlsext_servername_arg()
10482 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10484 openssl s_client has a new '-servername ...' option.
10486 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10487 '-key2 ...', '-servername_fatal' (subject to change). This allows
10488 testing the HostName extension for a specific single hostname ('-cert'
10489 and '-key' remain fallbacks for handshakes without HostName
10490 negotiation). If the unrecognized_name alert has to be sent, this by
10491 default is a warning; it becomes fatal with the '-servername_fatal'
10494 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10496 * Add AES and SSE2 assembly language support to VC++ build.
10500 * Mitigate attack on final subtraction in Montgomery reduction.
10504 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10505 (which previously caused an internal error).
10509 * Squeeze another 10% out of IGE mode when in != out.
10513 * AES IGE mode speedup.
10515 *Dean Gaudet (Google)*
10517 * Add the Korean symmetric 128-bit cipher SEED (see
10518 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10519 add SEED ciphersuites from RFC 4162:
10521 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10522 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10523 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10524 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10526 To minimize changes between patchlevels in the OpenSSL 0.9.8
10527 series, SEED remains excluded from compilation unless OpenSSL
10528 is configured with 'enable-seed'.
10530 *KISA, Bodo Moeller*
10532 * Mitigate branch prediction attacks, which can be practical if a
10533 single processor is shared, allowing a spy process to extract
10534 information. For detailed background information, see
10535 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10536 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10537 and Necessary Software Countermeasures"). The core of the change
10538 are new versions BN_div_no_branch() and
10539 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10540 respectively, which are slower, but avoid the security-relevant
10541 conditional branches. These are automatically called by BN_div()
10542 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10543 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10544 remove a conditional branch.
10546 BN_FLG_CONSTTIME is the new name for the previous
10547 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10548 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10549 in the exponent causes BN_mod_exp_mont() to use the alternative
10550 implementation in BN_mod_exp_mont_consttime().) The old name
10551 remains as a deprecated alias.
10553 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10554 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10555 constant-time implementations for more than just exponentiation.
10556 Here too the old name is kept as a deprecated alias.
10558 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10559 the BN_BLINDING structure gets an independent copy of the
10560 modulus. This means that the previous `BIGNUM *m` argument to
10561 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10562 essentially becomes `const BIGNUM *m`, although we can't actually
10563 change this in the header file before 0.9.9. It allows
10564 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10565 enable BN_FLG_CONSTTIME.
10567 *Matthew D Wood (Intel Corp)*
10569 * In the SSL/TLS server implementation, be strict about session ID
10570 context matching (which matters if an application uses a single
10571 external cache for different purposes). Previously,
10572 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10573 set. This did ensure strict client verification, but meant that,
10574 with applications using a single external cache for quite
10575 different requirements, clients could circumvent ciphersuite
10576 restrictions for a given session ID context by starting a session
10577 in a different context.
10581 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10582 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10583 authentication-only ciphersuites.
10587 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10588 not complete and could lead to a possible single byte overflow
10589 ([CVE-2007-5135]) [Ben Laurie]
10591 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10593 * Since AES128 and AES256 (and similarly Camellia128 and
10594 Camellia256) share a single mask bit in the logic of
10595 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10596 kludge to work properly if AES128 is available and AES256 isn't
10597 (or if Camellia128 is available and Camellia256 isn't).
10601 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10602 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10603 When a point or a seed is encoded in a BIT STRING, we need to
10604 prevent the removal of trailing zero bits to get the proper DER
10605 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10606 of a NamedBitList, for which trailing 0 bits need to be removed.)
10610 * Have SSL/TLS server implementation tolerate "mismatched" record
10611 protocol version while receiving ClientHello even if the
10612 ClientHello is fragmented. (The server can't insist on the
10613 particular protocol version it has chosen before the ServerHello
10614 message has informed the client about his choice.)
10618 * Add RFC 3779 support.
10620 *Rob Austein for ARIN, Ben Laurie*
10622 * Load error codes if they are not already present instead of using a
10623 static variable. This allows them to be cleanly unloaded and reloaded.
10624 Improve header file function name parsing.
10628 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10629 or CAPABILITY handshake as required by RFCs.
10631 *Goetz Babin-Ebell*
10633 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10635 * Introduce limits to prevent malicious keys being able to
10636 cause a denial of service. ([CVE-2006-2940])
10638 *Steve Henson, Bodo Moeller*
10640 * Fix ASN.1 parsing of certain invalid structures that can result
10641 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10643 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10644 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10646 * Fix SSL client code which could crash if connecting to a
10647 malicious SSLv2 server. ([CVE-2006-4343])
10649 *Tavis Ormandy and Will Drewry, Google Security Team*
10651 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10652 match only those. Before that, "AES256-SHA" would be interpreted
10653 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10654 the same strength classification in 0.9.7h) as we currently only
10655 have a single AES bit in the ciphersuite description bitmap.
10656 That change, however, also applied to ciphersuite strings such as
10657 "RC4-MD5" that intentionally matched multiple ciphersuites --
10658 namely, SSL 2.0 ciphersuites in addition to the more common ones
10659 from SSL 3.0/TLS 1.0.
10661 So we change the selection algorithm again: Naming an explicit
10662 ciphersuite selects this one ciphersuite, and any other similar
10663 ciphersuite (same bitmap) from *other* protocol versions.
10664 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10665 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10667 Since SSL 2.0 does not have any ciphersuites for which the
10668 128/256 bit distinction would be relevant, this works for now.
10669 The proper fix will be to use different bits for AES128 and
10670 AES256, which would have avoided the problems from the beginning;
10671 however, bits are scarce, so we can only do this in a new release
10672 (not just a patchlevel) when we can change the SSL_CIPHER
10673 definition to split the single 'unsigned long mask' bitmap into
10674 multiple values to extend the available space.
10678 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10680 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10681 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10683 * Add AES IGE and biIGE modes.
10687 * Change the Unix randomness entropy gathering to use poll() when
10688 possible instead of select(), since the latter has some
10689 undesirable limitations.
10691 *Darryl Miles via Richard Levitte and Bodo Moeller*
10693 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10694 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10695 cannot be implicitly activated as part of, e.g., the "AES" alias.
10696 However, please upgrade to OpenSSL 0.9.9[-dev] for
10697 non-experimental use of the ECC ciphersuites to get TLS extension
10698 support, which is required for curve and point format negotiation
10699 to avoid potential handshake problems.
10703 * Disable rogue ciphersuites:
10705 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10706 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10707 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10709 The latter two were purportedly from
10710 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10713 Also deactivate the remaining ciphersuites from
10714 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10715 unofficial, and the ID has long expired.
10719 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10720 dual-core machines) and other potential thread-safety issues.
10724 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10725 versions), which is now available for royalty-free use
10726 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10727 Also, add Camellia TLS ciphersuites from RFC 4132.
10729 To minimize changes between patchlevels in the OpenSSL 0.9.8
10730 series, Camellia remains excluded from compilation unless OpenSSL
10731 is configured with 'enable-camellia'.
10735 * Disable the padding bug check when compression is in use. The padding
10736 bug check assumes the first packet is of even length, this is not
10737 necessarily true if compression is enabled and can result in false
10738 positives causing handshake failure. The actual bug test is ancient
10739 code so it is hoped that implementations will either have fixed it by
10740 now or any which still have the bug do not support compression.
10744 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10746 * When applying a cipher rule check to see if string match is an explicit
10747 cipher suite and only match that one cipher suite if it is.
10751 * Link in manifests for VC++ if needed.
10753 *Austin Ziegler <halostatue@gmail.com>*
10755 * Update support for ECC-based TLS ciphersuites according to
10756 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10757 TLS extensions, which are supported starting with the 0.9.9
10758 branch, not in the OpenSSL 0.9.8 branch).
10762 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10763 opaque EVP_CIPHER_CTX handling.
10767 * Fixes and enhancements to zlib compression code. We now only use
10768 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10769 to conform with the standards mentioned here:
10770 <http://www.zlib.net/DLL_FAQ.txt>
10771 Static zlib linking now works on Windows and the new --with-zlib-include
10772 --with-zlib-lib options to Configure can be used to supply the location
10773 of the headers and library. Gracefully handle case where zlib library
10778 * Several fixes and enhancements to the OID generation code. The old code
10779 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10780 handle numbers larger than ULONG_MAX, truncated printing and had a
10781 non standard OBJ_obj2txt() behaviour.
10785 * Add support for building of engines under engine/ as shared libraries
10786 under VC++ build system.
10790 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10791 Hopefully, we will not see any false combination of paths any more.
10795 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10797 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10798 (part of SSL_OP_ALL). This option used to disable the
10799 countermeasure against man-in-the-middle protocol-version
10800 rollback in the SSL 2.0 server implementation, which is a bad
10801 idea. ([CVE-2005-2969])
10803 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10804 for Information Security, National Institute of Advanced Industrial
10805 Science and Technology [AIST], Japan)*
10807 * Add two function to clear and return the verify parameter flags.
10811 * Keep cipherlists sorted in the source instead of sorting them at
10812 runtime, thus removing the need for a lock.
10816 * Avoid some small subgroup attacks in Diffie-Hellman.
10818 *Nick Mathewson and Ben Laurie*
10820 * Add functions for well-known primes.
10824 * Extended Windows CE support.
10826 *Satoshi Nakamura and Andy Polyakov*
10828 * Initialize SSL_METHOD structures at compile time instead of during
10829 runtime, thus removing the need for a lock.
10833 * Make PKCS7_decrypt() work even if no certificate is supplied by
10834 attempting to decrypt each encrypted key in turn. Add support to
10839 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10841 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10844 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10848 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10849 key into the same file any more.
10853 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10857 * Add -utf8 command line and config file option to 'ca'.
10859 *Stefan <stf@udoma.org*
10861 * Removed the macro des_crypt(), as it seems to conflict with some
10862 libraries. Use DES_crypt().
10866 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10867 involves renaming the source and generated shared-libs for
10868 both. The engines will accept the corrected or legacy ids
10869 ('ncipher' and '4758_cca' respectively) when binding. NB,
10870 this only applies when building 'shared'.
10872 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10874 * Add attribute functions to EVP_PKEY structure. Modify
10875 PKCS12_create() to recognize a CSP name attribute and
10876 use it. Make -CSP option work again in pkcs12 utility.
10880 * Add new functionality to the bn blinding code:
10881 - automatic re-creation of the BN_BLINDING parameters after
10882 a fixed number of uses (currently 32)
10883 - add new function for parameter creation
10884 - introduce flags to control the update behaviour of the
10885 BN_BLINDING parameters
10886 - hide BN_BLINDING structure
10887 Add a second BN_BLINDING slot to the RSA structure to improve
10888 performance when a single RSA object is shared among several
10893 * Add support for DTLS.
10895 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10897 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10898 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10902 * Remove buggy and incomplete DH cert support from
10903 ssl/ssl_rsa.c and ssl/s3_both.c
10907 * Use SHA-1 instead of MD5 as the default digest algorithm for
10908 the `apps/openssl` commands.
10912 * Compile clean with "-Wall -Wmissing-prototypes
10913 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10914 DEBUG_SAFESTACK must also be set.
10918 * Change ./Configure so that certain algorithms can be disabled by default.
10919 The new counterpiece to "no-xxx" is "enable-xxx".
10921 The patented RC5 and MDC2 algorithms will now be disabled unless
10922 "enable-rc5" and "enable-mdc2", respectively, are specified.
10924 (IDEA remains enabled despite being patented. This is because IDEA
10925 is frequently required for interoperability, and there is no license
10926 fee for non-commercial use. As before, "no-idea" can be used to
10927 avoid this algorithm.)
10931 * Add processing of proxy certificates (see RFC 3820). This work was
10932 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10933 EGEE (Enabling Grids for E-science in Europe).
10937 * RC4 performance overhaul on modern architectures/implementations, such
10938 as Intel P4, IA-64 and AMD64.
10942 * New utility extract-section.pl. This can be used specify an alternative
10943 section number in a pod file instead of having to treat each file as
10944 a separate case in Makefile. This can be done by adding two lines to the
10947 =for comment openssl_section:XXX
10949 The blank line is mandatory.
10953 * New arguments -certform, -keyform and -pass for s_client and s_server
10954 to allow alternative format key and certificate files and passphrase
10959 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10960 update associated structures and add various utility functions.
10962 Add new policy related verify parameters, include policy checking in
10963 standard verify code. Enhance 'smime' application with extra parameters
10964 to support policy checking and print out.
10968 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10969 Nehemiah processors. These extensions support AES encryption in hardware
10970 as well as RNG (though RNG support is currently disabled).
10972 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10974 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10978 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10980 *Andy Polyakov and a number of other people*
10982 * Improved PowerPC platform support. Most notably BIGNUM assembler
10983 implementation contributed by IBM.
10985 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10987 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10988 exponent rather than 'unsigned long'. There is a corresponding change to
10989 the new 'rsa_keygen' element of the RSA_METHOD structure.
10991 *Jelte Jansen, Geoff Thorpe*
10993 * Functionality for creating the initial serial number file is now
10994 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10996 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10997 number file to 1, which is bound to cause problems. To avoid
10998 the problems while respecting compatibility between different 0.9.7
10999 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11000 CA.pl for serial number initialization. With the new release 0.9.8,
11001 we can fix the problem directly in the 'ca' utility.)
11005 * Reduced header interdependencies by declaring more opaque objects in
11006 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11007 give fewer recursive includes, which could break lazy source code - so
11008 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11009 developers should define this symbol when building and using openssl to
11010 ensure they track the recommended behaviour, interfaces, [etc], but
11011 backwards-compatible behaviour prevails when this isn't defined.
11015 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11019 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11020 This will generate a random key of the appropriate length based on the
11021 cipher context. The EVP_CIPHER can provide its own random key generation
11022 routine to support keys of a specific form. This is used in the des and
11023 3des routines to generate a key of the correct parity. Update S/MIME
11024 code to use new functions and hence generate correct parity DES keys.
11025 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11026 valid (weak or incorrect parity).
11030 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11031 as looking them up. This is useful when the verified structure may contain
11032 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11033 present unless the new PKCS7_NO_CRL flag is asserted.
11037 * Extend ASN1 oid configuration module. It now additionally accepts the
11040 shortName = some long name, 1.2.3.4
11044 * Reimplemented the BN_CTX implementation. There is now no more static
11045 limitation on the number of variables it can handle nor the depth of the
11046 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11047 information can now expand as required, and rather than having a single
11048 static array of bignums, BN_CTX now uses a linked-list of such arrays
11049 allowing it to expand on demand whilst maintaining the usefulness of
11050 BN_CTX's "bundling".
11054 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11055 to allow all RSA operations to function using a single BN_CTX.
11059 * Preliminary support for certificate policy evaluation and checking. This
11060 is initially intended to pass the tests outlined in "Conformance Testing
11061 of Relying Party Client Certificate Path Processing Logic" v1.07.
11065 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11066 remained unused and not that useful. A variety of other little bignum
11067 tweaks and fixes have also been made continuing on from the audit (see
11072 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11073 associated ASN1, EVP and SSL functions and old ASN1 macros.
11077 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11078 and this should never fail. So the return value from the use of
11079 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11080 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11084 * BN_CTX_get() should return zero-valued bignums, providing the same
11085 initialised value as BN_new().
11087 *Geoff Thorpe, suggested by Ulf Möller*
11089 * Support for inhibitAnyPolicy certificate extension.
11093 * An audit of the BIGNUM code is underway, for which debugging code is
11094 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11095 is considered valid when processing BIGNUMs, and causes execution to
11096 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11097 further steps are taken to deliberately pollute unused data in BIGNUM
11098 structures to try and expose faulty code further on. For now, openssl will
11099 (in its default mode of operation) continue to tolerate the inconsistent
11100 forms that it has tolerated in the past, but authors and packagers should
11101 consider trying openssl and their own applications when compiled with
11102 these debugging symbols defined. It will help highlight potential bugs in
11103 their own code, and will improve the test coverage for OpenSSL itself. At
11104 some point, these tighter rules will become openssl's default to improve
11105 maintainability, though the assert()s and other overheads will remain only
11106 in debugging configurations. See bn.h for more details.
11108 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11110 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11111 that can only be obtained through BN_CTX_new() (which implicitly
11112 initialises it). The presence of this function only made it possible
11113 to overwrite an existing structure (and cause memory leaks).
11117 * Because of the callback-based approach for implementing LHASH as a
11118 template type, lh_insert() adds opaque objects to hash-tables and
11119 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11120 to clean up those corresponding objects before destroying the hash table
11121 (and losing the object pointers). So some over-zealous constifications in
11122 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11123 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11124 prototyped to have "const" restrictions on the object pointers they are
11125 given (and so aren't required to cast them away any more).
11129 * The tmdiff.h API was so ugly and minimal that our own timing utility
11130 (speed) prefers to use its own implementation. The two implementations
11131 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11132 its object type properly exposed (MS_TM) instead of casting to/from
11133 `char *`. This may still change yet if someone realises MS_TM and
11135 aren't necessarily the greatest nomenclatures - but this is what was used
11136 internally to the implementation so I've used that for now.
11140 * Ensure that deprecated functions do not get compiled when
11141 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11142 the self-tests were still using deprecated key-generation functions so
11143 these have been updated also.
11147 * Reorganise PKCS#7 code to separate the digest location functionality
11148 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11149 New function PKCS7_set_digest() to set the digest type for PKCS#7
11150 digestedData type. Add additional code to correctly generate the
11151 digestedData type and add support for this type in PKCS7 initialization
11156 * New function PKCS7_set0_type_other() this initializes a PKCS7
11157 structure of type "other".
11161 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11162 sure the loop does correctly stop and breaking ("division by zero")
11163 modulus operations are not performed. The (pre-generated) prime
11164 table crypto/bn/bn_prime.h was already correct, but it could not be
11165 re-generated on some platforms because of the "division by zero"
11166 situation in the script.
11168 *Ralf S. Engelschall*
11170 * Update support for ECC-based TLS ciphersuites according to
11171 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11172 SHA-1 now is only used for "small" curves (where the
11173 representation of a field element takes up to 24 bytes); for
11174 larger curves, the field element resulting from ECDH is directly
11175 used as premaster secret.
11177 *Douglas Stebila (Sun Microsystems Laboratories)*
11179 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11180 curve secp160r1 to the tests.
11182 *Douglas Stebila (Sun Microsystems Laboratories)*
11184 * Add the possibility to load symbols globally with DSO.
11186 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11188 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11189 control of the error stack.
11193 * Add support for STORE in ENGINE.
11197 * Add the STORE type. The intention is to provide a common interface
11198 to certificate and key stores, be they simple file-based stores, or
11199 HSM-type store, or LDAP stores, or...
11200 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11204 * Add a generic structure called OPENSSL_ITEM. This can be used to
11205 pass a list of arguments to any function as well as provide a way
11206 for a function to pass data back to the caller.
11210 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11211 works like BUF_strdup() but can be used to duplicate a portion of
11212 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11217 * Add the function sk_find_ex() which works like sk_find(), but will
11218 return an index to an element even if an exact match couldn't be
11219 found. The index is guaranteed to point at the element where the
11220 searched-for key would be inserted to preserve sorting order.
11224 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11225 takes an extra flags argument for optional functionality. Currently,
11226 the following flags are defined:
11228 OBJ_BSEARCH_VALUE_ON_NOMATCH
11229 This one gets OBJ_bsearch_ex() to return a pointer to the first
11230 element where the comparing function returns a negative or zero
11233 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11234 This one gets OBJ_bsearch_ex() to return a pointer to the first
11235 element where the comparing function returns zero. This is useful
11236 if there are more than one element where the comparing function
11241 * Make it possible to create self-signed certificates with 'openssl ca'
11242 in such a way that the self-signed certificate becomes part of the
11243 CA database and uses the same mechanisms for serial number generation
11244 as all other certificate signing. The new flag '-selfsign' enables
11245 this functionality. Adapt CA.sh and CA.pl.in.
11249 * Add functionality to check the public key of a certificate request
11250 against a given private. This is useful to check that a certificate
11251 request can be signed by that key (self-signing).
11255 * Make it possible to have multiple active certificates with the same
11256 subject in the CA index file. This is done only if the keyword
11257 'unique_subject' is set to 'no' in the main CA section (default
11258 if 'CA_default') of the configuration file. The value is saved
11259 with the database itself in a separate index attribute file,
11260 named like the index file with '.attr' appended to the name.
11264 * Generate multi-valued AVAs using '+' notation in config files for
11269 * Support for nameConstraints certificate extension.
11273 * Support for policyConstraints certificate extension.
11277 * Support for policyMappings certificate extension.
11281 * Make sure the default DSA_METHOD implementation only uses its
11282 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11283 and change its own handlers to be NULL so as to remove unnecessary
11284 indirection. This lets alternative implementations fallback to the
11285 default implementation more easily.
11289 * Support for directoryName in GeneralName related extensions
11294 * Make it possible to link applications using Makefile.shared.
11295 Make that possible even when linking against static libraries!
11299 * Support for single pass processing for S/MIME signing. This now
11300 means that S/MIME signing can be done from a pipe, in addition
11301 cleartext signing (multipart/signed type) is effectively streaming
11302 and the signed data does not need to be all held in memory.
11304 This is done with a new flag PKCS7_STREAM. When this flag is set
11305 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11306 is done after the data is output (and digests calculated) in
11307 SMIME_write_PKCS7().
11311 * Add full support for -rpath/-R, both in shared libraries and
11312 applications, at least on the platforms where it's known how
11317 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11318 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11319 will now compute a table of multiples of the generator that
11320 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11321 faster (notably in the case of a single point multiplication,
11322 scalar * generator).
11324 *Nils Larsch, Bodo Moeller*
11326 * IPv6 support for certificate extensions. The various extensions
11327 which use the IP:a.b.c.d can now take IPv6 addresses using the
11328 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11333 * Added an ENGINE that implements RSA by performing private key
11334 exponentiations with the GMP library. The conversions to and from
11335 GMP's mpz_t format aren't optimised nor are any montgomery forms
11336 cached, and on x86 it appears OpenSSL's own performance has caught up.
11337 However there are likely to be other architectures where GMP could
11338 provide a boost. This ENGINE is not built in by default, but it can be
11339 specified at Configure time and should be accompanied by the necessary
11340 linker additions, eg;
11341 ./config -DOPENSSL_USE_GMP -lgmp
11345 * "openssl engine" will not display ENGINE/DSO load failure errors when
11346 testing availability of engines with "-t" - the old behaviour is
11347 produced by increasing the feature's verbosity with "-tt".
11351 * ECDSA routines: under certain error conditions uninitialized BN objects
11352 could be freed. Solution: make sure initialization is performed early
11353 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11358 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11359 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11360 software implementations. For DSA and DH, parameter generation can
11361 also be overridden by providing the appropriate method callbacks.
11365 * Change the "progress" mechanism used in key-generation and
11366 primality testing to functions that take a new BN_GENCB pointer in
11367 place of callback/argument pairs. The new API functions have `_ex`
11368 postfixes and the older functions are reimplemented as wrappers for
11369 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11370 declarations of the old functions to help (graceful) attempts to
11371 migrate to the new functions. Also, the new key-generation API
11372 functions operate on a caller-supplied key-structure and return
11373 success/failure rather than returning a key or NULL - this is to
11374 help make "keygen" another member function of RSA_METHOD etc.
11376 Example for using the new callback interface:
11378 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11379 void *my_arg = ...;
11382 BN_GENCB_set(&my_cb, my_callback, my_arg);
11384 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11385 /* For the meaning of a, b in calls to my_callback(), see the
11386 * documentation of the function that calls the callback.
11387 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11388 * my_callback should return 1 if it wants BN_is_prime_ex()
11389 * to continue, or 0 to stop.
11394 * Change the ZLIB compression method to be stateful, and make it
11395 available to TLS with the number defined in
11396 draft-ietf-tls-compression-04.txt.
11400 * Add the ASN.1 structures and functions for CertificatePair, which
11401 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11403 CertificatePair ::= SEQUENCE {
11404 forward [0] Certificate OPTIONAL,
11405 reverse [1] Certificate OPTIONAL,
11406 -- at least one of the pair shall be present -- }
11408 Also implement the PEM functions to read and write certificate
11409 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11411 This needed to be defined, mostly for the sake of the LDAP
11412 attribute crossCertificatePair, but may prove useful elsewhere as
11417 * Make it possible to inhibit symlinking of shared libraries in
11418 Makefile.shared, for Cygwin's sake.
11422 * Extend the BIGNUM API by creating a function
11423 void BN_set_negative(BIGNUM *a, int neg);
11424 and a macro that behave like
11425 int BN_is_negative(const BIGNUM *a);
11427 to avoid the need to access 'a->neg' directly in applications.
11431 * Implement fast modular reduction for pseudo-Mersenne primes
11432 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11433 EC_GROUP_new_curve_GFp() will now automatically use this
11436 *Nils Larsch <nla@trustcenter.de>*
11438 * Add new lock type (CRYPTO_LOCK_BN).
11442 * Change the ENGINE framework to automatically load engines
11443 dynamically from specific directories unless they could be
11444 found to already be built in or loaded. Move all the
11445 current engines except for the cryptodev one to a new
11446 directory engines/.
11447 The engines in engines/ are built as shared libraries if
11448 the "shared" options was given to ./Configure or ./config.
11449 Otherwise, they are inserted in libcrypto.a.
11450 /usr/local/ssl/engines is the default directory for dynamic
11451 engines, but that can be overridden at configure time through
11452 the usual use of --prefix and/or --openssldir, and at run
11453 time with the environment variable OPENSSL_ENGINES.
11455 *Geoff Thorpe and Richard Levitte*
11457 * Add Makefile.shared, a helper makefile to build shared
11458 libraries. Adapt Makefile.org.
11462 * Add version info to Win32 DLLs.
11464 *Peter 'Luna' Runestig" <peter@runestig.com>*
11466 * Add new 'medium level' PKCS#12 API. Certificates and keys
11467 can be added using this API to created arbitrary PKCS#12
11468 files while avoiding the low-level API.
11470 New options to PKCS12_create(), key or cert can be NULL and
11471 will then be omitted from the output file. The encryption
11472 algorithm NIDs can be set to -1 for no encryption, the mac
11473 iteration count can be set to 0 to omit the mac.
11475 Enhance pkcs12 utility by making the -nokeys and -nocerts
11476 options work when creating a PKCS#12 file. New option -nomac
11477 to omit the mac, NONE can be set for an encryption algorithm.
11478 New code is modified to use the enhanced PKCS12_create()
11479 instead of the low-level API.
11483 * Extend ASN1 encoder to support indefinite length constructed
11484 encoding. This can output sequences tags and octet strings in
11485 this form. Modify pk7_asn1.c to support indefinite length
11486 encoding. This is experimental and needs additional code to
11487 be useful, such as an ASN1 bio and some enhanced streaming
11490 Extend template encode functionality so that tagging is passed
11491 down to the template encoder.
11495 * Let 'openssl req' fail if an argument to '-newkey' is not
11496 recognized instead of using RSA as a default.
11500 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11501 As these are not official, they are not included in "ALL";
11502 the "ECCdraft" ciphersuite group alias can be used to select them.
11504 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11506 * Add ECDH engine support.
11508 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11510 * Add ECDH in new directory crypto/ecdh/.
11512 *Douglas Stebila (Sun Microsystems Laboratories)*
11514 * Let BN_rand_range() abort with an error after 100 iterations
11515 without success (which indicates a broken PRNG).
11519 * Change BN_mod_sqrt() so that it verifies that the input value
11520 is really the square of the return value. (Previously,
11521 BN_mod_sqrt would show GIGO behaviour.)
11525 * Add named elliptic curves over binary fields from X9.62, SECG,
11526 and WAP/WTLS; add OIDs that were still missing.
11528 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11530 * Extend the EC library for elliptic curves over binary fields
11531 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11534 EC_GF2m_simple_method
11538 EC_GROUP_new_curve_GF2m
11539 EC_GROUP_set_curve_GF2m
11540 EC_GROUP_get_curve_GF2m
11541 EC_POINT_set_affine_coordinates_GF2m
11542 EC_POINT_get_affine_coordinates_GF2m
11543 EC_POINT_set_compressed_coordinates_GF2m
11545 Point compression for binary fields is disabled by default for
11546 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11549 As binary polynomials are represented as BIGNUMs, various members
11550 of the EC_GROUP and EC_POINT data structures can be shared
11551 between the implementations for prime fields and binary fields;
11552 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11553 are essentially identical to their `..._GFp` counterparts.
11554 (For simplicity, the `..._GFp` prefix has been dropped from
11555 various internal method names.)
11557 An internal 'field_div' method (similar to 'field_mul' and
11558 'field_sqr') has been added; this is used only for binary fields.
11560 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11562 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11563 through methods ('mul', 'precompute_mult').
11565 The generic implementations (now internally called 'ec_wNAF_mul'
11566 and 'ec_wNAF_precomputed_mult') remain the default if these
11567 methods are undefined.
11569 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11571 * New function EC_GROUP_get_degree, which is defined through
11572 EC_METHOD. For curves over prime fields, this returns the bit
11573 length of the modulus.
11575 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11577 * New functions EC_GROUP_dup, EC_POINT_dup.
11578 (These simply call ..._new and ..._copy).
11580 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11582 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11583 Polynomials are represented as BIGNUMs (where the sign bit is not
11584 used) in the following functions [macros]:
11587 BN_GF2m_sub [= BN_GF2m_add]
11588 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11589 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11590 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11592 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11593 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11594 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11595 BN_GF2m_cmp [= BN_ucmp]
11597 (Note that only the 'mod' functions are actually for fields GF(2^m).
11598 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11600 For some functions, an the irreducible polynomial defining a
11601 field can be given as an 'unsigned int[]' with strictly
11602 decreasing elements giving the indices of those bits that are set;
11603 i.e., p[] represents the polynomial
11604 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11606 p[0] > p[1] > ... > p[k] = 0.
11607 This applies to the following functions:
11610 BN_GF2m_mod_mul_arr
11611 BN_GF2m_mod_sqr_arr
11612 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11613 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11614 BN_GF2m_mod_exp_arr
11615 BN_GF2m_mod_sqrt_arr
11616 BN_GF2m_mod_solve_quad_arr
11620 Conversion can be performed by the following functions:
11625 bntest.c has additional tests for binary polynomial arithmetic.
11627 Two implementations for BN_GF2m_mod_div() are available.
11628 The default algorithm simply uses BN_GF2m_mod_inv() and
11629 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11630 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11631 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11633 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11635 * Add new error code 'ERR_R_DISABLED' that can be used when some
11636 functionality is disabled at compile-time.
11638 *Douglas Stebila <douglas.stebila@sun.com>*
11640 * Change default behaviour of 'openssl asn1parse' so that more
11641 information is visible when viewing, e.g., a certificate:
11643 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11644 mode the content of non-printable OCTET STRINGs is output in a
11645 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11646 avoid the appearance of a printable string.
11648 *Nils Larsch <nla@trustcenter.de>*
11650 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11652 EC_GROUP_set_asn1_flag()
11653 EC_GROUP_get_asn1_flag()
11654 EC_GROUP_set_point_conversion_form()
11655 EC_GROUP_get_point_conversion_form()
11656 These control ASN1 encoding details:
11657 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11658 has been set to OPENSSL_EC_NAMED_CURVE.
11659 - Points are encoded in uncompressed form by default; options for
11660 asn1_for are as for point2oct, namely
11661 POINT_CONVERSION_COMPRESSED
11662 POINT_CONVERSION_UNCOMPRESSED
11663 POINT_CONVERSION_HYBRID
11665 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11667 EC_GROUP_set_seed()
11668 EC_GROUP_get0_seed()
11669 EC_GROUP_get_seed_len()
11670 This is used only for ASN1 purposes (so far).
11672 *Nils Larsch <nla@trustcenter.de>*
11674 * Add 'field_type' member to EC_METHOD, which holds the NID
11675 of the appropriate field type OID. The new function
11676 EC_METHOD_get_field_type() returns this value.
11678 *Nils Larsch <nla@trustcenter.de>*
11681 EC_POINT_point2bn()
11682 EC_POINT_bn2point()
11683 EC_POINT_point2hex()
11684 EC_POINT_hex2point()
11685 providing useful interfaces to EC_POINT_point2oct() and
11686 EC_POINT_oct2point().
11688 *Nils Larsch <nla@trustcenter.de>*
11690 * Change internals of the EC library so that the functions
11691 EC_GROUP_set_generator()
11692 EC_GROUP_get_generator()
11693 EC_GROUP_get_order()
11694 EC_GROUP_get_cofactor()
11695 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11696 to methods, which would lead to unnecessary code duplication when
11697 adding different types of curves.
11699 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11701 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11702 arithmetic, and such that modified wNAFs are generated
11703 (which avoid length expansion in many cases).
11707 * Add a function EC_GROUP_check_discriminant() (defined via
11708 EC_METHOD) that verifies that the curve discriminant is non-zero.
11710 Add a function EC_GROUP_check() that makes some sanity tests
11711 on a EC_GROUP, its generator and order. This includes
11712 EC_GROUP_check_discriminant().
11714 *Nils Larsch <nla@trustcenter.de>*
11716 * Add ECDSA in new directory crypto/ecdsa/.
11718 Add applications 'openssl ecparam' and 'openssl ecdsa'
11719 (these are based on 'openssl dsaparam' and 'openssl dsa').
11721 ECDSA support is also included in various other files across the
11722 library. Most notably,
11723 - 'openssl req' now has a '-newkey ecdsa:file' option;
11724 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11725 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11726 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11727 them suitable for ECDSA where domain parameters must be
11728 extracted before the specific public key;
11729 - ECDSA engine support has been added.
11731 *Nils Larsch <nla@trustcenter.de>*
11733 * Include some named elliptic curves, and add OIDs from X9.62,
11734 SECG, and WAP/WTLS. Each curve can be obtained from the new
11736 EC_GROUP_new_by_curve_name(),
11737 and the list of available named curves can be obtained with
11738 EC_get_builtin_curves().
11739 Also add a 'curve_name' member to EC_GROUP objects, which can be
11741 EC_GROUP_set_curve_name()
11742 EC_GROUP_get_curve_name()
11744 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11746 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11747 was actually never needed) and in BN_mul(). The removal in BN_mul()
11748 required a small change in bn_mul_part_recursive() and the addition
11749 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11750 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11751 bn_sub_words() and bn_add_words() except they take arrays with
11756 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11758 * Cleanse PEM buffers before freeing them since they may contain
11761 *Benjamin Bennett <ben@psc.edu>*
11763 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11764 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11765 authentication-only ciphersuites.
11769 * Since AES128 and AES256 share a single mask bit in the logic of
11770 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11771 kludge to work properly if AES128 is available and AES256 isn't.
11775 * Expand security boundary to match 1.1.1 module.
11779 * Remove redundant features: hash file source, editing of test vectors
11780 modify fipsld to use external fips_premain.c signature.
11784 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11785 run algorithm test programs.
11789 * Make algorithm test programs more tolerant of whitespace.
11793 * Have SSL/TLS server implementation tolerate "mismatched" record
11794 protocol version while receiving ClientHello even if the
11795 ClientHello is fragmented. (The server can't insist on the
11796 particular protocol version it has chosen before the ServerHello
11797 message has informed the client about his choice.)
11801 * Load error codes if they are not already present instead of using a
11802 static variable. This allows them to be cleanly unloaded and reloaded.
11806 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11808 * Introduce limits to prevent malicious keys being able to
11809 cause a denial of service. ([CVE-2006-2940])
11811 *Steve Henson, Bodo Moeller*
11813 * Fix ASN.1 parsing of certain invalid structures that can result
11814 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11816 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11817 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11819 * Fix SSL client code which could crash if connecting to a
11820 malicious SSLv2 server. ([CVE-2006-4343])
11822 *Tavis Ormandy and Will Drewry, Google Security Team*
11824 * Change ciphersuite string processing so that an explicit
11825 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11826 will no longer include "AES128-SHA"), and any other similar
11827 ciphersuite (same bitmap) from *other* protocol versions (so that
11828 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11829 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11830 changes from 0.9.8b and 0.9.8d.
11834 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11836 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11837 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11839 * Change the Unix randomness entropy gathering to use poll() when
11840 possible instead of select(), since the latter has some
11841 undesirable limitations.
11843 *Darryl Miles via Richard Levitte and Bodo Moeller*
11845 * Disable rogue ciphersuites:
11847 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11848 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11849 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11851 The latter two were purportedly from
11852 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11855 Also deactivate the remaining ciphersuites from
11856 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11857 unofficial, and the ID has long expired.
11861 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11862 dual-core machines) and other potential thread-safety issues.
11866 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11868 * Adapt fipsld and the build system to link against the validated FIPS
11869 module in FIPS mode.
11873 * Fixes for VC++ 2005 build under Windows.
11877 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11878 from a Windows bash shell such as MSYS. It is autodetected from the
11879 "config" script when run from a VC++ environment. Modify standard VC++
11880 build to use fipscanister.o from the GNU make build.
11884 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11886 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11887 The value now differs depending on if you build for FIPS or not.
11888 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11889 safely run with a non-FIPSed libcrypto, as it may crash because of
11890 the difference induced by this change.
11894 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11896 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11897 (part of SSL_OP_ALL). This option used to disable the
11898 countermeasure against man-in-the-middle protocol-version
11899 rollback in the SSL 2.0 server implementation, which is a bad
11900 idea. ([CVE-2005-2969])
11902 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11903 for Information Security, National Institute of Advanced Industrial
11904 Science and Technology [AIST, Japan)]*
11906 * Minimal support for X9.31 signatures and PSS padding modes. This is
11907 mainly for FIPS compliance and not fully integrated at this stage.
11911 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11912 the exponentiation using a fixed-length exponent. (Otherwise,
11913 the information leaked through timing could expose the secret key
11914 after many signatures; cf. Bleichenbacher's attack on DSA with
11919 * Make a new fixed-window mod_exp implementation the default for
11920 RSA, DSA, and DH private-key operations so that the sequence of
11921 squares and multiplies and the memory access pattern are
11922 independent of the particular secret key. This will mitigate
11923 cache-timing and potential related attacks.
11925 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11926 and this is automatically used by BN_mod_exp_mont() if the new flag
11927 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11928 will use this BN flag for private exponents unless the flag
11929 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11930 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11932 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11934 * Change the client implementation for SSLv23_method() and
11935 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11936 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11937 (Previously, the SSL 2.0 backwards compatible Client Hello
11938 message format would be used even with SSL_OP_NO_SSLv2.)
11942 * Add support for smime-type MIME parameter in S/MIME messages which some
11947 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11948 a threadsafe manner. Modify rsa code to use new function and add calls
11949 to dsa and dh code (which had race conditions before).
11953 * Include the fixed error library code in the C error file definitions
11954 instead of fixing them up at runtime. This keeps the error code
11955 structures constant.
11959 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11961 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11964 * Fixes for newer kerberos headers. NB: the casts are needed because
11965 the 'length' field is signed on one version and unsigned on another
11966 with no (?) obvious way to tell the difference, without these VC++
11967 complains. Also the "definition" of FAR (blank) is no longer included
11968 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11969 some needed definitions.
11973 * Undo Cygwin change.
11977 * Added support for proxy certificates according to RFC 3820.
11978 Because they may be a security thread to unaware applications,
11979 they must be explicitly allowed in run-time. See
11980 docs/HOWTO/proxy_certificates.txt for further information.
11984 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11986 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11987 server and client random values. Previously
11988 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11989 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11991 This change has negligible security impact because:
11993 1. Server and client random values still have 24 bytes of pseudo random
11996 2. Server and client random values are sent in the clear in the initial
11999 3. The master secret is derived using the premaster secret (48 bytes in
12000 size for static RSA ciphersuites) as well as client server and random
12003 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12006 *Stephen Henson, reported by UK NISCC*
12008 * Use Windows randomness collection on Cygwin.
12012 * Fix hang in EGD/PRNGD query when communication socket is closed
12013 prematurely by EGD/PRNGD.
12015 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12017 * Prompt for pass phrases when appropriate for PKCS12 input format.
12021 * Back-port of selected performance improvements from development
12022 branch, as well as improved support for PowerPC platforms.
12026 * Add lots of checks for memory allocation failure, error codes to indicate
12027 failure and freeing up memory if a failure occurs.
12029 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12031 * Add new -passin argument to dgst.
12035 * Perform some character comparisons of different types in X509_NAME_cmp:
12036 this is needed for some certificates that re-encode DNs into UTF8Strings
12037 (in violation of RFC3280) and can't or won't issue name rollover
12042 * Make an explicit check during certificate validation to see that
12043 the CA setting in each certificate on the chain is correct. As a
12044 side effect always do the following basic checks on extensions,
12045 not just when there's an associated purpose to the check:
12047 - if there is an unhandled critical extension (unless the user
12048 has chosen to ignore this fault)
12049 - if the path length has been exceeded (if one is set at all)
12050 - that certain extensions fit the associated purpose (if one has
12055 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12057 * Avoid a race condition when CRLs are checked in a multi threaded
12058 environment. This would happen due to the reordering of the revoked
12059 entries during signature checking and serial number lookup. Now the
12060 encoding is cached and the serial number sort performed under a lock.
12061 Add new STACK function sk_is_sorted().
12065 * Add Delta CRL to the extension code.
12069 * Various fixes to s3_pkt.c so alerts are sent properly.
12071 *David Holmes <d.holmes@f5.com>*
12073 * Reduce the chances of duplicate issuer name and serial numbers (in
12074 violation of RFC3280) using the OpenSSL certificate creation utilities.
12075 This is done by creating a random 64 bit value for the initial serial
12076 number when a serial number file is created or when a self signed
12077 certificate is created using 'openssl req -x509'. The initial serial
12078 number file is created using 'openssl x509 -next_serial' in CA.pl
12079 rather than being initialized to 1.
12083 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12085 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12086 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12088 *Joe Orton, Steve Henson*
12090 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12093 *Joe Orton, Steve Henson*
12095 * Make it possible to have multiple active certificates with the same
12096 subject in the CA index file. This is done only if the keyword
12097 'unique_subject' is set to 'no' in the main CA section (default
12098 if 'CA_default') of the configuration file. The value is saved
12099 with the database itself in a separate index attribute file,
12100 named like the index file with '.attr' appended to the name.
12104 * X509 verify fixes. Disable broken certificate workarounds when
12105 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12106 keyUsage extension present. Don't accept CRLs with unhandled critical
12107 extensions: since verify currently doesn't process CRL extensions this
12108 rejects a CRL with *any* critical extensions. Add new verify error codes
12113 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12114 A clarification of RFC2560 will require the use of OCTET STRINGs and
12115 some implementations cannot handle the current raw format. Since OpenSSL
12116 copies and compares OCSP nonces as opaque blobs without any attempt at
12117 parsing them this should not create any compatibility issues.
12121 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12122 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12123 this HMAC (and other) operations are several times slower than OpenSSL
12128 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12130 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12132 * Use the correct content when signing type "other".
12136 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12138 * Fix various bugs revealed by running the NISCC test suite:
12140 Stop out of bounds reads in the ASN1 code when presented with
12141 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12143 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12145 If verify callback ignores invalid public key errors don't try to check
12146 certificate signature with the NULL public key.
12150 * New -ignore_err option in ocsp application to stop the server
12151 exiting on the first error in a request.
12155 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12156 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12161 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12162 extra data after the compression methods not only for TLS 1.0
12163 but also for SSL 3.0 (as required by the specification).
12165 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12167 * Change X509_certificate_type() to mark the key as exported/exportable
12168 when it's 512 *bits* long, not 512 bytes.
12172 * Change AES_cbc_encrypt() so it outputs exact multiple of
12173 blocks during encryption.
12177 * Various fixes to base64 BIO and non blocking I/O. On write
12178 flushes were not handled properly if the BIO retried. On read
12179 data was not being buffered properly and had various logic bugs.
12180 This also affects blocking I/O when the data being decoded is a
12185 * Various S/MIME bugfixes and compatibility changes:
12186 output correct application/pkcs7 MIME type if
12187 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12188 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12189 of files as .eml work). Correctly handle very long lines in MIME
12194 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12196 * Countermeasure against the Klima-Pokorny-Rosa extension of
12197 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12198 a protocol version number mismatch like a decryption error
12199 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12203 * Turn on RSA blinding by default in the default implementation
12204 to avoid a timing attack. Applications that don't want it can call
12205 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12206 They would be ill-advised to do so in most cases.
12208 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12210 * Change RSA blinding code so that it works when the PRNG is not
12211 seeded (in this case, the secret RSA exponent is abused as
12212 an unpredictable seed -- if it is not unpredictable, there
12213 is no point in blinding anyway). Make RSA blinding thread-safe
12214 by remembering the creator's thread ID in rsa->blinding and
12215 having all other threads use local one-time blinding factors
12216 (this requires more computation than sharing rsa->blinding, but
12217 avoids excessive locking; and if an RSA object is not shared
12218 between threads, blinding will still be very fast).
12222 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12223 ENGINE as defaults for all supported algorithms irrespective of
12224 the 'flags' parameter. 'flags' is now honoured, so applications
12225 should make sure they are passing it correctly.
12229 * Target "mingw" now allows native Windows code to be generated in
12230 the Cygwin environment as well as with the MinGW compiler.
12234 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12236 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12237 via timing by performing a MAC computation even if incorrect
12238 block cipher padding has been found. This is a countermeasure
12239 against active attacks where the attacker has to distinguish
12240 between bad padding and a MAC verification error. ([CVE-2003-0078])
12242 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12243 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12244 Martin Vuagnoux (EPFL, Ilion)*
12246 * Make the no-err option work as intended. The intention with no-err
12247 is not to have the whole error stack handling routines removed from
12248 libcrypto, it's only intended to remove all the function name and
12249 reason texts, thereby removing some of the footprint that may not
12250 be interesting if those errors aren't displayed anyway.
12252 NOTE: it's still possible for any application or module to have its
12253 own set of error texts inserted. The routines are there, just not
12254 used by default when no-err is given.
12258 * Add support for FreeBSD on IA64.
12260 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12262 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12263 Kerberos function mit_des_cbc_cksum(). Before this change,
12264 the value returned by DES_cbc_cksum() was like the one from
12265 mit_des_cbc_cksum(), except the bytes were swapped.
12267 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12269 * Allow an application to disable the automatic SSL chain building.
12270 Before this a rather primitive chain build was always performed in
12271 ssl3_output_cert_chain(): an application had no way to send the
12272 correct chain if the automatic operation produced an incorrect result.
12274 Now the chain builder is disabled if either:
12276 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12278 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12280 The reasoning behind this is that an application would not want the
12281 auto chain building to take place if extra chain certificates are
12282 present and it might also want a means of sending no additional
12283 certificates (for example the chain has two certificates and the
12288 * Add the possibility to build without the ENGINE framework.
12290 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12292 * Under Win32 gmtime() can return NULL: check return value in
12293 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12297 * DSA routines: under certain error conditions uninitialized BN objects
12298 could be freed. Solution: make sure initialization is performed early
12299 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12300 Nils Larsch <nla@trustcenter.de> via PR#459)
12304 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12305 checked on reconnect on the client side, therefore session resumption
12306 could still fail with a "ssl session id is different" error. This
12307 behaviour is masked when SSL_OP_ALL is used due to
12308 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12309 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12310 followup to PR #377.
12314 * IA-32 assembler support enhancements: unified ELF targets, support
12315 for SCO/Caldera platforms, fix for Cygwin shared build.
12319 * Add support for FreeBSD on sparc64. As a consequence, support for
12320 FreeBSD on non-x86 processors is separate from x86 processors on
12321 the config script, much like the NetBSD support.
12323 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12325 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12327 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12330 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12331 code (06) was taken as the first octet of the session ID and the last
12332 octet was ignored consequently. As a result SSLv2 client side session
12333 caching could not have worked due to the session ID mismatch between
12335 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12340 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12341 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12346 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12347 seems that in spite of existing for more than a year, many application
12348 author have done nothing to provide the necessary callbacks, which
12349 means that this particular engine will not work properly anywhere.
12350 This is a very unfortunate situation which forces us, in the name
12351 of usability, to give the hw_ncipher.c a static lock, which is part
12353 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12354 appear in 0.9.8 or later. We EXPECT application authors to have
12355 dealt properly with this when 0.9.8 is released (unless we actually
12356 make such changes in the libcrypto locking code that changes will
12357 have to be made anyway).
12361 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12362 octets have been read, EOF or an error occurs. Without this change
12363 some truncated ASN1 structures will not produce an error.
12367 * Disable Heimdal support, since it hasn't been fully implemented.
12368 Still give the possibility to force the use of Heimdal, but with
12369 warnings and a request that patches get sent to openssl-dev.
12373 * Add the VC-CE target, introduce the WINCE sysname, and add
12374 INSTALL.WCE and appropriate conditionals to make it build.
12376 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12378 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12379 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12380 edit numbers of the version.
12382 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12384 * Introduce safe string copy and catenation functions
12385 (BUF_strlcpy() and BUF_strlcat()).
12387 *Ben Laurie (CHATS) and Richard Levitte*
12389 * Avoid using fixed-size buffers for one-line DNs.
12391 *Ben Laurie (CHATS)*
12393 * Add BUF_MEM_grow_clean() to avoid information leakage when
12394 resizing buffers containing secrets, and use where appropriate.
12396 *Ben Laurie (CHATS)*
12398 * Avoid using fixed size buffers for configuration file location.
12400 *Ben Laurie (CHATS)*
12402 * Avoid filename truncation for various CA files.
12404 *Ben Laurie (CHATS)*
12406 * Use sizeof in preference to magic numbers.
12408 *Ben Laurie (CHATS)*
12410 * Avoid filename truncation in cert requests.
12412 *Ben Laurie (CHATS)*
12414 * Add assertions to check for (supposedly impossible) buffer
12417 *Ben Laurie (CHATS)*
12419 * Don't cache truncated DNS entries in the local cache (this could
12420 potentially lead to a spoofing attack).
12422 *Ben Laurie (CHATS)*
12424 * Fix various buffers to be large enough for hex/decimal
12425 representations in a platform independent manner.
12427 *Ben Laurie (CHATS)*
12429 * Add CRYPTO_realloc_clean() to avoid information leakage when
12430 resizing buffers containing secrets, and use where appropriate.
12432 *Ben Laurie (CHATS)*
12434 * Add BIO_indent() to avoid much slightly worrying code to do
12437 *Ben Laurie (CHATS)*
12439 * Convert sprintf()/BIO_puts() to BIO_printf().
12441 *Ben Laurie (CHATS)*
12443 * buffer_gets() could terminate with the buffer only half
12446 *Ben Laurie (CHATS)*
12448 * Add assertions to prevent user-supplied crypto functions from
12449 overflowing internal buffers by having large block sizes, etc.
12451 *Ben Laurie (CHATS)*
12453 * New OPENSSL_assert() macro (similar to assert(), but enabled
12456 *Ben Laurie (CHATS)*
12458 * Eliminate unused copy of key in RC4.
12460 *Ben Laurie (CHATS)*
12462 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12464 *Ben Laurie (CHATS)*
12466 * Fix off-by-one error in EGD path.
12468 *Ben Laurie (CHATS)*
12470 * If RANDFILE path is too long, ignore instead of truncating.
12472 *Ben Laurie (CHATS)*
12474 * Eliminate unused and incorrectly sized X.509 structure
12477 *Ben Laurie (CHATS)*
12479 * Eliminate unused and dangerous function knumber().
12481 *Ben Laurie (CHATS)*
12483 * Eliminate unused and dangerous structure, KSSL_ERR.
12485 *Ben Laurie (CHATS)*
12487 * Protect against overlong session ID context length in an encoded
12488 session object. Since these are local, this does not appear to be
12491 *Ben Laurie (CHATS)*
12493 * Change from security patch (see 0.9.6e below) that did not affect
12494 the 0.9.6 release series:
12496 Remote buffer overflow in SSL3 protocol - an attacker could
12497 supply an oversized master key in Kerberos-enabled versions.
12500 *Ben Laurie (CHATS)*
12502 * Change the SSL kerb5 codes to match RFC 2712.
12506 * Make -nameopt work fully for req and add -reqopt switch.
12508 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12510 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12512 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12514 * Make sure tests can be performed even if the corresponding algorithms
12515 have been removed entirely. This was also the last step to make
12516 OpenSSL compilable with DJGPP under all reasonable conditions.
12518 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12520 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12521 to allow version independent disabling of normally unselected ciphers,
12522 which may be activated as a side-effect of selecting a single cipher.
12524 (E.g., cipher list string "RSA" enables ciphersuites that are left
12525 out of "ALL" because they do not provide symmetric encryption.
12526 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12528 *Lutz Jaenicke, Bodo Moeller*
12530 * Add appropriate support for separate platform-dependent build
12531 directories. The recommended way to make a platform-dependent
12532 build directory is the following (tested on Linux), maybe with
12535 # Place yourself outside of the OpenSSL source tree. In
12536 # this example, the environment variable OPENSSL_SOURCE
12537 # is assumed to contain the absolute OpenSSL source directory.
12538 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12539 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12540 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12541 mkdir -p `dirname $F`
12542 ln -s $OPENSSL_SOURCE/$F $F
12545 To be absolutely sure not to disturb the source tree, a "make clean"
12546 is a good thing. If it isn't successful, don't worry about it,
12547 it probably means the source directory is very clean.
12551 * Make sure any ENGINE control commands make local copies of string
12552 pointers passed to them whenever necessary. Otherwise it is possible
12553 the caller may have overwritten (or deallocated) the original string
12554 data when a later ENGINE operation tries to use the stored values.
12556 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12558 * Improve diagnostics in file reading and command-line digests.
12560 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12562 * Add AES modes CFB and OFB to the object database. Correct an
12563 error in AES-CFB decryption.
12567 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12568 allows existing EVP_CIPHER_CTX structures to be reused after
12569 calling `EVP_*Final()`. This behaviour is used by encryption
12570 BIOs and some applications. This has the side effect that
12571 applications must explicitly clean up cipher contexts with
12572 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12576 * Check the values of dna and dnb in bn_mul_recursive before calling
12577 bn_mul_comba (a non zero value means the a or b arrays do not contain
12578 n2 elements) and fallback to bn_mul_normal if either is not zero.
12582 * Fix escaping of non-ASCII characters when using the -subj option
12583 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12587 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12588 form for "surname", serialNumber has no short form.
12589 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12590 therefore remove "mail" short name for "internet 7".
12591 The OID for unique identifiers in X509 certificates is
12592 x500UniqueIdentifier, not uniqueIdentifier.
12593 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12597 * Add an "init" command to the ENGINE config module and auto initialize
12598 ENGINEs. Without any "init" command the ENGINE will be initialized
12599 after all ctrl commands have been executed on it. If init=1 the
12600 ENGINE is initialized at that point (ctrls before that point are run
12601 on the uninitialized ENGINE and after on the initialized one). If
12602 init=0 then the ENGINE will not be initialized at all.
12606 * Fix the 'app_verify_callback' interface so that the user-defined
12607 argument is actually passed to the callback: In the
12608 SSL_CTX_set_cert_verify_callback() prototype, the callback
12609 declaration has been changed from
12612 int (*cb)(X509_STORE_CTX *,void *);
12613 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12614 i=s->ctx->app_verify_callback(&ctx)
12615 has been changed into
12616 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12618 To update applications using SSL_CTX_set_cert_verify_callback(),
12619 a dummy argument can be added to their callback functions.
12621 *D. K. Smetters <smetters@parc.xerox.com>*
12623 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12625 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12627 * Add and OPENSSL_LOAD_CONF define which will cause
12628 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12629 This allows older applications to transparently support certain
12630 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12631 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12632 load the config file and OPENSSL_add_all_algorithms_conf() which will
12633 always load it have also been added.
12637 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12638 Adjust NIDs and EVP layer.
12640 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12642 * Config modules support in openssl utility.
12644 Most commands now load modules from the config file,
12645 though in a few (such as version) this isn't done
12646 because it couldn't be used for anything.
12648 In the case of ca and req the config file used is
12649 the same as the utility itself: that is the -config
12650 command line option can be used to specify an
12655 * Move default behaviour from OPENSSL_config(). If appname is NULL
12656 use "openssl_conf" if filename is NULL use default openssl config file.
12660 * Add an argument to OPENSSL_config() to allow the use of an alternative
12661 config section name. Add a new flag to tolerate a missing config file
12662 and move code to CONF_modules_load_file().
12666 * Support for crypto accelerator cards from Accelerated Encryption
12667 Processing, www.aep.ie. (Use engine 'aep')
12668 The support was copied from 0.9.6c [engine] and adapted/corrected
12669 to work with the new engine framework.
12671 *AEP Inc. and Richard Levitte*
12673 * Support for SureWare crypto accelerator cards from Baltimore
12674 Technologies. (Use engine 'sureware')
12675 The support was copied from 0.9.6c [engine] and adapted
12676 to work with the new engine framework.
12680 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12681 make the newer ENGINE framework commands for the CHIL engine work.
12683 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12685 * Make it possible to produce shared libraries on ReliantUNIX.
12687 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12689 * Add the configuration target debug-linux-ppro.
12690 Make 'openssl rsa' use the general key loading routines
12691 implemented in `apps.c`, and make those routines able to
12692 handle the key format FORMAT_NETSCAPE and the variant
12695 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12697 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12699 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12701 * Add -keyform to rsautl, and document -engine.
12703 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12705 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12706 BIO_R_NO_SUCH_FILE error code rather than the generic
12707 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12711 * Add new functions
12712 ERR_peek_last_error
12713 ERR_peek_last_error_line
12714 ERR_peek_last_error_line_data.
12715 These are similar to
12717 ERR_peek_error_line
12718 ERR_peek_error_line_data,
12719 but report on the latest error recorded rather than the first one
12720 still in the error queue.
12722 *Ben Laurie, Bodo Moeller*
12724 * default_algorithms option in ENGINE config module. This allows things
12726 default_algorithms = ALL
12727 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12731 * Preliminary ENGINE config module.
12735 * New experimental application configuration code.
12739 * Change the AES code to follow the same name structure as all other
12740 symmetric ciphers, and behave the same way. Move everything to
12741 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12743 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12745 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12747 *Ben Laurie and Theo de Raadt*
12749 * Add option to output public keys in req command.
12751 *Massimiliano Pala madwolf@openca.org*
12753 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12754 (up to about 10% better than before for P-192 and P-224).
12758 * New functions/macros
12760 SSL_CTX_set_msg_callback(ctx, cb)
12761 SSL_CTX_set_msg_callback_arg(ctx, arg)
12762 SSL_set_msg_callback(ssl, cb)
12763 SSL_set_msg_callback_arg(ssl, arg)
12765 to request calling a callback function
12767 void cb(int write_p, int version, int content_type,
12768 const void *buf, size_t len, SSL *ssl, void *arg)
12770 whenever a protocol message has been completely received
12771 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12772 protocol version according to which the SSL library interprets
12773 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12774 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12775 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12776 specification (change_cipher_spec(20), alert(21), handshake(22)).
12777 'buf' and 'len' point to the actual message, 'ssl' to the
12778 SSL object, and 'arg' is the application-defined value set by
12779 SSL[_CTX]_set_msg_callback_arg().
12781 'openssl s_client' and 'openssl s_server' have new '-msg' options
12782 to enable a callback that displays all protocol messages.
12786 * Change the shared library support so shared libraries are built as
12787 soon as the corresponding static library is finished, and thereby get
12788 openssl and the test programs linked against the shared library.
12789 This still only happens when the keyword "shard" has been given to
12790 the configuration scripts.
12792 NOTE: shared library support is still an experimental thing, and
12793 backward binary compatibility is still not guaranteed.
12795 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12797 * Add support for Subject Information Access extension.
12799 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12801 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12802 additional bytes when new memory had to be allocated, not just
12803 when reusing an existing buffer.
12807 * New command line and configuration option 'utf8' for the req command.
12808 This allows field values to be specified as UTF8 strings.
12812 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12813 runs for the former and machine-readable output for the latter.
12817 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12818 of the e-mail address in the DN (i.e., it will go into a certificate
12819 extension only). The new configuration file option 'email_in_dn = no'
12820 has the same effect.
12822 *Massimiliano Pala madwolf@openca.org*
12824 * Change all functions with names starting with `des_` to be starting
12825 with `DES_` instead. Add wrappers that are compatible with libdes,
12826 but are named `_ossl_old_des_*`. Finally, add macros that map the
12827 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12828 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12829 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12832 Since we provide two compatibility mappings, the user needs to
12833 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12834 compatibility is desired. The default (i.e., when that macro
12835 isn't defined) is OpenSSL 0.9.6c compatibility.
12837 There are also macros that enable and disable the support of old
12838 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12839 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12840 are defined, the default will apply: to support the old des routines.
12842 In either case, one must include openssl/des.h to get the correct
12843 definitions. Do not try to just include openssl/des_old.h, that
12846 NOTE: This is a major break of an old API into a new one. Software
12847 authors are encouraged to switch to the `DES_` style functions. Some
12848 time in the future, des_old.h and the libdes compatibility functions
12849 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12850 default), and then completely removed.
12854 * Test for certificates which contain unsupported critical extensions.
12855 If such a certificate is found during a verify operation it is
12856 rejected by default: this behaviour can be overridden by either
12857 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12858 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12859 X509_supported_extension() has also been added which returns 1 if a
12860 particular extension is supported.
12864 * Modify the behaviour of EVP cipher functions in similar way to digests
12865 to retain compatibility with existing code.
12869 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12870 compatibility with existing code. In particular the 'ctx' parameter does
12871 not have to be to be initialized before the call to EVP_DigestInit() and
12872 it is tidied up after a call to EVP_DigestFinal(). New function
12873 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12874 EVP_MD_CTX_copy() changed to not require the destination to be
12875 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12876 requires the destination to be valid.
12878 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12879 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12883 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12884 so that complete 'Handshake' protocol structures are kept in memory
12885 instead of overwriting 'msg_type' and 'length' with 'body' data.
12889 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12891 *Massimo Santin via Richard Levitte*
12893 * Major restructuring to the underlying ENGINE code. This includes
12894 reduction of linker bloat, separation of pure "ENGINE" manipulation
12895 (initialisation, etc) from functionality dealing with implementations
12896 of specific crypto interfaces. This change also introduces integrated
12897 support for symmetric ciphers and digest implementations - so ENGINEs
12898 can now accelerate these by providing EVP_CIPHER and EVP_MD
12899 implementations of their own. This is detailed in
12900 [crypto/engine/README.md](crypto/engine/README.md)
12901 as it couldn't be adequately described here. However, there are a few
12902 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12903 were changed in the original introduction of ENGINE code have now
12904 reverted back - the hooking from this code to ENGINE is now a good
12905 deal more passive and at run-time, operations deal directly with
12906 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12907 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12908 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12909 they were not being used by the framework as there is no concept of a
12910 BIGNUM_METHOD and they could not be generalised to the new
12911 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12912 ENGINE_cpy() has been removed as it cannot be consistently defined in
12917 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12921 * Change mkdef.pl to sort symbols that get the same entry number,
12922 and make sure the automatically generated functions `ERR_load_*`
12923 become part of libeay.num as well.
12927 * New function SSL_renegotiate_pending(). This returns true once
12928 renegotiation has been requested (either SSL_renegotiate() call
12929 or HelloRequest/ClientHello received from the peer) and becomes
12930 false once a handshake has been completed.
12931 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12932 sends a HelloRequest, but does not ensure that a handshake takes
12933 place. SSL_renegotiate_pending() is useful for checking if the
12934 client has followed the request.)
12938 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12939 By default, clients may request session resumption even during
12940 renegotiation (if session ID contexts permit); with this option,
12941 session resumption is possible only in the first handshake.
12943 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12944 more bits available for options that should not be part of
12945 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12949 * Add some demos for certificate and certificate request creation.
12953 * Make maximum certificate chain size accepted from the peer application
12954 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12955 "Douglas E. Engert" <deengert@anl.gov>.
12959 * Add support for shared libraries for Unixware-7
12960 (Boyd Lynn Gerber <gerberb@zenez.com>).
12964 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12965 be done prior to destruction. Use this to unload error strings from
12966 ENGINEs that load their own error strings. NB: This adds two new API
12967 functions to "get" and "set" this destroy handler in an ENGINE.
12971 * Alter all existing ENGINE implementations (except "openssl" and
12972 "openbsd") to dynamically instantiate their own error strings. This
12973 makes them more flexible to be built both as statically-linked ENGINEs
12974 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12975 Also, add stub code to each that makes building them as self-contained
12976 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12980 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12981 implementations into applications that are completely implemented in
12982 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12983 commands that can be used to configure what shared-library to load and
12984 to control aspects of the way it is handled. Also, made an update to
12985 the [README-Engine.md](README-Engine.md) file
12986 that brings its information up-to-date and
12987 provides some information and instructions on the "dynamic" ENGINE
12988 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12992 * Make it possible to unload ranges of ERR strings with a new
12993 "ERR_unload_strings" function.
12997 * Add a copy() function to EVP_MD.
13001 * Make EVP_MD routines take a context pointer instead of just the
13002 md_data void pointer.
13006 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13007 that the digest can only process a single chunk of data
13008 (typically because it is provided by a piece of
13009 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13010 is only going to provide a single chunk of data, and hence the
13011 framework needn't accumulate the data for oneshot drivers.
13015 * As with "ERR", make it possible to replace the underlying "ex_data"
13016 functions. This change also alters the storage and management of global
13017 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13018 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13019 index counters. The API functions that use this state have been changed
13020 to take a "class_index" rather than pointers to the class's local STACK
13021 and counter, and there is now an API function to dynamically create new
13022 classes. This centralisation allows us to (a) plug a lot of the
13023 thread-safety problems that existed, and (b) makes it possible to clean
13024 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13025 such data would previously have always leaked in application code and
13026 workarounds were in place to make the memory debugging turn a blind eye
13027 to it. Application code that doesn't use this new function will still
13028 leak as before, but their memory debugging output will announce it now
13029 rather than letting it slide.
13031 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13032 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13033 has a return value to indicate success or failure.
13037 * Make it possible to replace the underlying "ERR" functions such that the
13038 global state (2 LHASH tables and 2 locks) is only used by the "default"
13039 implementation. This change also adds two functions to "get" and "set"
13040 the implementation prior to it being automatically set the first time
13041 any other ERR function takes place. Ie. an application can call "get",
13042 pass the return value to a module it has just loaded, and that module
13043 can call its own "set" function using that value. This means the
13044 module's "ERR" operations will use (and modify) the error state in the
13045 application and not in its own statically linked copy of OpenSSL code.
13049 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13050 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13051 the operation, and provides a more encapsulated way for external code
13052 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13053 to use these functions rather than manually incrementing the counts.
13055 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13059 * Add EVP test program.
13063 * Add symmetric cipher support to ENGINE. Expect the API to change!
13067 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13068 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13069 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13070 These allow a CRL to be built without having to access X509_CRL fields
13071 directly. Modify 'ca' application to use new functions.
13075 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13076 bug workarounds. Rollback attack detection is a security feature.
13077 The problem will only arise on OpenSSL servers when TLSv1 is not
13078 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13079 Software authors not wanting to support TLSv1 will have special reasons
13080 for their choice and can explicitly enable this option.
13082 *Bodo Moeller, Lutz Jaenicke*
13084 * Rationalise EVP so it can be extended: don't include a union of
13085 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13086 (similar to those existing for EVP_CIPHER_CTX).
13091 EVP_MD_CTX_init(&md); /* new function call */
13092 EVP_DigestInit(&md, EVP_sha1());
13093 EVP_DigestUpdate(&md, in, len);
13094 EVP_DigestFinal(&md, out, NULL);
13095 EVP_MD_CTX_cleanup(&md); /* new function call */
13099 * Make DES key schedule conform to the usual scheme, as well as
13100 correcting its structure. This means that calls to DES functions
13101 now have to pass a pointer to a des_key_schedule instead of a
13102 plain des_key_schedule (which was actually always a pointer
13105 des_key_schedule ks;
13107 des_set_key_checked(..., &ks);
13108 des_ncbc_encrypt(..., &ks, ...);
13110 (Note that a later change renames 'des_...' into 'DES_...'.)
13114 * Initial reduction of linker bloat: the use of some functions, such as
13115 PEM causes large amounts of unused functions to be linked in due to
13116 poor organisation. For example pem_all.c contains every PEM function
13117 which has a knock on effect of linking in large amounts of (unused)
13118 ASN1 code. Grouping together similar functions and splitting unrelated
13119 functions prevents this.
13123 * Cleanup of EVP macros.
13127 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13128 correct `_ecb suffix`.
13132 * Add initial OCSP responder support to ocsp application. The
13133 revocation information is handled using the text based index
13134 use by the ca application. The responder can either handle
13135 requests generated internally, supplied in files (for example
13136 via a CGI script) or using an internal minimal server.
13140 * Add configuration choices to get zlib compression for TLS.
13144 * Changes to Kerberos SSL for RFC 2712 compliance:
13145 1. Implemented real KerberosWrapper, instead of just using
13146 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13147 2. Implemented optional authenticator field of KerberosWrapper.
13149 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13150 and authenticator structs; see crypto/krb5/.
13152 Generalized Kerberos calls to support multiple Kerberos libraries.
13153 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13154 via Richard Levitte*
13156 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13157 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13158 values for each of the key sizes rather than having just
13159 parameters (and 'speed' generating keys each time).
13163 * Speed up EVP routines.
13166 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13167 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13168 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13169 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13171 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13172 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13173 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13176 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13178 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13182 * Added the OS2-EMX target.
13184 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13186 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13187 New functions to support `NCONF` routines in extension code.
13188 New function `CONF_set_nconf()`
13189 to allow functions which take an `NCONF` to also handle the old `LHASH`
13190 structure: this means that the old `CONF` compatible routines can be
13191 retained (in particular w.rt. extensions) without having to duplicate the
13192 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13196 * Enhance the general user interface with mechanisms for inner control
13197 and with possibilities to have yes/no kind of prompts.
13201 * Change all calls to low-level digest routines in the library and
13202 applications to use EVP. Add missing calls to HMAC_cleanup() and
13203 don't assume HMAC_CTX can be copied using memcpy().
13205 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13207 * Add the possibility to control engines through control names but with
13208 arbitrary arguments instead of just a string.
13209 Change the key loaders to take a UI_METHOD instead of a callback
13210 function pointer. NOTE: this breaks binary compatibility with earlier
13211 versions of OpenSSL [engine].
13212 Adapt the nCipher code for these new conditions and add a card insertion
13217 * Enhance the general user interface with mechanisms to better support
13218 dialog box interfaces, application-defined prompts, the possibility
13219 to use defaults (for example default passwords from somewhere else)
13220 and interrupts/cancellations.
13224 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13225 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13229 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13230 tidy up some unnecessarily weird code in 'sk_new()').
13232 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13234 * Change the key loading routines for ENGINEs to use the same kind
13235 callback (pem_password_cb) as all other routines that need this
13240 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13241 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13242 than this minimum value is recommended.
13246 * New random seeder for OpenVMS, using the system process statistics
13247 that are easily reachable.
13251 * Windows apparently can't transparently handle global
13252 variables defined in DLLs. Initialisations such as:
13254 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13256 won't compile. This is used by the any applications that need to
13257 declare their own ASN1 modules. This was fixed by adding the option
13258 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13259 needed for static libraries under Win32.
13263 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13264 setting of purpose and trust fields. New X509_STORE trust and
13265 purpose functions and tidy up setting in other SSL functions.
13269 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13270 structure. These are inherited by X509_STORE_CTX when it is
13271 initialised. This allows various defaults to be set in the
13272 X509_STORE structure (such as flags for CRL checking and custom
13273 purpose or trust settings) for functions which only use X509_STORE_CTX
13274 internally such as S/MIME.
13276 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13277 trust settings if they are not set in X509_STORE. This allows X509_STORE
13278 purposes and trust (in S/MIME for example) to override any set by default.
13280 Add command line options for CRL checking to smime, s_client and s_server
13285 * Initial CRL based revocation checking. If the CRL checking flag(s)
13286 are set then the CRL is looked up in the X509_STORE structure and
13287 its validity and signature checked, then if the certificate is found
13288 in the CRL the verify fails with a revoked error.
13290 Various new CRL related callbacks added to X509_STORE_CTX structure.
13292 Command line options added to 'verify' application to support this.
13294 This needs some additional work, such as being able to handle multiple
13295 CRLs with different times, extension based lookup (rather than just
13296 by subject name) and ultimately more complete V2 CRL extension
13301 * Add a general user interface API (crypto/ui/). This is designed
13302 to replace things like des_read_password and friends (backward
13303 compatibility functions using this new API are provided).
13304 The purpose is to remove prompting functions from the DES code
13305 section as well as provide for prompting through dialog boxes in
13306 a window system and the like.
13310 * Add "ex_data" support to ENGINE so implementations can add state at a
13311 per-structure level rather than having to store it globally.
13315 * Make it possible for ENGINE structures to be copied when retrieved by
13316 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13317 This causes the "original" ENGINE structure to act like a template,
13318 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13319 operational state can be localised to each ENGINE structure, despite the
13320 fact they all share the same "methods". New ENGINE structures returned in
13321 this case have no functional references and the return value is the single
13322 structural reference. This matches the single structural reference returned
13323 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13328 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13329 needs to match any other type at all we need to manually clear the
13334 * Changes to the "openssl engine" utility to include;
13335 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13336 about an ENGINE's available control commands.
13337 - executing control commands from command line arguments using the
13338 '-pre' and '-post' switches. '-post' is only used if '-t' is
13339 specified and the ENGINE is successfully initialised. The syntax for
13340 the individual commands are colon-separated, for example;
13341 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13345 * New dynamic control command support for ENGINEs. ENGINEs can now
13346 declare their own commands (numbers), names (strings), descriptions,
13347 and input types for run-time discovery by calling applications. A
13348 subset of these commands are implicitly classed as "executable"
13349 depending on their input type, and only these can be invoked through
13350 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13351 can be based on user input, config files, etc). The distinction is
13352 that "executable" commands cannot return anything other than a boolean
13353 result and can only support numeric or string input, whereas some
13354 discoverable commands may only be for direct use through
13355 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13356 pointers, or other custom uses. The "executable" commands are to
13357 support parameterisations of ENGINE behaviour that can be
13358 unambiguously defined by ENGINEs and used consistently across any
13359 OpenSSL-based application. Commands have been added to all the
13360 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13361 control over shared-library paths without source code alterations.
13365 * Changed all ENGINE implementations to dynamically allocate their
13366 ENGINEs rather than declaring them statically. Apart from this being
13367 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13368 this also allows the implementations to compile without using the
13369 internal engine_int.h header.
13373 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13374 'const' value. Any code that should be able to modify a RAND_METHOD
13375 should already have non-const pointers to it (ie. they should only
13376 modify their own ones).
13380 * Made a variety of little tweaks to the ENGINE code.
13381 - "atalla" and "ubsec" string definitions were moved from header files
13382 to C code. "nuron" string definitions were placed in variables
13383 rather than hard-coded - allowing parameterisation of these values
13384 later on via ctrl() commands.
13385 - Removed unused "#if 0"'d code.
13386 - Fixed engine list iteration code so it uses ENGINE_free() to release
13387 structural references.
13388 - Constified the RAND_METHOD element of ENGINE structures.
13389 - Constified various get/set functions as appropriate and added
13390 missing functions (including a catch-all ENGINE_cpy that duplicates
13391 all ENGINE values onto a new ENGINE except reference counts/state).
13392 - Removed NULL parameter checks in get/set functions. Setting a method
13393 or function to NULL is a way of cancelling out a previously set
13394 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13395 and doesn't justify the extra error symbols and code.
13396 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13397 flags from engine_int.h to engine.h.
13398 - Changed prototypes for ENGINE handler functions (init(), finish(),
13399 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13403 * Implement binary inversion algorithm for BN_mod_inverse in addition
13404 to the algorithm using long division. The binary algorithm can be
13405 used only if the modulus is odd. On 32-bit systems, it is faster
13406 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13407 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13408 up to 450 bits. In 64-bit environments, the binary algorithm
13409 appears to be advantageous for much longer moduli; here we use it
13410 for moduli up to 2048 bits.
13414 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13415 could not support the combine flag in choice fields.
13419 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13420 extensions from a certificate request to the certificate.
13424 * Allow multiple 'certopt' and 'nameopt' options to be separated
13425 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13426 file: this allows the display of the certificate about to be
13427 signed to be customised, to allow certain fields to be included
13428 or excluded and extension details. The old system didn't display
13429 multicharacter strings properly, omitted fields not in the policy
13430 and couldn't display additional details such as extensions.
13434 * Function EC_POINTs_mul for multiple scalar multiplication
13435 of an arbitrary number of elliptic curve points
13436 \sum scalars[i]*points[i],
13437 optionally including the generator defined for the EC_GROUP:
13438 scalar*generator + \sum scalars[i]*points[i].
13440 EC_POINT_mul is a simple wrapper function for the typical case
13441 that the point list has just one item (besides the optional
13446 * First EC_METHODs for curves over GF(p):
13448 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13449 operations and provides various method functions that can also
13450 operate with faster implementations of modular arithmetic.
13452 EC_GFp_mont_method() reuses most functions that are part of
13453 EC_GFp_simple_method, but uses Montgomery arithmetic.
13455 *Bodo Moeller; point addition and point doubling
13456 implementation directly derived from source code provided by
13457 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13459 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13460 crypto/ec/ec_lib.c):
13462 Curves are EC_GROUP objects (with an optional group generator)
13463 based on EC_METHODs that are built into the library.
13465 Points are EC_POINT objects based on EC_GROUP objects.
13467 Most of the framework would be able to handle curves over arbitrary
13468 finite fields, but as there are no obvious types for fields other
13469 than GF(p), some functions are limited to that for now.
13473 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13474 that the file contains a complete HTTP response.
13478 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13479 change the def and num file printf format specifier from "%-40sXXX"
13480 to "%-39s XXX". The latter will always guarantee a space after the
13481 field while the former will cause them to run together if the field
13482 is 40 of more characters long.
13486 * Constify the cipher and digest 'method' functions and structures
13487 and modify related functions to take constant EVP_MD and EVP_CIPHER
13492 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13493 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13497 * Modify `EVP_Digest*()` routines so they now return values. Although the
13498 internal software routines can never fail additional hardware versions
13503 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13505 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13506 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13509 ERR_R_NESTED_ASN1_ERROR
13511 ERR_R_MISSING_ASN1_EOS
13512 were 4 .. 9, conflicting with
13513 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13515 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13516 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13518 Add new error code 'ERR_R_INTERNAL_ERROR'.
13522 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13527 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13528 sets the subject name for a new request or supersedes the
13529 subject name in a given request. Formats that can be parsed are
13530 'CN=Some Name, OU=myOU, C=IT'
13532 'CN=Some Name/OU=myOU/C=IT'.
13534 Add options '-batch' and '-verbose' to 'openssl req'.
13536 *Massimiliano Pala <madwolf@hackmasters.net>*
13538 * Introduce the possibility to access global variables through
13539 functions on platform were that's the best way to handle exporting
13540 global variables in shared libraries. To enable this functionality,
13541 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13542 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13543 is normally done by Configure or something similar).
13545 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13546 in the source file (foo.c) like this:
13548 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13549 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13551 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13552 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13554 OPENSSL_DECLARE_GLOBAL(int,foo);
13555 #define foo OPENSSL_GLOBAL_REF(foo)
13556 OPENSSL_DECLARE_GLOBAL(double,bar);
13557 #define bar OPENSSL_GLOBAL_REF(bar)
13559 The #defines are very important, and therefore so is including the
13560 header file everywhere where the defined globals are used.
13562 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13563 of ASN.1 items, but that structure is a bit different.
13565 The largest change is in util/mkdef.pl which has been enhanced with
13566 better and easier to understand logic to choose which symbols should
13567 go into the Windows .def files as well as a number of fixes and code
13568 cleanup (among others, algorithm keywords are now sorted
13569 lexicographically to avoid constant rewrites).
13573 * In BN_div() keep a copy of the sign of 'num' before writing the
13574 result to 'rm' because if rm==num the value will be overwritten
13575 and produce the wrong result if 'num' is negative: this caused
13576 problems with BN_mod() and BN_nnmod().
13580 * Function OCSP_request_verify(). This checks the signature on an
13581 OCSP request and verifies the signer certificate. The signer
13582 certificate is just checked for a generic purpose and OCSP request
13587 * Add OCSP_check_validity() function to check the validity of OCSP
13588 responses. OCSP responses are prepared in real time and may only
13589 be a few seconds old. Simply checking that the current time lies
13590 between thisUpdate and nextUpdate max reject otherwise valid responses
13591 caused by either OCSP responder or client clock inaccuracy. Instead
13592 we allow thisUpdate and nextUpdate to fall within a certain period of
13593 the current time. The age of the response can also optionally be
13594 checked. Two new options -validity_period and -status_age added to
13599 * If signature or public key algorithm is unrecognized print out its
13600 OID rather that just UNKNOWN.
13604 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13605 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13606 ID to be generated from the issuer certificate alone which can then be
13607 passed to OCSP_id_issuer_cmp().
13611 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13612 ASN1 modules to export functions returning ASN1_ITEM pointers
13613 instead of the ASN1_ITEM structures themselves. This adds several
13614 new macros which allow the underlying ASN1 function/structure to
13615 be accessed transparently. As a result code should not use ASN1_ITEM
13616 references directly (such as &X509_it) but instead use the relevant
13617 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13618 use of the new ASN1 code on platforms where exporting structures
13619 is problematical (for example in shared libraries) but exporting
13620 functions returning pointers to structures is not.
13624 * Add support for overriding the generation of SSL/TLS session IDs.
13625 These callbacks can be registered either in an SSL_CTX or per SSL.
13626 The purpose of this is to allow applications to control, if they wish,
13627 the arbitrary values chosen for use as session IDs, particularly as it
13628 can be useful for session caching in multiple-server environments. A
13629 command-line switch for testing this (and any client code that wishes
13630 to use such a feature) has been added to "s_server".
13632 *Geoff Thorpe, Lutz Jaenicke*
13634 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13635 of the form `#if defined(...) || defined(...) || ...` and
13636 `#if !defined(...) && !defined(...) && ...`. This also avoids
13637 the growing number of special cases it was previously handling.
13641 * Make all configuration macros available for application by making
13642 sure they are available in opensslconf.h, by giving them names starting
13643 with `OPENSSL_` to avoid conflicts with other packages and by making
13644 sure e_os2.h will cover all platform-specific cases together with
13646 Additionally, it is now possible to define configuration/platform-
13647 specific names (called "system identities"). In the C code, these
13648 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13649 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13650 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13655 * New option -set_serial to 'req' and 'x509' this allows the serial
13656 number to use to be specified on the command line. Previously self
13657 signed certificates were hard coded with serial number 0 and the
13658 CA options of 'x509' had to use a serial number in a file which was
13663 * New options to 'ca' utility to support V2 CRL entry extensions.
13664 Currently CRL reason, invalidity date and hold instruction are
13665 supported. Add new CRL extensions to V3 code and some new objects.
13669 * New function EVP_CIPHER_CTX_set_padding() this is used to
13670 disable standard block padding (aka PKCS#5 padding) in the EVP
13671 API, which was previously mandatory. This means that the data is
13672 not padded in any way and so the total length much be a multiple
13673 of the block size, otherwise an error occurs.
13677 * Initial (incomplete) OCSP SSL support.
13681 * New function OCSP_parse_url(). This splits up a URL into its host,
13682 port and path components: primarily to parse OCSP URLs. New -url
13683 option to ocsp utility.
13687 * New nonce behavior. The return value of OCSP_check_nonce() now
13688 reflects the various checks performed. Applications can decide
13689 whether to tolerate certain situations such as an absent nonce
13690 in a response when one was present in a request: the ocsp application
13691 just prints out a warning. New function OCSP_add1_basic_nonce()
13692 this is to allow responders to include a nonce in a response even if
13693 the request is nonce-less.
13697 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13698 skipped when using openssl x509 multiple times on a single input file,
13699 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13703 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13704 set string type: to handle setting ASN1_TIME structures. Fix ca
13705 utility to correctly initialize revocation date of CRLs.
13709 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13710 the clients preferred ciphersuites and rather use its own preferences.
13711 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13712 Internet Explorer by ensuring unchanged hash method during stepup.
13713 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13717 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13718 to aes and add a new 'exist' option to print out symbols that don't
13723 * Additional options to ocsp utility to allow flags to be set and
13724 additional certificates supplied.
13728 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13729 OCSP client a number of certificate to only verify the response
13734 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13735 handle the new API. Currently only ECB, CBC modes supported. Add new
13738 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13739 Encryption Standard (AES) Ciphersuites for Transport Layer
13740 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13741 not enabled by default and were not part of the "ALL" ciphersuite
13742 alias because they were not yet official; they could be
13743 explicitly requested by specifying the "AESdraft" ciphersuite
13744 group alias. In the final release of OpenSSL 0.9.7, the group
13745 alias is called "AES" and is part of "ALL".)
13747 *Ben Laurie, Steve Henson, Bodo Moeller*
13749 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13750 request to response.
13754 * Functions for OCSP responders. OCSP_request_onereq_count(),
13755 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13756 extract information from a certificate request. OCSP_response_create()
13757 creates a response and optionally adds a basic response structure.
13758 OCSP_basic_add1_status() adds a complete single response to a basic
13759 response and returns the OCSP_SINGLERESP structure just added (to allow
13760 extensions to be included for example). OCSP_basic_add1_cert() adds a
13761 certificate to a basic response and OCSP_basic_sign() signs a basic
13762 response with various flags. New helper functions ASN1_TIME_check()
13763 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13764 (converts ASN1_TIME to GeneralizedTime).
13768 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13769 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13770 structure from a certificate. X509_pubkey_digest() digests the public_key
13771 contents: this is used in various key identifiers.
13775 * Make sk_sort() tolerate a NULL argument.
13777 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13779 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13780 passed by the function are trusted implicitly. If any of them signed the
13781 response then it is assumed to be valid and is not verified.
13785 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13786 to data. This was previously part of the PKCS7 ASN1 code. This
13787 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13788 *Steve Henson, reported by Kenneth R. Robinette
13789 <support@securenetterm.com>*
13791 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13792 routines: without these tracing memory leaks is very painful.
13793 Fix leaks in PKCS12 and PKCS7 routines.
13797 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13798 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13799 effectively meant GeneralizedTime would never be used. Now it
13800 is initialised to -1 but X509_time_adj() now has to check the value
13801 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13802 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13803 *Steve Henson, reported by Kenneth R. Robinette
13804 <support@securenetterm.com>*
13806 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13807 result in a zero length in the ASN1_INTEGER structure which was
13808 not consistent with the structure when d2i_ASN1_INTEGER() was used
13809 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13810 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13811 where it did not print out a minus for negative ASN1_INTEGER.
13815 * Add summary printout to ocsp utility. The various functions which
13816 convert status values to strings have been renamed to:
13817 OCSP_response_status_str(), OCSP_cert_status_str() and
13818 OCSP_crl_reason_str() and are no longer static. New options
13819 to verify nonce values and to disable verification. OCSP response
13820 printout format cleaned up.
13824 * Add additional OCSP certificate checks. These are those specified
13825 in RFC2560. This consists of two separate checks: the CA of the
13826 certificate being checked must either be the OCSP signer certificate
13827 or the issuer of the OCSP signer certificate. In the latter case the
13828 OCSP signer certificate must contain the OCSP signing extended key
13829 usage. This check is performed by attempting to match the OCSP
13830 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13831 in the OCSP_CERTID structures of the response.
13835 * Initial OCSP certificate verification added to OCSP_basic_verify()
13836 and related routines. This uses the standard OpenSSL certificate
13837 verify routines to perform initial checks (just CA validity) and
13838 to obtain the certificate chain. Then additional checks will be
13839 performed on the chain. Currently the root CA is checked to see
13840 if it is explicitly trusted for OCSP signing. This is used to set
13841 a root CA as a global signing root: that is any certificate that
13842 chains to that CA is an acceptable OCSP signing certificate.
13846 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13847 extensions from a separate configuration file.
13848 As when reading extensions from the main configuration file,
13849 the '-extensions ...' option may be used for specifying the
13852 *Massimiliano Pala <madwolf@comune.modena.it>*
13854 * New OCSP utility. Allows OCSP requests to be generated or
13855 read. The request can be sent to a responder and the output
13856 parsed, outputted or printed in text form. Not complete yet:
13857 still needs to check the OCSP response validity.
13861 * New subcommands for 'openssl ca':
13862 `openssl ca -status <serial>` prints the status of the cert with
13863 the given serial number (according to the index file).
13864 `openssl ca -updatedb` updates the expiry status of certificates
13867 *Massimiliano Pala <madwolf@comune.modena.it>*
13869 * New '-newreq-nodes' command option to CA.pl. This is like
13870 '-newreq', but calls 'openssl req' with the '-nodes' option
13871 so that the resulting key is not encrypted.
13873 *Damien Miller <djm@mindrot.org>*
13875 * New configuration for the GNU Hurd.
13877 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13879 * Initial code to implement OCSP basic response verify. This
13880 is currently incomplete. Currently just finds the signer's
13881 certificate and verifies the signature on the response.
13885 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13886 value of OPENSSLDIR. This is available via the new '-d' option
13887 to 'openssl version', and is also included in 'openssl version -a'.
13891 * Allowing defining memory allocation callbacks that will be given
13892 file name and line number information in additional arguments
13893 (a `const char*` and an int). The basic functionality remains, as
13894 well as the original possibility to just replace malloc(),
13895 realloc() and free() by functions that do not know about these
13896 additional arguments. To register and find out the current
13897 settings for extended allocation functions, the following
13898 functions are provided:
13900 CRYPTO_set_mem_ex_functions
13901 CRYPTO_set_locked_mem_ex_functions
13902 CRYPTO_get_mem_ex_functions
13903 CRYPTO_get_locked_mem_ex_functions
13905 These work the same way as CRYPTO_set_mem_functions and friends.
13906 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13907 extended allocation function is enabled.
13908 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13909 a conventional allocation function is enabled.
13911 *Richard Levitte, Bodo Moeller*
13913 * Finish off removing the remaining LHASH function pointer casts.
13914 There should no longer be any prototype-casting required when using
13915 the LHASH abstraction, and any casts that remain are "bugs". See
13916 the callback types and macros at the head of lhash.h for details
13917 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13921 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13922 If /dev/[u]random devices are not available or do not return enough
13923 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13925 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13926 /etc/entropy will be queried once each in this sequence, querying stops
13927 when enough entropy was collected without querying more sockets.
13931 * Change the Unix RAND_poll() variant to be able to poll several
13932 random devices, as specified by DEVRANDOM, until a sufficient amount
13933 of data has been collected. We spend at most 10 ms on each file
13934 (select timeout) and read in non-blocking mode. DEVRANDOM now
13935 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13936 (previously it was just the string "/dev/urandom"), so on typical
13937 platforms the 10 ms delay will never occur.
13938 Also separate out the Unix variant to its own file, rand_unix.c.
13939 For VMS, there's a currently-empty rand_vms.c.
13943 * Move OCSP client related routines to ocsp_cl.c. These
13944 provide utility functions which an application needing
13945 to issue a request to an OCSP responder and analyse the
13946 response will typically need: as opposed to those which an
13947 OCSP responder itself would need which will be added later.
13949 OCSP_request_sign() signs an OCSP request with an API similar
13950 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13951 response. OCSP_response_get1_basic() extracts basic response
13952 from response. OCSP_resp_find_status(): finds and extracts status
13953 information from an OCSP_CERTID structure (which will be created
13954 when the request structure is built). These are built from lower
13955 level functions which work on OCSP_SINGLERESP structures but
13956 won't normally be used unless the application wishes to examine
13957 extensions in the OCSP response for example.
13959 Replace nonce routines with a pair of functions.
13960 OCSP_request_add1_nonce() adds a nonce value and optionally
13961 generates a random value. OCSP_check_nonce() checks the
13962 validity of the nonce in an OCSP response.
13966 * Change function OCSP_request_add() to OCSP_request_add0_id().
13967 This doesn't copy the supplied OCSP_CERTID and avoids the
13968 need to free up the newly created id. Change return type
13969 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13970 This can then be used to add extensions to the request.
13971 Deleted OCSP_request_new(), since most of its functionality
13972 is now in OCSP_REQUEST_new() (and the case insensitive name
13973 clash) apart from the ability to set the request name which
13974 will be added elsewhere.
13978 * Update OCSP API. Remove obsolete extensions argument from
13979 various functions. Extensions are now handled using the new
13980 OCSP extension code. New simple OCSP HTTP function which
13981 can be used to send requests and parse the response.
13985 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13986 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13987 uses the special reorder version of SET OF to sort the attributes
13988 and reorder them to match the encoded order. This resolves a long
13989 standing problem: a verify on a PKCS7 structure just after signing
13990 it used to fail because the attribute order did not match the
13991 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13992 it uses the received order. This is necessary to tolerate some broken
13993 software that does not order SET OF. This is handled by encoding
13994 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13995 to produce the required SET OF.
13999 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14000 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14001 files to get correct declarations of the ASN.1 item variables.
14005 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14006 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14007 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14008 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14009 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14010 ASN1_ITEM and no wrapper functions.
14014 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14015 replace the old function pointer based I/O routines. Change most of
14016 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14020 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14021 lines, recognize more "algorithms" that can be deselected, and make
14022 it complain about algorithm deselection that isn't recognised.
14026 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14027 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14028 to use new functions. Add NO_ASN1_OLD which can be set to remove
14029 some old style ASN1 functions: this can be used to determine if old
14030 code will still work when these eventually go away.
14034 * New extension functions for OCSP structures, these follow the
14035 same conventions as certificates and CRLs.
14039 * New function X509V3_add1_i2d(). This automatically encodes and
14040 adds an extension. Its behaviour can be customised with various
14041 flags to append, replace or delete. Various wrappers added for
14042 certificates and CRLs.
14046 * Fix to avoid calling the underlying ASN1 print routine when
14047 an extension cannot be parsed. Correct a typo in the
14048 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14052 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14053 entries for variables.
14057 * Add functionality to `apps/openssl.c` for detecting locking
14058 problems: As the program is single-threaded, all we have
14059 to do is register a locking callback using an array for
14060 storing which locks are currently held by the program.
14064 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14065 SSL_get_ex_data_X509_STORE_idx(), which is used in
14066 ssl_verify_cert_chain() and thus can be called at any time
14067 during TLS/SSL handshakes so that thread-safety is essential.
14068 Unfortunately, the ex_data design is not at all suited
14069 for multi-threaded use, so it probably should be abolished.
14073 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14075 *Broadcom, tweaked and integrated by Geoff Thorpe*
14077 * Move common extension printing code to new function
14078 X509V3_print_extensions(). Reorganise OCSP print routines and
14079 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14083 * New function X509_signature_print() to remove duplication in some
14088 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14089 set (this was treated exactly the same as SET OF previously). This
14090 is used to reorder the STACK representing the structure to match the
14091 encoding. This will be used to get round a problem where a PKCS7
14092 structure which was signed could not be verified because the STACK
14093 order did not reflect the encoded order.
14097 * Reimplement the OCSP ASN1 module using the new code.
14101 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14102 for its ASN1 operations. The old style function pointers still exist
14103 for now but they will eventually go away.
14107 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14108 completely replaces the old ASN1 functionality with a table driven
14109 encoder and decoder which interprets an ASN1_ITEM structure describing
14110 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14111 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14112 has also been converted to the new form.
14116 * Change BN_mod_exp_recp so that negative moduli are tolerated
14117 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14118 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14119 for negative moduli.
14123 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14124 of not touching the result's sign bit.
14128 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14133 * Changed the LHASH code to use prototypes for callbacks, and created
14134 macros to declare and implement thin (optionally static) functions
14135 that provide type-safety and avoid function pointer casting for the
14136 type-specific callbacks.
14140 * Added Kerberos Cipher Suites to be used with TLS, as written in
14142 *Veers Staats <staatsvr@asc.hpc.mil>,
14143 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14145 * Reformat the FAQ so the different questions and answers can be divided
14146 in sections depending on the subject.
14150 * Have the zlib compression code load ZLIB.DLL dynamically under
14155 * New function BN_mod_sqrt for computing square roots modulo a prime
14156 (using the probabilistic Tonelli-Shanks algorithm unless
14157 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14158 be handled deterministically).
14160 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14162 * Make BN_mod_inverse faster by explicitly handling small quotients
14163 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14164 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14168 * New function BN_kronecker.
14172 * Fix BN_gcd so that it works on negative inputs; the result is
14173 positive unless both parameters are zero.
14174 Previously something reasonably close to an infinite loop was
14175 possible because numbers could be growing instead of shrinking
14176 in the implementation of Euclid's algorithm.
14180 * Fix BN_is_word() and BN_is_one() macros to take into account the
14181 sign of the number in question.
14183 Fix BN_is_word(a,w) to work correctly for w == 0.
14185 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14186 because its test if the absolute value of 'a' equals 'w'.
14187 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14188 it exists mostly for use in the implementations of BN_is_zero(),
14189 BN_is_one(), and BN_is_word().
14193 * New function BN_swap.
14197 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14198 the exponentiation functions are more likely to produce reasonable
14199 results on negative inputs.
14203 * Change BN_mod_mul so that the result is always non-negative.
14204 Previously, it could be negative if one of the factors was negative;
14205 I don't think anyone really wanted that behaviour.
14209 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14210 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14211 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14212 and add new functions:
14221 BN_mod_lshift1_quick
14223 BN_mod_lshift_quick
14225 These functions always generate non-negative results.
14227 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14228 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14230 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14231 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14232 be reduced modulo `m`.
14234 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14237 The following entry accidentally appeared in the CHANGES file
14238 distributed with OpenSSL 0.9.7. The modifications described in
14239 it do *not* apply to OpenSSL 0.9.7.
14241 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14242 was actually never needed) and in BN_mul(). The removal in BN_mul()
14243 required a small change in bn_mul_part_recursive() and the addition
14244 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14245 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14246 bn_sub_words() and bn_add_words() except they take arrays with
14252 * In 'openssl passwd', verify passwords read from the terminal
14253 unless the '-salt' option is used (which usually means that
14254 verification would just waste user's time since the resulting
14255 hash is going to be compared with some given password hash)
14256 or the new '-noverify' option is used.
14258 This is an incompatible change, but it does not affect
14259 non-interactive use of 'openssl passwd' (passwords on the command
14260 line, '-stdin' option, '-in ...' option) and thus should not
14261 cause any problems.
14265 * Remove all references to RSAref, since there's no more need for it.
14269 * Make DSO load along a path given through an environment variable
14270 (SHLIB_PATH) with shl_load().
14274 * Constify the ENGINE code as a result of BIGNUM constification.
14275 Also constify the RSA code and most things related to it. In a
14276 few places, most notable in the depth of the ASN.1 code, ugly
14277 casts back to non-const were required (to be solved at a later
14282 * Make it so the openssl application has all engines loaded by default.
14286 * Constify the BIGNUM routines a little more.
14290 * Add the following functions:
14292 ENGINE_load_cswift()
14294 ENGINE_load_atalla()
14295 ENGINE_load_nuron()
14296 ENGINE_load_builtin_engines()
14298 That way, an application can itself choose if external engines that
14299 are built-in in OpenSSL shall ever be used or not. The benefit is
14300 that applications won't have to be linked with libdl or other dso
14301 libraries unless it's really needed.
14303 Changed 'openssl engine' to load all engines on demand.
14304 Changed the engine header files to avoid the duplication of some
14305 declarations (they differed!).
14309 * 'openssl engine' can now list capabilities.
14313 * Better error reporting in 'openssl engine'.
14317 * Never call load_dh_param(NULL) in s_server.
14321 * Add engine application. It can currently list engines by name and
14322 identity, and test if they are actually available.
14326 * Improve RPM specification file by forcing symbolic linking and making
14327 sure the installed documentation is also owned by root.root.
14329 *Damien Miller <djm@mindrot.org>*
14331 * Give the OpenSSL applications more possibilities to make use of
14332 keys (public as well as private) handled by engines.
14336 * Add OCSP code that comes from CertCo.
14340 * Add VMS support for the Rijndael code.
14344 * Added untested support for Nuron crypto accelerator.
14348 * Add support for external cryptographic devices. This code was
14349 previously distributed separately as the "engine" branch.
14351 *Geoff Thorpe, Richard Levitte*
14353 * Rework the filename-translation in the DSO code. It is now possible to
14354 have far greater control over how a "name" is turned into a filename
14355 depending on the operating environment and any oddities about the
14356 different shared library filenames on each system.
14360 * Support threads on FreeBSD-elf in Configure.
14364 * Fix for SHA1 assembly problem with MASM: it produces
14365 warnings about corrupt line number information when assembling
14366 with debugging information. This is caused by the overlapping
14369 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14372 NCONF_get_number() has no error checking at all. As a replacement,
14373 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14374 promoted strongly. The old NCONF_get_number is kept around for
14375 binary backward compatibility.
14376 Make it possible for methods to load from something other than a BIO,
14377 by providing a function pointer that is given a name instead of a BIO.
14378 For example, this could be used to load configuration data from an
14383 * Fix for non blocking accept BIOs. Added new I/O special reason
14384 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14385 with non blocking I/O was not possible because no retry code was
14386 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14391 * Added the beginnings of Rijndael support.
14395 * Fix for bug in DirectoryString mask setting. Add support for
14396 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14397 to allow certificate printing to more controllable, additional
14398 'certopt' option to 'x509' to allow new printing options to be
14403 * Clean old EAY MD5 hack from e_os.h.
14407 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14409 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14410 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14412 *Joe Orton, Steve Henson*
14414 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14416 * Fix additional bug revealed by the NISCC test suite:
14418 Stop bug triggering large recursion when presented with
14419 certain ASN.1 tags ([CVE-2003-0851])
14423 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14425 * Fix various bugs revealed by running the NISCC test suite:
14427 Stop out of bounds reads in the ASN1 code when presented with
14428 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14430 If verify callback ignores invalid public key errors don't try to check
14431 certificate signature with the NULL public key.
14435 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14436 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14441 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14442 extra data after the compression methods not only for TLS 1.0
14443 but also for SSL 3.0 (as required by the specification).
14445 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14447 * Change X509_certificate_type() to mark the key as exported/exportable
14448 when it's 512 *bits* long, not 512 bytes.
14452 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14454 * Countermeasure against the Klima-Pokorny-Rosa extension of
14455 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14456 a protocol version number mismatch like a decryption error
14457 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14461 * Turn on RSA blinding by default in the default implementation
14462 to avoid a timing attack. Applications that don't want it can call
14463 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14464 They would be ill-advised to do so in most cases.
14466 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14468 * Change RSA blinding code so that it works when the PRNG is not
14469 seeded (in this case, the secret RSA exponent is abused as
14470 an unpredictable seed -- if it is not unpredictable, there
14471 is no point in blinding anyway). Make RSA blinding thread-safe
14472 by remembering the creator's thread ID in rsa->blinding and
14473 having all other threads use local one-time blinding factors
14474 (this requires more computation than sharing rsa->blinding, but
14475 avoids excessive locking; and if an RSA object is not shared
14476 between threads, blinding will still be very fast).
14480 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14482 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14483 via timing by performing a MAC computation even if incorrect
14484 block cipher padding has been found. This is a countermeasure
14485 against active attacks where the attacker has to distinguish
14486 between bad padding and a MAC verification error. ([CVE-2003-0078])
14488 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14489 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14490 Martin Vuagnoux (EPFL, Ilion)*
14492 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14494 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14495 memory from its contents. This is done with a counter that will
14496 place alternating values in each byte. This can be used to solve
14497 two issues: 1) the removal of calls to memset() by highly optimizing
14498 compilers, and 2) cleansing with other values than 0, since those can
14499 be read through on certain media, for example a swap space on disk.
14503 * Bugfix: client side session caching did not work with external caching,
14504 because the session->cipher setting was not restored when reloading
14505 from the external cache. This problem was masked, when
14506 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14507 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14511 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14512 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14514 *Zeev Lieber <zeev-l@yahoo.com>*
14516 * Undo an undocumented change introduced in 0.9.6e which caused
14517 repeated calls to OpenSSL_add_all_ciphers() and
14518 OpenSSL_add_all_digests() to be ignored, even after calling
14523 * Change the default configuration reader to deal with last line not
14524 being properly terminated.
14528 * Change X509_NAME_cmp() so it applies the special rules on handling
14529 DN values that are of type PrintableString, as well as RDNs of type
14530 emailAddress where the value has the type ia5String.
14532 *stefank@valicert.com via Richard Levitte*
14534 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14535 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14536 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14537 the bitwise-OR of the two for use by the majority of applications
14538 wanting this behaviour, and update the docs. The documented
14539 behaviour and actual behaviour were inconsistent and had been
14540 changing anyway, so this is more a bug-fix than a behavioural
14543 *Geoff Thorpe, diagnosed by Nadav Har'El*
14545 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14546 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14550 * Fix initialization code race conditions in
14551 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14552 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14553 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14554 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14555 ssl2_get_cipher_by_char(),
14556 ssl3_get_cipher_by_char().
14558 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14560 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14561 the cached sessions are flushed, as the remove_cb() might use ex_data
14562 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14563 (see [openssl.org #212]).
14565 *Geoff Thorpe, Lutz Jaenicke*
14567 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14568 length, instead of the encoding length to d2i_ASN1_OBJECT.
14572 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14574 * [In 0.9.6g-engine release:]
14575 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14577 *Lynn Gazis <lgazis@rainbow.com>*
14579 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14581 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14582 and get fix the header length calculation.
14583 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14584 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14586 * Use proper error handling instead of 'assertions' in buffer
14587 overflow checks added in 0.9.6e. This prevents DoS (the
14588 assertions could call abort()).
14590 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14592 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14594 * Add various sanity checks to asn1_get_length() to reject
14595 the ASN1 length bytes if they exceed sizeof(long), will appear
14596 negative or the content length exceeds the length of the
14599 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14601 * Fix cipher selection routines: ciphers without encryption had no flags
14602 for the cipher strength set and where therefore not handled correctly
14603 by the selection routines (PR #130).
14607 * Fix EVP_dsa_sha macro.
14612 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14613 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14614 that was added in OpenSSL 0.9.6d.
14616 As the countermeasure turned out to be incompatible with some
14617 broken SSL implementations, the new option is part of SSL_OP_ALL.
14618 SSL_OP_ALL is usually employed when compatibility with weird SSL
14619 implementations is desired (e.g. '-bugs' option to 's_client' and
14620 's_server'), so the new option is automatically set in many
14625 * Changes in security patch:
14627 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14628 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14629 Air Force Materiel Command, USAF, under agreement number
14632 * Add various sanity checks to asn1_get_length() to reject
14633 the ASN1 length bytes if they exceed sizeof(long), will appear
14634 negative or the content length exceeds the length of the
14635 supplied buffer. ([CVE-2002-0659])
14637 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14639 * Assertions for various potential buffer overflows, not known to
14640 happen in practice.
14642 *Ben Laurie (CHATS)*
14644 * Various temporary buffers to hold ASCII versions of integers were
14645 too small for 64 bit platforms. ([CVE-2002-0655])
14646 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14648 * Remote buffer overflow in SSL3 protocol - an attacker could
14649 supply an oversized session ID to a client. ([CVE-2002-0656])
14651 *Ben Laurie (CHATS)*
14653 * Remote buffer overflow in SSL2 protocol - an attacker could
14654 supply an oversized client master key. ([CVE-2002-0656])
14656 *Ben Laurie (CHATS)*
14658 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14660 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14661 encoded as NULL) with id-dsa-with-sha1.
14663 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14665 * Check various `X509_...()` return values in `apps/req.c`.
14667 *Nils Larsch <nla@trustcenter.de>*
14669 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14670 an end-of-file condition would erroneously be flagged, when the CRLF
14671 was just at the end of a processed block. The bug was discovered when
14672 processing data through a buffering memory BIO handing the data to a
14673 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14674 <ptsekov@syntrex.com> and Nedelcho Stanev.
14678 * Implement a countermeasure against a vulnerability recently found
14679 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14680 before application data chunks to avoid the use of known IVs
14681 with data potentially chosen by the attacker.
14685 * Fix length checks in ssl3_get_client_hello().
14689 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14690 to prevent ssl3_read_internal() from incorrectly assuming that
14691 ssl3_read_bytes() found application data while handshake
14692 processing was enabled when in fact s->s3->in_read_app_data was
14693 merely automatically cleared during the initial handshake.
14695 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14697 * Fix object definitions for Private and Enterprise: they were not
14698 recognized in their shortname (=lowercase) representation. Extend
14699 obj_dat.pl to issue an error when using undefined keywords instead
14700 of silently ignoring the problem (Svenning Sorensen
14701 <sss@sss.dnsalias.net>).
14705 * Fix DH_generate_parameters() so that it works for 'non-standard'
14706 generators, i.e. generators other than 2 and 5. (Previously, the
14707 code did not properly initialise the 'add' and 'rem' values to
14708 BN_generate_prime().)
14710 In the new general case, we do not insist that 'generator' is
14711 actually a primitive root: This requirement is rather pointless;
14712 a generator of the order-q subgroup is just as good, if not
14717 * Map new X509 verification errors to alerts. Discovered and submitted by
14718 Tom Wu <tom@arcot.com>.
14722 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14723 returning non-zero before the data has been completely received
14724 when using non-blocking I/O.
14726 *Bodo Moeller; problem pointed out by John Hughes*
14728 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14730 *Ben Laurie, Lutz Jaenicke*
14732 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14733 Yoram Zahavi <YoramZ@gilian.com>).
14737 * Add information about CygWin 1.3 and on, and preserve proper
14738 configuration for the versions before that.
14740 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14742 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14743 check whether we deal with a copy of a session and do not delete from
14744 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14745 <izhar@checkpoint.com>.
14749 * Do not store session data into the internal session cache, if it
14750 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14751 flag is set). Proposed by Aslam <aslam@funk.com>.
14755 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14760 * [In 0.9.6d-engine release:]
14761 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14763 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14765 * Add the configuration target linux-s390x.
14767 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14769 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14770 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14771 variable as an indication that a ClientHello message has been
14772 received. As the flag value will be lost between multiple
14773 invocations of ssl3_accept when using non-blocking I/O, the
14774 function may not be aware that a handshake has actually taken
14775 place, thus preventing a new session from being added to the
14778 To avoid this problem, we now set s->new_session to 2 instead of
14779 using a local variable.
14781 *Lutz Jaenicke, Bodo Moeller*
14783 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14784 if the SSL_R_LENGTH_MISMATCH error is detected.
14786 *Geoff Thorpe, Bodo Moeller*
14788 * New 'shared_ldflag' column in Configure platform table.
14792 * Fix EVP_CIPHER_mode macro.
14794 *"Dan S. Camper" <dan@bti.net>*
14796 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14797 type, we must throw them away by setting rr->length to 0.
14799 *D P Chang <dpc@qualys.com>*
14801 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14803 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14804 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14805 worked incorrectly for those cases where range = `10..._2` and
14806 `3*range` is two bits longer than range.)
14810 * Only add signing time to PKCS7 structures if it is not already
14815 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14816 OBJ_ld_ce should be OBJ_id_ce.
14817 Also some ip-pda OIDs in crypto/objects/objects.txt were
14818 incorrect (cf. RFC 3039).
14820 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14822 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14823 returns early because it has nothing to do.
14825 *Andy Schneider <andy.schneider@bjss.co.uk>*
14827 * [In 0.9.6c-engine release:]
14828 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14830 *Andy Schneider <andy.schneider@bjss.co.uk>*
14832 * [In 0.9.6c-engine release:]
14833 Add support for Cryptographic Appliance's keyserver technology.
14834 (Use engine 'keyclient')
14836 *Cryptographic Appliances and Geoff Thorpe*
14838 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14839 is called via tools/c89.sh because arguments have to be
14840 rearranged (all '-L' options must appear before the first object
14843 *Richard Shapiro <rshapiro@abinitio.com>*
14845 * [In 0.9.6c-engine release:]
14846 Add support for Broadcom crypto accelerator cards, backported
14849 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14851 * [In 0.9.6c-engine release:]
14852 Add support for SureWare crypto accelerator cards from
14853 Baltimore Technologies. (Use engine 'sureware')
14855 *Baltimore Technologies and Mark Cox*
14857 * [In 0.9.6c-engine release:]
14858 Add support for crypto accelerator cards from Accelerated
14859 Encryption Processing, www.aep.ie. (Use engine 'aep')
14861 *AEP Inc. and Mark Cox*
14863 * Add a configuration entry for gcc on UnixWare.
14865 *Gary Benson <gbenson@redhat.com>*
14867 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14868 messages are stored in a single piece (fixed-length part and
14869 variable-length part combined) and fix various bugs found on the way.
14873 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14874 instead. BIO_gethostbyname() does not know what timeouts are
14875 appropriate, so entries would stay in cache even when they have
14877 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14879 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14880 faced with a pathologically small ClientHello fragment that does
14881 not contain client_version: Instead of aborting with an error,
14882 simply choose the highest available protocol version (i.e.,
14883 TLS 1.0 unless it is disabled). In practice, ClientHello
14884 messages are never sent like this, but this change gives us
14885 strictly correct behaviour at least for TLS.
14889 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14890 never resets s->method to s->ctx->method when called from within
14891 one of the SSL handshake functions.
14893 *Bodo Moeller; problem pointed out by Niko Baric*
14895 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14896 (sent using the client's version number) if client_version is
14897 smaller than the protocol version in use. Also change
14898 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14899 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14900 the client will at least see that alert.
14904 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14909 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14910 client receives HelloRequest while in a handshake.
14912 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14914 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14915 should end in 'break', not 'goto end' which circumvents various
14916 cleanups done in state SSL_ST_OK. But session related stuff
14917 must be disabled for SSL_ST_OK in the case that we just sent a
14920 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14921 before just sending a HelloRequest.
14923 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14925 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14926 reveal whether illegal block cipher padding was found or a MAC
14927 verification error occurred. (Neither SSLerr() codes nor alerts
14928 are directly visible to potential attackers, but the information
14929 may leak via logfiles.)
14931 Similar changes are not required for the SSL 2.0 implementation
14932 because the number of padding bytes is sent in clear for SSL 2.0,
14933 and the extra bytes are just ignored. However ssl/s2_pkt.c
14934 failed to verify that the purported number of padding bytes is in
14939 * Add OpenUNIX-8 support including shared libraries
14940 (Boyd Lynn Gerber <gerberb@zenez.com>).
14944 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14945 'wristwatch attack' using huge encoding parameters (cf.
14946 James H. Manger's CRYPTO 2001 paper). Note that the
14947 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14948 encoding parameters and hence was not vulnerable.
14952 * BN_sqr() bug fix.
14954 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14956 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14957 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14958 followed by modular reduction.
14960 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14962 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14963 equivalent based on BN_pseudo_rand() instead of BN_rand().
14967 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14968 This function was broken, as the check for a new client hello message
14969 to handle SGC did not allow these large messages.
14970 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14974 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14978 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14979 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14983 * Rework the configuration and shared library support for Tru64 Unix.
14984 The configuration part makes use of modern compiler features and
14985 still retains old compiler behavior for those that run older versions
14986 of the OS. The shared library support part includes a variant that
14987 uses the RPATH feature, and is available through the special
14988 configuration target "alpha-cc-rpath", which will never be selected
14991 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14993 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14994 with the same message size as in ssl3_get_certificate_request().
14995 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14996 messages might inadvertently be reject as too long.
14998 *Petr Lampa <lampa@fee.vutbr.cz>*
15000 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15004 * Modified SSL library such that the verify_callback that has been set
15005 specifically for an SSL object with SSL_set_verify() is actually being
15006 used. Before the change, a verify_callback set with this function was
15007 ignored and the verify_callback() set in the SSL_CTX at the time of
15008 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15009 to allow the necessary settings.
15013 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15014 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15015 done automatically (in contradiction to the requirements of the C
15016 standard). This made problems when used from OpenSSH.
15020 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15021 dh->length and always used
15023 BN_rand_range(priv_key, dh->p).
15025 BN_rand_range() is not necessary for Diffie-Hellman, and this
15026 specific range makes Diffie-Hellman unnecessarily inefficient if
15027 dh->length (recommended exponent length) is much smaller than the
15028 length of dh->p. We could use BN_rand_range() if the order of
15029 the subgroup was stored in the DH structure, but we only have
15034 BN_rand(priv_key, l, ...)
15036 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15043 RSA_eay_public_encrypt
15044 RSA_eay_private_decrypt
15045 RSA_eay_private_encrypt (signing)
15046 RSA_eay_public_decrypt (signature verification)
15048 (default implementations for RSA_public_encrypt,
15049 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15050 always reject numbers >= n.
15054 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15055 to synchronize access to 'locking_thread'. This is necessary on
15056 systems where access to 'locking_thread' (an 'unsigned long'
15057 variable) is not atomic.
15061 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15062 *before* setting the 'crypto_lock_rand' flag. The previous code had
15063 a race condition if 0 is a valid thread ID.
15065 *Travis Vitek <vitek@roguewave.com>*
15067 * Add support for shared libraries under Irix.
15069 *Albert Chin-A-Young <china@thewrittenword.com>*
15071 * Add configuration option to build on Linux on both big-endian and
15072 little-endian MIPS.
15074 *Ralf Baechle <ralf@uni-koblenz.de>*
15076 * Add the possibility to create shared libraries on HP-UX.
15080 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15082 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15083 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15084 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15085 PRNG state recovery was possible based on the output of
15086 one PRNG request appropriately sized to gain knowledge on
15087 'md' followed by enough consecutive 1-byte PRNG requests
15088 to traverse all of 'state'.
15090 1. When updating 'md_local' (the current thread's copy of 'md')
15091 during PRNG output generation, hash all of the previous
15092 'md_local' value, not just the half used for PRNG output.
15094 2. Make the number of bytes from 'state' included into the hash
15095 independent from the number of PRNG bytes requested.
15097 The first measure alone would be sufficient to avoid
15098 Markku-Juhani's attack. (Actually it had never occurred
15099 to me that the half of 'md_local' used for chaining was the
15100 half from which PRNG output bytes were taken -- I had always
15101 assumed that the secret half would be used.) The second
15102 measure makes sure that additional data from 'state' is never
15103 mixed into 'md_local' in small portions; this heuristically
15104 further strengthens the PRNG.
15108 * Fix crypto/bn/asm/mips3.s.
15112 * When only the key is given to "enc", the IV is undefined. Print out
15113 an error message in this case.
15117 * Handle special case when X509_NAME is empty in X509 printing routines.
15121 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15122 positive and less than q.
15126 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15127 used: it isn't thread safe and the add_lock_callback should handle
15130 *Paul Rose <Paul.Rose@bridge.com>*
15132 * Verify that incoming data obeys the block size in
15133 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15139 *Ulf Möller, Bodo Möller*
15141 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15142 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15143 when fixing the server behaviour for backwards-compatible 'client
15144 hello' messages. (Note that the attack is impractical against
15145 SSL 3.0 and TLS 1.0 anyway because length and version checking
15146 means that the probability of guessing a valid ciphertext is
15147 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15150 Before 0.9.5, the countermeasure (hide the error by generating a
15151 random 'decryption result') did not work properly because
15152 ERR_clear_error() was missing, meaning that SSL_get_error() would
15153 detect the supposedly ignored error.
15155 Both problems are now fixed.
15159 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15160 (previously it was 1024).
15164 * Fix for compatibility mode trust settings: ignore trust settings
15165 unless some valid trust or reject settings are present.
15169 * Fix for blowfish EVP: its a variable length cipher.
15173 * Fix various bugs related to DSA S/MIME verification. Handle missing
15174 parameters in DSA public key structures and return an error in the
15175 DSA routines if parameters are absent.
15179 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15180 in the current directory if neither $RANDFILE nor $HOME was set.
15181 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15182 caused some confusion to Windows users who haven't defined $HOME.
15183 Thus RAND_file_name() is changed again: e_os.h can define a
15184 DEFAULT_HOME, which will be used if $HOME is not set.
15185 For Windows, we use "C:"; on other platforms, we still require
15186 environment variables.
15188 * Move 'if (!initialized) RAND_poll()' into regions protected by
15189 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15190 having multiple threads call RAND_poll() concurrently.
15194 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15195 combination of a flag and a thread ID variable.
15196 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15197 flag), *other* threads can enter ssleay_add_bytes without obeying
15198 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15199 that they do not hold after the first thread unsets add_do_not_lock).
15203 * Change bctest again: '-x' expressions are not available in all
15204 versions of 'test'.
15208 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15210 * Fix a couple of memory leaks in PKCS7_dataDecode()
15212 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15214 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15215 the default extension for executables, if any. Also, make the perl
15216 scripts that use symlink() to test if it really exists and use "cp"
15217 if it doesn't. All this made OpenSSL compilable and installable in
15222 * Fix for asn1_GetSequence() for indefinite length constructed data.
15223 If SEQUENCE is length is indefinite just set c->slen to the total
15224 amount of data available.
15226 *Steve Henson, reported by shige@FreeBSD.org*
15228 *This change does not apply to 0.9.7.*
15230 * Change bctest to avoid here-documents inside command substitution
15231 (workaround for FreeBSD /bin/sh bug).
15232 For compatibility with Ultrix, avoid shell functions (introduced
15233 in the bctest version that searches along $PATH).
15237 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15238 with des_encrypt() defined on some operating systems, like Solaris
15243 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15244 On the Importance of Eliminating Errors in Cryptographic
15245 Computations, J. Cryptology 14 (2001) 2, 101-119,
15246 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15250 * MIPS assembler BIGNUM division bug fix.
15254 * Disabled incorrect Alpha assembler code.
15258 * Fix PKCS#7 decode routines so they correctly update the length
15259 after reading an EOC for the EXPLICIT tag.
15263 *This change does not apply to 0.9.7.*
15265 * Fix bug in PKCS#12 key generation routines. This was triggered
15266 if a 3DES key was generated with a 0 initial byte. Include
15267 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15268 (but broken) behaviour.
15272 * Enhance bctest to search for a working bc along $PATH and print
15275 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15277 * Fix memory leaks in err.c: free err_data string if necessary;
15278 don't write to the wrong index in ERR_set_error_data.
15282 * Implement ssl23_peek (analogous to ssl23_read), which previously
15287 * Replace rdtsc with `_emit` statements for VC++ version 5.
15289 *Jeremy Cooper <jeremy@baymoo.org>*
15291 * Make it possible to reuse SSLv2 sessions.
15295 * In copy_email() check for >= 0 as a return value for
15296 X509_NAME_get_index_by_NID() since 0 is a valid index.
15298 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15300 * Avoid coredump with unsupported or invalid public keys by checking if
15301 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15302 PKCS7_verify() fails with non detached data.
15306 * Don't use getenv in library functions when run as setuid/setgid.
15307 New function OPENSSL_issetugid().
15311 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15312 due to incorrect handling of multi-threading:
15314 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15316 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15318 3. Count how many times MemCheck_off() has been called so that
15319 nested use can be treated correctly. This also avoids
15320 inband-signalling in the previous code (which relied on the
15321 assumption that thread ID 0 is impossible).
15325 * Add "-rand" option also to s_client and s_server.
15329 * Fix CPU detection on Irix 6.x.
15330 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15331 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15333 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15338 *This change does not apply to 0.9.7.*
15340 * Use the cached encoding of an X509_NAME structure rather than
15341 copying it. This is apparently the reason for the libsafe "errors"
15342 but the code is actually correct.
15346 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15347 Bleichenbacher's DSA attack.
15348 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15349 to be set and top=0 forces the highest bit to be set; top=-1 is new
15350 and leaves the highest bit random.
15352 *Ulf Moeller, Bodo Moeller*
15354 * In the `NCONF_...`-based implementations for `CONF_...` queries
15355 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15356 a temporary CONF structure with the data component set to NULL
15357 (which gives segmentation faults in lh_retrieve).
15358 Instead, use NULL for the CONF pointer in CONF_get_string and
15359 CONF_get_number (which may use environment variables) and directly
15360 return NULL from CONF_get_section.
15364 * Fix potential buffer overrun for EBCDIC.
15368 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15369 keyUsage if basicConstraints absent for a CA.
15373 * Make SMIME_write_PKCS7() write mail header values with a format that
15374 is more generally accepted (no spaces before the semicolon), since
15375 some programs can't parse those values properly otherwise. Also make
15376 sure BIO's that break lines after each write do not create invalid
15381 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15382 macros previously used would not encode an empty SEQUENCE OF
15383 and break the signature.
15387 *This change does not apply to 0.9.7.*
15389 * Zero the premaster secret after deriving the master secret in
15394 * Add some EVP_add_digest_alias registrations (as found in
15395 OpenSSL_add_all_digests()) to SSL_library_init()
15396 aka OpenSSL_add_ssl_algorithms(). This provides improved
15397 compatibility with peers using X.509 certificates
15398 with unconventional AlgorithmIdentifier OIDs.
15402 * Fix for Irix with NO_ASM.
15404 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15406 * ./config script fixes.
15408 *Ulf Moeller, Richard Levitte*
15410 * Fix 'openssl passwd -1'.
15414 * Change PKCS12_key_gen_asc() so it can cope with non null
15415 terminated strings whose length is passed in the passlen
15416 parameter, for example from PEM callbacks. This was done
15417 by adding an extra length parameter to asc2uni().
15419 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15421 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15422 call failed, free the DSA structure.
15426 * Fix to uni2asc() to cope with zero length Unicode strings.
15427 These are present in some PKCS#12 files.
15431 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15432 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15433 when writing a 32767 byte record.
15435 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15437 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15438 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15440 (RSA objects have a reference count access to which is protected
15441 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15442 so they are meant to be shared between threads.)
15443 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15444 "Reddie, Steven" <Steven.Reddie@ca.com>*
15446 * Fix a deadlock in CRYPTO_mem_leaks().
15450 * Use better test patterns in bntest.
15454 * rand_win.c fix for Borland C.
15458 * BN_rshift bugfix for n == 0.
15462 * Add a 'bctest' script that checks for some known 'bc' bugs
15463 so that 'make test' does not abort just because 'bc' is broken.
15467 * Store verify_result within SSL_SESSION also for client side to
15468 avoid potential security hole. (Re-used sessions on the client side
15469 always resulted in verify_result==X509_V_OK, not using the original
15470 result of the server certificate verification.)
15474 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15475 SSL3_RT_APPLICATION_DATA, return 0.
15476 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15481 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15482 releases, have been re-implemented by renaming the previous
15483 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15484 and ssl3_read_internal, respectively, and adding 'peek' parameters
15485 to them. The new ssl[23]_{read,peek} functions are calls to
15486 ssl[23]_read_internal with the 'peek' flag set appropriately.
15487 A 'peek' parameter has also been added to ssl3_read_bytes, which
15488 does the actual work for ssl3_read_internal.
15492 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15493 the method-specific "init()" handler. Also clean up ex_data after
15494 calling the method-specific "finish()" handler. Previously, this was
15495 happening the other way round.
15499 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15500 The previous value, 12, was not always sufficient for BN_mod_exp().
15504 * Make sure that shared libraries get the internal name engine with
15505 the full version number and not just 0. This should mark the
15506 shared libraries as not backward compatible. Of course, this should
15507 be changed again when we can guarantee backward binary compatibility.
15511 * Fix typo in get_cert_by_subject() in by_dir.c
15513 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15515 * Rework the system to generate shared libraries:
15517 - Make note of the expected extension for the shared libraries and
15518 if there is a need for symbolic links from for example libcrypto.so.0
15519 to libcrypto.so.0.9.7. There is extended info in Configure for
15522 - Make as few rebuilds of the shared libraries as possible.
15524 - Still avoid linking the OpenSSL programs with the shared libraries.
15526 - When installing, install the shared libraries separately from the
15531 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15533 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15534 and not in SSL_clear because the latter is also used by the
15535 accept/connect functions; previously, the settings made by
15536 SSL_set_read_ahead would be lost during the handshake.
15538 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15540 * Correct util/mkdef.pl to be selective about disabled algorithms.
15541 Previously, it would create entries for disabled algorithms no
15546 * Added several new manual pages for SSL_* function.
15550 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15552 * In ssl23_get_client_hello, generate an error message when faced
15553 with an initial SSL 3.0/TLS record that is too small to contain the
15554 first two bytes of the ClientHello message, i.e. client_version.
15555 (Note that this is a pathologic case that probably has never happened
15556 in real life.) The previous approach was to use the version number
15557 from the record header as a substitute; but our protocol choice
15558 should not depend on that one because it is not authenticated
15559 by the Finished messages.
15563 * More robust randomness gathering functions for Windows.
15565 *Jeffrey Altman <jaltman@columbia.edu>*
15567 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15568 not set then we don't setup the error code for issuer check errors
15569 to avoid possibly overwriting other errors which the callback does
15570 handle. If an application does set the flag then we assume it knows
15571 what it is doing and can handle the new informational codes
15576 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15577 a general "ANY" type, as such it should be able to decode anything
15578 including tagged types. However it didn't check the class so it would
15579 wrongly interpret tagged types in the same way as their universal
15580 counterpart and unknown types were just rejected. Changed so that the
15581 tagged and unknown types are handled in the same way as a SEQUENCE:
15582 that is the encoding is stored intact. There is also a new type
15583 "V_ASN1_OTHER" which is used when the class is not universal, in this
15584 case we have no idea what the actual type is so we just lump them all
15589 * On VMS, stdout may very well lead to a file that is written to
15590 in a record-oriented fashion. That means that every write() will
15591 write a separate record, which will be read separately by the
15592 programs trying to read from it. This can be very confusing.
15594 The solution is to put a BIO filter in the way that will buffer
15595 text until a linefeed is reached, and then write everything a
15596 line at a time, so every record written will be an actual line,
15597 not chunks of lines and not (usually doesn't happen, but I've
15598 seen it once) several lines in one record. BIO_f_linebuffer() is
15601 Currently, it's a VMS-only method, because that's where it has
15602 been tested well enough.
15606 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15607 it can return incorrect results.
15608 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15609 but it was in 0.9.6-beta[12].)
15613 * Disable the check for content being present when verifying detached
15614 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15615 include zero length content when signing messages.
15619 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15620 BIO_ctrl (for BIO pairs).
15624 * Add DSO method for VMS.
15628 * Bug fix: Montgomery multiplication could produce results with the
15633 * Add RPM specification openssl.spec and modify it to build three
15634 packages. The default package contains applications, application
15635 documentation and run-time libraries. The devel package contains
15636 include files, static libraries and function documentation. The
15637 doc package contains the contents of the doc directory. The original
15638 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15642 * Add a large number of documentation files for many SSL routines.
15644 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15646 * Add a configuration entry for Sony News 4.
15648 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15650 * Don't set the two most significant bits to one when generating a
15651 random number < q in the DSA library.
15655 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15656 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15657 the underlying transport is blocking) if a handshake took place.
15658 (The default behaviour is needed by applications such as s_client
15659 and s_server that use select() to determine when to use SSL_read;
15660 but for applications that know in advance when to expect data, it
15661 just makes things more complicated.)
15665 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15670 * Add a few more EBCDIC conditionals that make `req` and `x509`
15671 work better on such systems.
15673 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15675 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15676 Update PKCS12_parse() so it copies the friendlyName and the
15677 keyid to the certificates aux info.
15681 * Fix bug in PKCS7_verify() which caused an infinite loop
15682 if there was more than one signature.
15684 *Sven Uszpelkat <su@celocom.de>*
15686 * Major change in util/mkdef.pl to include extra information
15687 about each symbol, as well as presenting variables as well
15688 as functions. This change means that there's n more need
15689 to rebuild the .num files when some algorithms are excluded.
15693 * Allow the verify time to be set by an application,
15694 rather than always using the current time.
15698 * Phase 2 verify code reorganisation. The certificate
15699 verify code now looks up an issuer certificate by a
15700 number of criteria: subject name, authority key id
15701 and key usage. It also verifies self signed certificates
15702 by the same criteria. The main comparison function is
15703 X509_check_issued() which performs these checks.
15705 Lot of changes were necessary in order to support this
15706 without completely rewriting the lookup code.
15708 Authority and subject key identifier are now cached.
15710 The LHASH 'certs' is X509_STORE has now been replaced
15711 by a STACK_OF(X509_OBJECT). This is mainly because an
15712 LHASH can't store or retrieve multiple objects with
15713 the same hash value.
15715 As a result various functions (which were all internal
15716 use only) have changed to handle the new X509_STORE
15717 structure. This will break anything that messed round
15718 with X509_STORE internally.
15720 The functions X509_STORE_add_cert() now checks for an
15721 exact match, rather than just subject name.
15723 The X509_STORE API doesn't directly support the retrieval
15724 of multiple certificates matching a given criteria, however
15725 this can be worked round by performing a lookup first
15726 (which will fill the cache with candidate certificates)
15727 and then examining the cache for matches. This is probably
15728 the best we can do without throwing out X509_LOOKUP
15729 entirely (maybe later...).
15731 The X509_VERIFY_CTX structure has been enhanced considerably.
15733 All certificate lookup operations now go via a get_issuer()
15734 callback. Although this currently uses an X509_STORE it
15735 can be replaced by custom lookups. This is a simple way
15736 to bypass the X509_STORE hackery necessary to make this
15737 work and makes it possible to use more efficient techniques
15738 in future. A very simple version which uses a simple
15739 STACK for its trusted certificate store is also provided
15740 using X509_STORE_CTX_trusted_stack().
15742 The verify_cb() and verify() callbacks now have equivalents
15743 in the X509_STORE_CTX structure.
15745 X509_STORE_CTX also has a 'flags' field which can be used
15746 to customise the verify behaviour.
15750 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15751 excludes S/MIME capabilities.
15755 * When a certificate request is read in keep a copy of the
15756 original encoding of the signed data and use it when outputting
15757 again. Signatures then use the original encoding rather than
15758 a decoded, encoded version which may cause problems if the
15759 request is improperly encoded.
15763 * For consistency with other BIO_puts implementations, call
15764 buffer_write(b, ...) directly in buffer_puts instead of calling
15767 In BIO_puts, increment b->num_write as in BIO_write.
15769 *Peter.Sylvester@EdelWeb.fr*
15771 * Fix BN_mul_word for the case where the word is 0. (We have to use
15772 BN_zero, we may not return a BIGNUM with an array consisting of
15773 words set to zero.)
15777 * Avoid calling abort() from within the library when problems are
15778 detected, except if preprocessor symbols have been defined
15779 (such as REF_CHECK, BN_DEBUG etc.).
15783 * New openssl application 'rsautl'. This utility can be
15784 used for low-level RSA operations. DER public key
15785 BIO/fp routines also added.
15789 * New Configure entry and patches for compiling on QNX 4.
15791 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15793 * A demo state-machine implementation was sponsored by
15794 Nuron (<http://www.nuron.com/>) and is now available in
15795 demos/state_machine.
15799 * New options added to the 'dgst' utility for signature
15800 generation and verification.
15804 * Unrecognized PKCS#7 content types are now handled via a
15805 catch all ASN1_TYPE structure. This allows unsupported
15806 types to be stored as a "blob" and an application can
15807 encode and decode it manually.
15811 * Fix various signed/unsigned issues to make a_strex.c
15812 compile under VC++.
15814 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15816 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15817 length if passed a buffer. ASN1_INTEGER_to_BN failed
15818 if passed a NULL BN and its argument was negative.
15820 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15822 * Modification to PKCS#7 encoding routines to output definite
15823 length encoding. Since currently the whole structures are in
15824 memory there's not real point in using indefinite length
15825 constructed encoding. However if OpenSSL is compiled with
15826 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15830 * Added BIO_vprintf() and BIO_vsnprintf().
15834 * Added more prefixes to parse for in the strings written
15835 through a logging bio, to cover all the levels that are available
15836 through syslog. The prefixes are now:
15838 PANIC, EMERG, EMR => LOG_EMERG
15839 ALERT, ALR => LOG_ALERT
15840 CRIT, CRI => LOG_CRIT
15841 ERROR, ERR => LOG_ERR
15842 WARNING, WARN, WAR => LOG_WARNING
15843 NOTICE, NOTE, NOT => LOG_NOTICE
15844 INFO, INF => LOG_INFO
15845 DEBUG, DBG => LOG_DEBUG
15847 and as before, if none of those prefixes are present at the
15848 beginning of the string, LOG_ERR is chosen.
15850 On Win32, the `LOG_*` levels are mapped according to this:
15852 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15853 LOG_WARNING => EVENTLOG_WARNING_TYPE
15854 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15858 * Made it possible to reconfigure with just the configuration
15859 argument "reconf" or "reconfigure". The command line arguments
15860 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15861 and are retrieved from there when reconfiguring.
15867 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15869 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15873 * The obj_dat.pl script was messing up the sorting of object
15874 names. The reason was that it compared the quoted version
15875 of strings as a result "OCSP" > "OCSP Signing" because
15876 " > SPACE. Changed script to store unquoted versions of
15877 names and add quotes on output. It was also omitting some
15878 names from the lookup table if they were given a default
15879 value (that is if SN is missing it is given the same
15880 value as LN and vice versa), these are now added on the
15881 grounds that if an object has a name we should be able to
15882 look it up. Finally added warning output when duplicate
15883 short or long names are found.
15887 * Changes needed for Tandem NSK.
15889 *Scott Uroff <scott@xypro.com>*
15891 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15892 RSA_padding_check_SSLv23(), special padding was never detected
15893 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15894 version rollback attacks was not effective.
15896 In s23_clnt.c, don't use special rollback-attack detection padding
15897 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15898 client; similarly, in s23_srvr.c, don't do the rollback check if
15899 SSL 2.0 is the only protocol enabled in the server.
15903 * Make it possible to get hexdumps of unprintable data with 'openssl
15904 asn1parse'. By implication, the functions ASN1_parse_dump() and
15905 BIO_dump_indent() are added.
15909 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15910 these print out strings and name structures based on various
15911 flags including RFC2253 support and proper handling of
15912 multibyte characters. Added options to the 'x509' utility
15913 to allow the various flags to be set.
15917 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15918 Also change the functions X509_cmp_current_time() and
15919 X509_gmtime_adj() work with an ASN1_TIME structure,
15920 this will enable certificates using GeneralizedTime in validity
15921 dates to be checked.
15925 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15926 negative public key encodings) on by default,
15927 NO_NEG_PUBKEY_BUG can be set to disable it.
15931 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15932 content octets. An i2c_ASN1_OBJECT is unnecessary because
15933 the encoding can be trivially obtained from the structure.
15937 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15938 not read locks (`CRYPTO_r_[un]lock`).
15942 * A first attempt at creating official support for shared
15943 libraries through configuration. I've kept it so the
15944 default is static libraries only, and the OpenSSL programs
15945 are always statically linked for now, but there are
15946 preparations for dynamic linking in place.
15947 This has been tested on Linux and Tru64.
15951 * Randomness polling function for Win9x, as described in:
15952 Peter Gutmann, Software Generation of Practically Strong
15957 * Fix so PRNG is seeded in req if using an already existing
15962 * New options to smime application. -inform and -outform
15963 allow alternative formats for the S/MIME message including
15964 PEM and DER. The -content option allows the content to be
15965 specified separately. This should allow things like Netscape
15966 form signing output easier to verify.
15970 * Fix the ASN1 encoding of tags using the 'long form'.
15974 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15975 STRING types. These convert content octets to and from the
15976 underlying type. The actual tag and length octets are
15977 already assumed to have been read in and checked. These
15978 are needed because all other string types have virtually
15979 identical handling apart from the tag. By having versions
15980 of the ASN1 functions that just operate on content octets
15981 IMPLICIT tagging can be handled properly. It also allows
15982 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15983 and ASN1_INTEGER are identical apart from the tag.
15987 * Change the handling of OID objects as follows:
15989 - New object identifiers are inserted in objects.txt, following
15990 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15991 - objects.pl is used to process obj_mac.num and create a new
15993 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15996 This is currently kind of a hack, and the perl code in objects.pl
15997 isn't very elegant, but it works as I intended. The simplest way
15998 to check that it worked correctly is to look in obj_dat.h and
15999 check the array nid_objs and make sure the objects haven't moved
16000 around (this is important!). Additions are OK, as well as
16001 consistent name changes.
16005 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16009 * Addition of the command line parameter '-rand file' to 'openssl req'.
16010 The given file adds to whatever has already been seeded into the
16011 random pool through the RANDFILE configuration file option or
16012 environment variable, or the default random state file.
16016 * mkstack.pl now sorts each macro group into lexical order.
16017 Previously the output order depended on the order the files
16018 appeared in the directory, resulting in needless rewriting
16023 * Patches to make OpenSSL compile under Win32 again. Mostly
16024 work arounds for the VC++ problem that it treats func() as
16025 func(void). Also stripped out the parts of mkdef.pl that
16026 added extra typesafe functions: these no longer exist.
16030 * Reorganisation of the stack code. The macros are now all
16031 collected in safestack.h . Each macro is defined in terms of
16032 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16033 DEBUG_SAFESTACK is now handled in terms of function casts,
16034 this has the advantage of retaining type safety without the
16035 use of additional functions. If DEBUG_SAFESTACK is not defined
16036 then the non typesafe macros are used instead. Also modified the
16037 mkstack.pl script to handle the new form. Needs testing to see
16038 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16039 the default if no major problems. Similar behaviour for ASN1_SET_OF
16040 and PKCS12_STACK_OF.
16044 * When some versions of IIS use the 'NET' form of private key the
16045 key derivation algorithm is different. Normally MD5(password) is
16046 used as a 128 bit RC4 key. In the modified case
16047 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16048 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16049 as the old Netscape_RSA functions except they have an additional
16050 'sgckey' parameter which uses the modified algorithm. Also added
16051 an -sgckey command line option to the rsa utility. Thanks to
16052 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16053 algorithm to openssl-dev.
16057 * The evp_local.h macros were using 'c.##kname' which resulted in
16058 invalid expansion on some systems (SCO 5.0.5 for example).
16059 Corrected to 'c.kname'.
16061 *Phillip Porch <root@theporch.com>*
16063 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16064 a STACK of email addresses from a certificate or request, these look
16065 in the subject name and the subject alternative name extensions and
16066 omit any duplicate addresses.
16070 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16071 This makes DSA verification about 2 % faster.
16075 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16076 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16077 plus overhead for 1024 bit moduli).
16078 This makes exponentiations about 0.5 % faster for 1024 bit
16079 exponents (as measured by "openssl speed rsa2048").
16083 * Rename memory handling macros to avoid conflicts with other
16085 Malloc => OPENSSL_malloc
16086 Malloc_locked => OPENSSL_malloc_locked
16087 Realloc => OPENSSL_realloc
16088 Free => OPENSSL_free
16092 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16093 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16097 * CygWin32 support.
16099 *John Jarvie <jjarvie@newsguy.com>*
16101 * The type-safe stack code has been rejigged. It is now only compiled
16102 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16103 by default all type-specific stack functions are "#define"d back to
16104 standard stack functions. This results in more streamlined output
16105 but retains the type-safety checking possibilities of the original
16110 * The STACK code has been cleaned up, and certain type declarations
16111 that didn't make a lot of sense have been brought in line. This has
16112 also involved a cleanup of sorts in safestack.h to more correctly
16113 map type-safe stack functions onto their plain stack counterparts.
16114 This work has also resulted in a variety of "const"ifications of
16115 lots of the code, especially `_cmp` operations which should normally
16116 be prototyped with "const" parameters anyway.
16120 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16121 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16122 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16123 where all of 'md' is used each time the PRNG is used, but 'state'
16124 is used only indexed by a cyclic counter. As entropy may not be
16125 well distributed from the beginning, 'md' is important as a
16126 chaining variable. However, the output function chains only half
16127 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16128 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16129 in all of 'state' being rewritten, with the new values depending
16130 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16134 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16135 the handshake is continued after ssl_verify_cert_chain();
16136 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16137 can lead to 'unexplainable' connection aborts later.
16139 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16141 * Major EVP API cipher revision.
16142 Add hooks for extra EVP features. This allows various cipher
16143 parameters to be set in the EVP interface. Support added for variable
16144 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16145 setting of RC2 and RC5 parameters.
16147 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16150 Remove lots of duplicated code from the EVP library. For example *every*
16151 cipher init() function handles the 'iv' in the same way according to the
16152 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16153 for CFB and OFB modes they zero ctx->num.
16155 New functionality allows removal of S/MIME code RC2 hack.
16157 Most of the routines have the same form and so can be declared in terms
16160 By shifting this to the top level EVP_CipherInit() it can be removed from
16161 all individual ciphers. If the cipher wants to handle IVs or keys
16162 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16165 Change lots of functions like EVP_EncryptUpdate() to now return a
16166 value: although software versions of the algorithms cannot fail
16167 any installed hardware versions can.
16171 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16172 this option is set, tolerate broken clients that send the negotiated
16173 protocol version number instead of the requested protocol version
16178 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16179 i.e. non-zero for export ciphersuites, zero otherwise.
16180 Previous versions had this flag inverted, inconsistent with
16181 rsa_tmp_cb (..._TMP_RSA_CB).
16183 *Bodo Moeller; problem reported by Amit Chopra*
16185 * Add missing DSA library text string. Work around for some IIS
16186 key files with invalid SEQUENCE encoding.
16190 * Add a document (doc/standards.txt) that list all kinds of standards
16191 and so on that are implemented in OpenSSL.
16195 * Enhance c_rehash script. Old version would mishandle certificates
16196 with the same subject name hash and wouldn't handle CRLs at all.
16197 Added -fingerprint option to crl utility, to support new c_rehash
16202 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16206 * Fix for SSL server purpose checking. Server checking was
16207 rejecting certificates which had extended key usage present
16208 but no ssl client purpose.
16210 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16212 * Make PKCS#12 code work with no password. The PKCS#12 spec
16213 is a little unclear about how a blank password is handled.
16214 Since the password in encoded as a BMPString with terminating
16215 double NULL a zero length password would end up as just the
16216 double NULL. However no password at all is different and is
16217 handled differently in the PKCS#12 key generation code. NS
16218 treats a blank password as zero length. MSIE treats it as no
16219 password on export: but it will try both on import. We now do
16220 the same: PKCS12_parse() tries zero length and no password if
16221 the password is set to "" or NULL (NULL is now a valid password:
16222 it wasn't before) as does the pkcs12 application.
16226 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16227 perror when PEM_read_bio_X509_REQ fails, the error message must
16228 be obtained from the error queue.
16232 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16233 it in ERR_remove_state if appropriate, and change ERR_get_state
16234 accordingly to avoid race conditions (this is necessary because
16235 thread_hash is no longer constant once set).
16239 * Bugfix for linux-elf makefile.one.
16243 * RSA_get_default_method() will now cause a default
16244 RSA_METHOD to be chosen if one doesn't exist already.
16245 Previously this was only set during a call to RSA_new()
16246 or RSA_new_method(NULL) meaning it was possible for
16247 RSA_get_default_method() to return NULL.
16251 * Added native name translation to the existing DSO code
16252 that will convert (if the flag to do so is set) filenames
16253 that are sufficiently small and have no path information
16254 into a canonical native form. Eg. "blah" converted to
16255 "libblah.so" or "blah.dll" etc.
16259 * New function ERR_error_string_n(e, buf, len) which is like
16260 ERR_error_string(e, buf), but writes at most 'len' bytes
16261 including the 0 terminator. For ERR_error_string_n, 'buf'
16264 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16266 * CONF library reworked to become more general. A new CONF
16267 configuration file reader "class" is implemented as well as a
16268 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16269 old `CONF_*` functions are still there, but are reimplemented to
16270 work in terms of the new functions. Also, a set of functions
16271 to handle the internal storage of the configuration data is
16272 provided to make it easier to write new configuration file
16273 reader "classes" (I can definitely see something reading a
16274 configuration file in XML format, for example), called `_CONF_*`,
16275 or "the configuration storage API"...
16277 The new configuration file reading functions are:
16279 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16280 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16282 NCONF_default, NCONF_WIN32
16284 NCONF_dump_fp, NCONF_dump_bio
16286 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16287 NCONF_new creates a new CONF object. This works in the same way
16288 as other interfaces in OpenSSL, like the BIO interface.
16289 `NCONF_dump_*` dump the internal storage of the configuration file,
16290 which is useful for debugging. All other functions take the same
16291 arguments as the old `CONF_*` functions with the exception of the
16292 first that must be a `CONF *` instead of a `LHASH *`.
16294 To make it easier to use the new classes with the old `CONF_*` functions,
16295 the function CONF_set_default_method is provided.
16299 * Add '-tls1' option to 'openssl ciphers', which was already
16300 mentioned in the documentation but had not been implemented.
16301 (This option is not yet really useful because even the additional
16302 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16306 * Initial DSO code added into libcrypto for letting OpenSSL (and
16307 OpenSSL-based applications) load shared libraries and bind to
16308 them in a portable way.
16310 *Geoff Thorpe, with contributions from Richard Levitte*
16312 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16314 * Make sure _lrotl and _lrotr are only used with MSVC.
16316 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16317 (the default implementation of RAND_status).
16319 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16320 to '-clrext' (= clear extensions), as intended and documented.
16321 *Bodo Moeller; inconsistency pointed out by Michael Attili
16322 <attili@amaxo.com>*
16324 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16325 was larger than the MD block size.
16327 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16329 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16330 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16331 using the passed key: if the passed key was a private key the result
16332 of X509_print(), for example, would be to print out all the private key
16337 * des_quad_cksum() byte order bug fix.
16338 *Ulf Möller, using the problem description in krb4-0.9.7, where
16339 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16341 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16344 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16346 * For easily testing in shell scripts whether some command
16347 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16348 returns with exit code 0 iff no command of the given name is available.
16349 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16350 the output goes to stdout and nothing is printed to stderr.
16351 Additional arguments are always ignored.
16353 Since for each cipher there is a command of the same name,
16354 the 'no-cipher' compilation switches can be tested this way.
16356 ('openssl no-XXX' is not able to detect pseudo-commands such
16357 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16361 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16365 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16366 is set; it will be thrown away anyway because each handshake creates
16368 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16369 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16370 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16371 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16375 * New s_client option -ign_eof: EOF at stdin is ignored, and
16376 'Q' and 'R' lose their special meanings (quit/renegotiate).
16377 This is part of what -quiet does; unlike -quiet, -ign_eof
16378 does not suppress any output.
16382 * Add compatibility options to the purpose and trust code. The
16383 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16384 accepts a certificate or CA, this was the previous behaviour,
16385 with all the associated security issues.
16387 X509_TRUST_COMPAT is the old trust behaviour: only and
16388 automatically trust self signed roots in certificate store. A
16389 new trust setting X509_TRUST_DEFAULT is used to specify that
16390 a purpose has no associated trust setting and it should instead
16391 use the value in the default purpose.
16395 * Fix the PKCS#8 DSA private key code so it decodes keys again
16396 and fix a memory leak.
16400 * In util/mkerr.pl (which implements 'make errors'), preserve
16401 reason strings from the previous version of the .c file, as
16402 the default to have only downcase letters (and digits) in
16403 automatically generated reasons codes is not always appropriate.
16407 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16408 using strerror. Previously, ERR_reason_error_string() returned
16409 library names as reason strings for SYSerr; but SYSerr is a special
16410 case where small numbers are errno values, not library numbers.
16414 * Add '-dsaparam' option to 'openssl dhparam' application. This
16415 converts DSA parameters into DH parameters. (When creating parameters,
16416 DSA_generate_parameters is used.)
16420 * Include 'length' (recommended exponent length) in C code generated
16421 by 'openssl dhparam -C'.
16425 * The second argument to set_label in perlasm was already being used
16426 so couldn't be used as a "file scope" flag. Moved to third argument
16431 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16432 instead of RAND_bytes for encryption IVs and salts.
16436 * Include RAND_status() into RAND_METHOD instead of implementing
16437 it only for md_rand.c Otherwise replacing the PRNG by calling
16438 RAND_set_rand_method would be impossible.
16442 * Don't let DSA_generate_key() enter an infinite loop if the random
16443 number generation fails.
16447 * New 'rand' application for creating pseudo-random output.
16451 * Added configuration support for Linux/IA64
16453 *Rolf Haberrecker <rolf@suse.de>*
16455 * Assembler module support for Mingw32.
16459 * Shared library support for HPUX (in shlib/).
16461 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16463 * Shared library support for Solaris gcc.
16465 *Lutz Behnke <behnke@trustcenter.de>*
16467 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16469 * PKCS7_encrypt() was adding text MIME headers twice because they
16470 were added manually and by SMIME_crlf_copy().
16474 * In bntest.c don't call BN_rand with zero bits argument.
16476 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16478 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16479 case was implemented. This caused BN_div_recp() to fail occasionally.
16483 * Add an optional second argument to the set_label() in the perl
16484 assembly language builder. If this argument exists and is set
16485 to 1 it signals that the assembler should use a symbol whose
16486 scope is the entire file, not just the current function. This
16487 is needed with MASM which uses the format label:: for this scope.
16489 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16491 * Change the ASN1 types so they are typedefs by default. Before
16492 almost all types were #define'd to ASN1_STRING which was causing
16493 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16498 * Change names of new functions to the new get1/get0 naming
16499 convention: After 'get1', the caller owns a reference count
16500 and has to call `..._free`; 'get0' returns a pointer to some
16501 data structure without incrementing reference counters.
16502 (Some of the existing 'get' functions increment a reference
16503 counter, some don't.)
16504 Similarly, 'set1' and 'add1' functions increase reference
16505 counters or duplicate objects.
16509 * Allow for the possibility of temp RSA key generation failure:
16510 the code used to assume it always worked and crashed on failure.
16514 * Fix potential buffer overrun problem in BIO_printf().
16515 *Ulf Möller, using public domain code by Patrick Powell; problem
16516 pointed out by David Sacerdote <das33@cornell.edu>*
16518 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16519 RAND_egd() and RAND_status(). In the command line application,
16520 the EGD socket can be specified like a seed file using RANDFILE
16525 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16526 Some CAs (e.g. Verisign) distribute certificates in this form.
16530 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16531 list to exclude them. This means that no special compilation option
16532 is needed to use anonymous DH: it just needs to be included in the
16537 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16538 EVP_MD_type. The old functionality is available in a new macro called
16539 EVP_MD_md(). Change code that uses it and update docs.
16543 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16544 where the `void *` argument is replaced by a function pointer argument.
16545 Previously `void *` was abused to point to functions, which works on
16546 many platforms, but is not correct. As these functions are usually
16547 called by macros defined in OpenSSL header files, most source code
16548 should work without changes.
16552 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16553 sections with information on -D... compiler switches used for
16554 compiling the library so that applications can see them. To enable
16555 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16556 must be defined. E.g.,
16557 #define OPENSSL_ALGORITHM_DEFINES
16558 #include <openssl/opensslconf.h>
16559 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16561 *Richard Levitte, Ulf and Bodo Möller*
16563 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16568 * Change the 'other' type in certificate aux info to a STACK_OF
16569 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16570 the required ASN1 format: arbitrary types determined by an OID.
16574 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16575 argument to 'req'. This is not because the function is newer or
16576 better than others it just uses the work 'NEW' in the certificate
16577 request header lines. Some software needs this.
16581 * Reorganise password command line arguments: now passwords can be
16582 obtained from various sources. Delete the PEM_cb function and make
16583 it the default behaviour: i.e. if the callback is NULL and the
16584 usrdata argument is not NULL interpret it as a null terminated pass
16585 phrase. If usrdata and the callback are NULL then the pass phrase
16586 is prompted for as usual.
16590 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16591 the support is automatically enabled. The resulting binaries will
16592 autodetect the card and use it if present.
16594 *Ben Laurie and Compaq Inc.*
16596 * Work around for Netscape hang bug. This sends certificate request
16597 and server done in one record. Since this is perfectly legal in the
16598 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16599 the bugs/SSLv3 entry for more info.
16603 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16607 * Add -rand argument to smime and pkcs12 applications and read/write
16612 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16616 * Add command line password options to the remaining applications.
16620 * Bug fix for BN_div_recp() for numerators with an even number of
16625 * More tests in bntest.c, and changed test_bn output.
16629 * ./config recognizes MacOS X now.
16633 * Bug fix for BN_div() when the first words of num and divisor are
16634 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16638 * Add support for various broken PKCS#8 formats, and command line
16639 options to produce them.
16643 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16644 get temporary BIGNUMs from a BN_CTX.
16648 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16653 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16654 include a #define from the old name to the new. The original intent
16655 was that statically linked binaries could for example just call
16656 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16657 link with digests. This never worked because SSLeay_add_all_digests()
16658 and SSLeay_add_all_ciphers() were in the same source file so calling
16659 one would link with the other. They are now in separate source files.
16663 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16667 * Use a less unusual form of the Miller-Rabin primality test (it used
16668 a binary algorithm for exponentiation integrated into the Miller-Rabin
16669 loop, our standard modexp algorithms are faster).
16673 * Support for the EBCDIC character set completed.
16675 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16677 * Source code cleanups: use const where appropriate, eliminate casts,
16678 use `void *` instead of `char *` in lhash.
16682 * Bugfix: ssl3_send_server_key_exchange was not restartable
16683 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16684 this the server could overwrite ephemeral keys that the client
16689 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16690 using 50 iterations of the Rabin-Miller test.
16692 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16693 iterations of the Rabin-Miller test as required by the appendix
16694 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16695 As BN_is_prime_fasttest includes trial division, DSA parameter
16696 generation becomes much faster.
16698 This implies a change for the callback functions in DSA_is_prime
16699 and DSA_generate_parameters: The callback function is called once
16700 for each positive witness in the Rabin-Miller test, not just
16701 occasionally in the inner loop; and the parameters to the
16702 callback function now provide an iteration count for the outer
16703 loop rather than for the current invocation of the inner loop.
16704 DSA_generate_parameters additionally can call the callback
16705 function with an 'iteration count' of -1, meaning that a
16706 candidate has passed the trial division test (when q is generated
16707 from an application-provided seed, trial division is skipped).
16711 * New function BN_is_prime_fasttest that optionally does trial
16712 division before starting the Rabin-Miller test and has
16713 an additional BN_CTX * argument (whereas BN_is_prime always
16714 has to allocate at least one BN_CTX).
16715 'callback(1, -1, cb_arg)' is called when a number has passed the
16716 trial division stage.
16720 * Fix for bug in CRL encoding. The validity dates weren't being handled
16725 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16729 * New function BN_pseudo_rand().
16733 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16734 bignum version of BN_from_montgomery() with the working code from
16735 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16740 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16741 made it impossible to use the same SSL_SESSION data structure in
16742 SSL2 clients in multiple threads.
16746 * The return value of RAND_load_file() no longer counts bytes obtained
16747 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16748 to seed the PRNG (previously an explicit byte count was required).
16750 *Ulf Möller, Bodo Möller*
16752 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16753 used `char *` instead of `void *` and had casts all over the place.
16757 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16761 * Retain source code compatibility for BN_prime_checks macro:
16762 BN_is_prime(..., BN_prime_checks, ...) now uses
16763 BN_prime_checks_for_size to determine the appropriate number of
16764 Rabin-Miller iterations.
16768 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16769 DH_CHECK_P_NOT_SAFE_PRIME.
16770 (Check if this is true? OpenPGP calls them "strong".)
16774 * Merge the functionality of "dh" and "gendh" programs into a new program
16775 "dhparam". The old programs are retained for now but will handle DH keys
16776 (instead of parameters) in future.
16780 * Make the ciphers, s_server and s_client programs check the return values
16781 when a new cipher list is set.
16785 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16786 ciphers. Before when the 56bit ciphers were enabled the sorting was
16789 The syntax for the cipher sorting has been extended to support sorting by
16790 cipher-strength (using the strength_bits hard coded in the tables).
16791 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16793 Fix a bug in the cipher-command parser: when supplying a cipher command
16794 string with an "undefined" symbol (neither command nor alphanumeric
16795 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16796 an error is flagged.
16798 Due to the strength-sorting extension, the code of the
16799 ssl_create_cipher_list() function was completely rearranged. I hope that
16800 the readability was also increased :-)
16802 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16804 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16805 for the first serial number and places 2 in the serial number file. This
16806 avoids problems when the root CA is created with serial number zero and
16807 the first user certificate has the same issuer name and serial number
16812 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16813 the new code. Add documentation for this stuff.
16817 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16818 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16819 structures and behave in an analogous way to the X509v3 functions:
16820 they shouldn't be called directly but wrapper functions should be used
16823 So we also now have some wrapper functions that call the X509at functions
16824 when passed certificate requests. (TO DO: similar things can be done with
16825 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16826 things. Some of these need some d2i or i2d and print functionality
16827 because they handle more complex structures.)
16831 * Add missing #ifndefs that caused missing symbols when building libssl
16832 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16833 NO_RSA in `ssl/s2*.c`.
16835 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16837 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16838 has a return value which indicates the quality of the random data
16839 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16840 error queue. New function RAND_pseudo_bytes() generates output that is
16841 guaranteed to be unique but not unpredictable. RAND_add is like
16842 RAND_seed, but takes an extra argument for an entropy estimate
16843 (RAND_seed always assumes full entropy).
16847 * Do more iterations of Rabin-Miller probable prime test (specifically,
16848 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16849 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16850 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16851 false-positive rate of at most 2^-80 for random input.
16855 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16859 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16860 in the 0.9.5 release), this returns the chain
16861 from an X509_CTX structure with a dup of the stack and all
16862 the X509 reference counts upped: so the stack will exist
16863 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16866 Also make SSL_SESSION_print() print out the verify return
16871 * Add manpage for the pkcs12 command. Also change the default
16872 behaviour so MAC iteration counts are used unless the new
16873 -nomaciter option is used. This improves file security and
16874 only older versions of MSIE (4.0 for example) need it.
16878 * Honor the no-xxx Configure options when creating .DEF files.
16882 * Add PKCS#10 attributes to field table: challengePassword,
16883 unstructuredName and unstructuredAddress. These are taken from
16884 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16885 international characters are used.
16887 More changes to X509_ATTRIBUTE code: allow the setting of types
16888 based on strings. Remove the 'loc' parameter when adding
16889 attributes because these will be a SET OF encoding which is sorted
16894 * Initial changes to the 'req' utility to allow request generation
16895 automation. This will allow an application to just generate a template
16896 file containing all the field values and have req construct the
16899 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16900 used all over the place including certificate requests and PKCS#7
16901 structures. They are currently handled manually where necessary with
16902 some primitive wrappers for PKCS#7. The new functions behave in a
16903 manner analogous to the X509 extension functions: they allow
16904 attributes to be looked up by NID and added.
16906 Later something similar to the X509V3 code would be desirable to
16907 automatically handle the encoding, decoding and printing of the
16908 more complex types. The string types like challengePassword can
16909 be handled by the string table functions.
16911 Also modified the multi byte string table handling. Now there is
16912 a 'global mask' which masks out certain types. The table itself
16913 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16914 is useful when for example there is only one permissible type
16915 (as in countryName) and using the mask might result in no valid
16920 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16921 SSL_get_peer_finished to allow applications to obtain the latest
16922 Finished messages sent to the peer or expected from the peer,
16923 respectively. (SSL_get_peer_finished is usually the Finished message
16924 actually received from the peer, otherwise the protocol will be aborted.)
16926 As the Finished message are message digests of the complete handshake
16927 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16928 be used for external authentication procedures when the authentication
16929 provided by SSL/TLS is not desired or is not enough.
16933 * Enhanced support for Alpha Linux is added. Now ./config checks if
16934 the host supports BWX extension and if Compaq C is present on the
16935 $PATH. Just exploiting of the BWX extension results in 20-30%
16936 performance kick for some algorithms, e.g. DES and RC4 to mention
16937 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16942 * Add support for MS "fast SGC". This is arguably a violation of the
16943 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16944 weak crypto and after checking the certificate is SGC a second one
16945 with strong crypto. MS SGC stops the first handshake after receiving
16946 the server certificate message and sends a second client hello. Since
16947 a server will typically do all the time consuming operations before
16948 expecting any further messages from the client (server key exchange
16949 is the most expensive) there is little difference between the two.
16951 To get OpenSSL to support MS SGC we have to permit a second client
16952 hello message after we have sent server done. In addition we have to
16953 reset the MAC if we do get this second client hello.
16957 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16958 if a DER encoded private key is RSA or DSA traditional format. Changed
16959 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16960 format DER encoded private key. Newer code should use PKCS#8 format which
16961 has the key type encoded in the ASN1 structure. Added DER private key
16962 support to pkcs8 application.
16966 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16967 ciphersuites has been selected (as required by the SSL 3/TLS 1
16968 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16969 is set, we interpret this as a request to violate the specification
16970 (the worst that can happen is a handshake failure, and 'correct'
16971 behaviour would result in a handshake failure anyway).
16975 * In SSL_CTX_add_session, take into account that there might be multiple
16976 SSL_SESSION structures with the same session ID (e.g. when two threads
16977 concurrently obtain them from an external cache).
16978 The internal cache can handle only one SSL_SESSION with a given ID,
16979 so if there's a conflict, we now throw out the old one to achieve
16984 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16985 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16986 some routines that use cipher OIDs: some ciphers do not have OIDs
16987 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16992 * Simplify the trust setting structure and code. Now we just have
16993 two sequences of OIDs for trusted and rejected settings. These will
16994 typically have values the same as the extended key usage extension
16995 and any application specific purposes.
16997 The trust checking code now has a default behaviour: it will just
16998 check for an object with the same NID as the passed id. Functions can
16999 be provided to override either the default behaviour or the behaviour
17000 for a given id. SSL client, server and email already have functions
17001 in place for compatibility: they check the NID and also return "trusted"
17002 if the certificate is self signed.
17006 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17007 traditional format into an EVP_PKEY structure.
17011 * Add a password callback function PEM_cb() which either prompts for
17012 a password if usr_data is NULL or otherwise assumes it is a null
17013 terminated password. Allow passwords to be passed on command line
17014 environment or config files in a few more utilities.
17018 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17019 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17020 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17021 Update documentation.
17025 * Support for ASN1 "NULL" type. This could be handled before by using
17026 ASN1_TYPE but there wasn't any function that would try to read a NULL
17027 and produce an error if it couldn't. For compatibility we also have
17028 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17029 don't allocate anything because they don't need to.
17033 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17036 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17038 * Rebuild of the memory allocation routines used by OpenSSL code and
17039 possibly others as well. The purpose is to make an interface that
17040 provide hooks so anyone can build a separate set of allocation and
17041 deallocation routines to be used by OpenSSL, for example memory
17042 pool implementations, or something else, which was previously hard
17043 since Malloc(), Realloc() and Free() were defined as macros having
17044 the values malloc, realloc and free, respectively (except for Win32
17045 compilations). The same is provided for memory debugging code.
17046 OpenSSL already comes with functionality to find memory leaks, but
17047 this gives people a chance to debug other memory problems.
17049 With these changes, a new set of functions and macros have appeared:
17051 CRYPTO_set_mem_debug_functions() [F]
17052 CRYPTO_get_mem_debug_functions() [F]
17053 CRYPTO_dbg_set_options() [F]
17054 CRYPTO_dbg_get_options() [F]
17055 CRYPTO_malloc_debug_init() [M]
17057 The memory debug functions are NULL by default, unless the library
17058 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17059 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17060 gives the standard debugging functions that come with OpenSSL) or
17061 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17062 provided by the library user) must be used. When the standard
17063 debugging functions are used, CRYPTO_dbg_set_options can be used to
17064 request additional information:
17065 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17066 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17068 Also, things like CRYPTO_set_mem_functions will always give the
17069 expected result (the new set of functions is used for allocation
17070 and deallocation) at all times, regardless of platform and compiler
17073 To finish it up, some functions that were never use in any other
17074 way than through macros have a new API and new semantic:
17076 CRYPTO_dbg_malloc()
17077 CRYPTO_dbg_realloc()
17080 All macros of value have retained their old syntax.
17082 *Richard Levitte and Bodo Moeller*
17084 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17085 ordering of SMIMECapabilities wasn't in "strength order" and there
17086 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17091 * Some ASN1 types with illegal zero length encoding (INTEGER,
17092 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17094 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17096 * Merge in my S/MIME library for OpenSSL. This provides a simple
17097 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17098 functionality to handle multipart/signed properly) and a utility
17099 called 'smime' to call all this stuff. This is based on code I
17100 originally wrote for Celo who have kindly allowed it to be
17101 included in OpenSSL.
17105 * Add variants des_set_key_checked and des_set_key_unchecked of
17106 des_set_key (aka des_key_sched). Global variable des_check_key
17107 decides which of these is called by des_set_key; this way
17108 des_check_key behaves as it always did, but applications and
17109 the library itself, which was buggy for des_check_key == 1,
17110 have a cleaner way to pick the version they need.
17114 * New function PKCS12_newpass() which changes the password of a
17119 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17120 dynamic mix. In both cases the ids can be used as an index into the
17121 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17122 functions so they accept a list of the field values and the
17123 application doesn't need to directly manipulate the X509_TRUST
17128 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17133 * Modify the way the V3 extension code looks up extensions. This now
17134 works in a similar way to the object code: we have some "standard"
17135 extensions in a static table which is searched with OBJ_bsearch()
17136 and the application can add dynamic ones if needed. The file
17137 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17138 updated whenever a new extension is added to the core code and kept
17139 in ext_nid order. There is a simple program 'tabtest.c' which checks
17140 this. New extensions are not added too often so this file can readily
17141 be maintained manually.
17143 There are two big advantages in doing things this way. The extensions
17144 can be looked up immediately and no longer need to be "added" using
17145 X509V3_add_standard_extensions(): this function now does nothing.
17146 Side note: I get *lots* of email saying the extension code doesn't
17147 work because people forget to call this function.
17148 Also no dynamic allocation is done unless new extensions are added:
17149 so if we don't add custom extensions there is no need to call
17150 X509V3_EXT_cleanup().
17154 * Modify enc utility's salting as follows: make salting the default. Add a
17155 magic header, so unsalted files fail gracefully instead of just decrypting
17156 to garbage. This is because not salting is a big security hole, so people
17157 should be discouraged from doing it.
17161 * Fixes and enhancements to the 'x509' utility. It allowed a message
17162 digest to be passed on the command line but it only used this
17163 parameter when signing a certificate. Modified so all relevant
17164 operations are affected by the digest parameter including the
17165 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17166 DSA key was used because it didn't fix the digest.
17170 * Initial certificate chain verify code. Currently tests the untrusted
17171 certificates for consistency with the verify purpose (which is set
17172 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17174 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17175 this is because it will reject chains with invalid extensions whereas
17176 every previous version of OpenSSL and SSLeay made no checks at all.
17178 Trust code: checks the root CA for the relevant trust settings. Trust
17179 settings have an initial value consistent with the verify purpose: e.g.
17180 if the verify purpose is for SSL client use it expects the CA to be
17181 trusted for SSL client use. However the default value can be changed to
17182 permit custom trust settings: one example of this would be to only trust
17183 certificates from a specific "secure" set of CAs.
17185 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17186 which should be used for version portability: especially since the
17187 verify structure is likely to change more often now.
17189 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17190 to set them. If not set then assume SSL clients will verify SSL servers
17193 Two new options to the verify program: -untrusted allows a set of
17194 untrusted certificates to be passed in and -purpose which sets the
17195 intended purpose of the certificate. If a purpose is set then the
17196 new chain verify code is used to check extension consistency.
17200 * Support for the authority information access extension.
17204 * Modify RSA and DSA PEM read routines to transparently handle
17205 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17206 public keys in a format compatible with certificate
17207 SubjectPublicKeyInfo structures. Unfortunately there were already
17208 functions called *_PublicKey_* which used various odd formats so
17209 these are retained for compatibility: however the DSA variants were
17210 never in a public release so they have been deleted. Changed dsa/rsa
17211 utilities to handle the new format: note no releases ever handled public
17212 keys so we should be OK.
17214 The primary motivation for this change is to avoid the same fiasco
17215 that dogs private keys: there are several incompatible private key
17216 formats some of which are standard and some OpenSSL specific and
17217 require various evil hacks to allow partial transparent handling and
17218 even then it doesn't work with DER formats. Given the option anything
17219 other than PKCS#8 should be dumped: but the other formats have to
17220 stay in the name of compatibility.
17222 With public keys and the benefit of hindsight one standard format
17223 is used which works with EVP_PKEY, RSA or DSA structures: though
17224 it clearly returns an error if you try to read the wrong kind of key.
17226 Added a -pubkey option to the 'x509' utility to output the public key.
17227 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17228 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17229 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17230 that do the same as the `EVP_PKEY_assign_*()` except they up the
17231 reference count of the added key (they don't "swallow" the
17236 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17237 CRLs would fail if the file contained no certificates or no CRLs:
17238 added a new function to read in both types and return the number
17239 read: this means that if none are read it will be an error. The
17240 DER versions of the certificate and CRL reader would always fail
17241 because it isn't possible to mix certificates and CRLs in DER format
17242 without choking one or the other routine. Changed this to just read
17243 a certificate: this is the best we can do. Also modified the code
17244 in `apps/verify.c` to take notice of return codes: it was previously
17245 attempting to read in certificates from NULL pointers and ignoring
17246 any errors: this is one reason why the cert and CRL reader seemed
17247 to work. It doesn't check return codes from the default certificate
17248 routines: these may well fail if the certificates aren't installed.
17252 * Code to support otherName option in GeneralName.
17256 * First update to verify code. Change the verify utility
17257 so it warns if it is passed a self signed certificate:
17258 for consistency with the normal behaviour. X509_verify
17259 has been modified to it will now verify a self signed
17260 certificate if *exactly* the same certificate appears
17261 in the store: it was previously impossible to trust a
17262 single self signed certificate. This means that:
17263 openssl verify ss.pem
17264 now gives a warning about a self signed certificate but
17265 openssl verify -CAfile ss.pem ss.pem
17270 * For servers, store verify_result in SSL_SESSION data structure
17271 (and add it to external session representation).
17272 This is needed when client certificate verifications fails,
17273 but an application-provided verification callback (set by
17274 SSL_CTX_set_cert_verify_callback) allows accepting the session
17275 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17276 but returns 1): When the session is reused, we have to set
17277 ssl->verify_result to the appropriate error code to avoid
17280 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17282 * Fix a bug in the new PKCS#7 code: it didn't consider the
17283 case in PKCS7_dataInit() where the signed PKCS7 structure
17284 didn't contain any existing data because it was being created.
17286 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17288 * Add a salt to the key derivation routines in enc.c. This
17289 forms the first 8 bytes of the encrypted file. Also add a
17290 -S option to allow a salt to be input on the command line.
17294 * New function X509_cmp(). Oddly enough there wasn't a function
17295 to compare two certificates. We do this by working out the SHA1
17296 hash and comparing that. X509_cmp() will be needed by the trust
17301 * SSL_get1_session() is like SSL_get_session(), but increments
17302 the reference count in the SSL_SESSION returned.
17304 *Geoff Thorpe <geoff@eu.c2.net>*
17306 * Fix for 'req': it was adding a null to request attributes.
17307 Also change the X509_LOOKUP and X509_INFO code to handle
17308 certificate auxiliary information.
17312 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17317 * Add the possibility to add extra information to the memory leak
17318 detecting output, to form tracebacks, showing from where each
17319 allocation was originated: CRYPTO_push_info("constant string") adds
17320 the string plus current file name and line number to a per-thread
17321 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17322 is like calling CYRPTO_pop_info() until the stack is empty.
17323 Also updated memory leak detection code to be multi-thread-safe.
17327 * Add options -text and -noout to pkcs7 utility and delete the
17328 encryption options which never did anything. Update docs.
17332 * Add options to some of the utilities to allow the pass phrase
17333 to be included on either the command line (not recommended on
17334 OSes like Unix) or read from the environment. Update the
17335 manpages and fix a few bugs.
17339 * Add a few manpages for some of the openssl commands.
17343 * Fix the -revoke option in ca. It was freeing up memory twice,
17344 leaking and not finding already revoked certificates.
17348 * Extensive changes to support certificate auxiliary information.
17349 This involves the use of X509_CERT_AUX structure and X509_AUX
17350 functions. An X509_AUX function such as PEM_read_X509_AUX()
17351 can still read in a certificate file in the usual way but it
17352 will also read in any additional "auxiliary information". By
17353 doing things this way a fair degree of compatibility can be
17354 retained: existing certificates can have this information added
17355 using the new 'x509' options.
17357 Current auxiliary information includes an "alias" and some trust
17358 settings. The trust settings will ultimately be used in enhanced
17359 certificate chain verification routines: currently a certificate
17360 can only be trusted if it is self signed and then it is trusted
17365 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17366 The problem was that one of the replacement routines had not been working
17367 since SSLeay releases. For now the offending routine has been replaced
17368 with non-optimised assembler. Even so, this now gives around 95%
17369 performance improvement for 1024 bit RSA signs.
17373 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17374 handling. Most clients have the effective key size in bits equal to
17375 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17376 A few however don't do this and instead use the size of the decrypted key
17377 to determine the RC2 key length and the AlgorithmIdentifier to determine
17378 the effective key length. In this case the effective key length can still
17379 be 40 bits but the key length can be 168 bits for example. This is fixed
17380 by manually forcing an RC2 key into the EVP_PKEY structure because the
17381 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17382 the key length and effective key length are equal.
17386 * Add a bunch of functions that should simplify the creation of
17387 X509_NAME structures. Now you should be able to do:
17388 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17389 and have it automatically work out the correct field type and fill in
17390 the structures. The more adventurous can try:
17391 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17392 and it will (hopefully) work out the correct multibyte encoding.
17396 * Change the 'req' utility to use the new field handling and multibyte
17397 copy routines. Before the DN field creation was handled in an ad hoc
17398 way in req, ca, and x509 which was rather broken and didn't support
17399 BMPStrings or UTF8Strings. Since some software doesn't implement
17400 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17401 using the dirstring_type option. See the new comment in the default
17402 openssl.cnf for more info.
17406 * Make crypto/rand/md_rand.c more robust:
17407 - Assure unique random numbers after fork().
17408 - Make sure that concurrent threads access the global counter and
17409 md serializably so that we never lose entropy in them
17410 or use exactly the same state in multiple threads.
17411 Access to the large state is not always serializable because
17412 the additional locking could be a performance killer, and
17413 md should be large enough anyway.
17417 * New file `apps/app_rand.c` with commonly needed functionality
17418 for handling the random seed file.
17420 Use the random seed file in some applications that previously did not:
17422 dsaparam -genkey (which also ignored its '-rand' option),
17425 x509 (when signing).
17426 Except on systems with /dev/urandom, it is crucial to have a random
17427 seed file at least for key creation, DSA signing, and for DH exchanges;
17428 for RSA signatures we could do without one.
17430 gendh and gendsa (unlike genrsa) used to read only the first byte
17431 of each file listed in the '-rand' option. The function as previously
17432 found in genrsa is now in app_rand.c and is used by all programs
17433 that support '-rand'.
17437 * In RAND_write_file, use mode 0600 for creating files;
17438 don't just chmod when it may be too late.
17442 * Report an error from X509_STORE_load_locations
17443 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17447 * New function ASN1_mbstring_copy() this copies a string in either
17448 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17449 into an ASN1_STRING type. A mask of permissible types is passed
17450 and it chooses the "minimal" type to use or an error if not type
17455 * Add function equivalents to the various macros in asn1.h. The old
17456 macros are retained with an `M_` prefix. Code inside the library can
17457 use the `M_` macros. External code (including the openssl utility)
17458 should *NOT* in order to be "shared library friendly".
17462 * Add various functions that can check a certificate's extensions
17463 to see if it usable for various purposes such as SSL client,
17464 server or S/MIME and CAs of these types. This is currently
17465 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17466 verification. Also added a -purpose flag to x509 utility to
17467 print out all the purposes.
17471 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17476 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17477 for, obtain and decode and extension and obtain its critical flag.
17478 This allows all the necessary extension code to be handled in a
17479 single function call.
17483 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17484 platforms. See crypto/rc4/rc4_enc.c for further details.
17488 * New -noout option to asn1parse. This causes no output to be produced
17489 its main use is when combined with -strparse and -out to extract data
17490 from a file (which may not be in ASN.1 format).
17494 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17495 when producing the local key id.
17497 *Richard Levitte <levitte@stacken.kth.se>*
17499 * New option -dhparam in s_server. This allows a DH parameter file to be
17500 stated explicitly. If it is not stated then it tries the first server
17501 certificate file. The previous behaviour hard coded the filename
17506 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17507 a public key to be input or output. For example:
17508 openssl rsa -in key.pem -pubout -out pubkey.pem
17509 Also added necessary DSA public key functions to handle this.
17513 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17514 in the message. This was handled by allowing
17515 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17517 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17519 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17520 to the end of the strings whereas this didn't. This would cause problems
17521 if strings read with d2i_ASN1_bytes() were later modified.
17523 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17525 * Fix for base64 decode bug. When a base64 bio reads only one line of
17526 data and it contains EOF it will end up returning an error. This is
17527 caused by input 46 bytes long. The cause is due to the way base64
17528 BIOs find the start of base64 encoded data. They do this by trying a
17529 trial decode on each line until they find one that works. When they
17530 do a flag is set and it starts again knowing it can pass all the
17531 data directly through the decoder. Unfortunately it doesn't reset
17532 the context it uses. This means that if EOF is reached an attempt
17533 is made to pass two EOFs through the context and this causes the
17534 resulting error. This can also cause other problems as well. As is
17535 usual with these problems it takes *ages* to find and the fix is
17536 trivial: move one line.
17538 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17540 * Ugly workaround to get s_client and s_server working under Windows. The
17541 old code wouldn't work because it needed to select() on sockets and the
17542 tty (for keypresses and to see if data could be written). Win32 only
17543 supports select() on sockets so we select() with a 1s timeout on the
17544 sockets and then see if any characters are waiting to be read, if none
17545 are present then we retry, we also assume we can always write data to
17546 the tty. This isn't nice because the code then blocks until we've
17547 received a complete line of data and it is effectively polling the
17548 keyboard at 1s intervals: however it's quite a bit better than not
17549 working at all :-) A dedicated Windows application might handle this
17550 with an event loop for example.
17554 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17555 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17556 will be called when RSA_sign() and RSA_verify() are used. This is useful
17557 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17558 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17559 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17560 This necessitated the support of an extra signature type NID_md5_sha1
17561 for SSL signatures and modifications to the SSL library to use it instead
17562 of calling RSA_public_decrypt() and RSA_private_encrypt().
17566 * Add new -verify -CAfile and -CApath options to the crl program, these
17567 will lookup a CRL issuers certificate and verify the signature in a
17568 similar way to the verify program. Tidy up the crl program so it
17569 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17570 less strict. It will now permit CRL extensions even if it is not
17571 a V2 CRL: this will allow it to tolerate some broken CRLs.
17575 * Initialize all non-automatic variables each time one of the openssl
17576 sub-programs is started (this is necessary as they may be started
17577 multiple times from the "OpenSSL>" prompt).
17579 *Lennart Bang, Bodo Moeller*
17581 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17582 removing all other RSA functionality (this is what NO_RSA does). This
17583 is so (for example) those in the US can disable those operations covered
17584 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17589 * Non-copying interface to BIO pairs.
17590 (still largely untested)
17594 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17595 ASCII string. This was handled independently in various places before.
17599 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17600 UTF8 strings a character at a time.
17604 * Use client_version from client hello to select the protocol
17605 (s23_srvr.c) and for RSA client key exchange verification
17606 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17610 * Add various utility functions to handle SPKACs, these were previously
17611 handled by poking round in the structure internals. Added new function
17612 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17613 print, verify and generate SPKACs. Based on an original idea from
17614 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17618 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17622 * Allow the config file extension section to be overwritten on the
17623 command line. Based on an original idea from Massimiliano Pala
17624 <madwolf@comune.modena.it>. The new option is called -extensions
17625 and can be applied to ca, req and x509. Also -reqexts to override
17626 the request extensions in req and -crlexts to override the crl extensions
17631 * Add new feature to the SPKAC handling in ca. Now you can include
17632 the same field multiple times by preceding it by "XXXX." for example:
17635 this is the same syntax as used in the req config file.
17639 * Allow certificate extensions to be added to certificate requests. These
17640 are specified in a 'req_extensions' option of the req section of the
17641 config file. They can be printed out with the -text option to req but
17642 are otherwise ignored at present.
17646 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17647 data read consists of only the final block it would not decrypted because
17648 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17649 A misplaced 'break' also meant the decrypted final block might not be
17650 copied until the next read.
17654 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17655 a few extra parameters to the DH structure: these will be useful if
17656 for example we want the value of 'q' or implement X9.42 DH.
17660 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17661 provides hooks that allow the default DSA functions or functions on a
17662 "per key" basis to be replaced. This allows hardware acceleration and
17663 hardware key storage to be handled without major modification to the
17664 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17665 associated functions.
17669 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17670 as "read only": it can't be written to and the buffer it points to will
17671 not be freed. Reading from a read only BIO is much more efficient than
17672 a normal memory BIO. This was added because there are several times when
17673 an area of memory needs to be read from a BIO. The previous method was
17674 to create a memory BIO and write the data to it, this results in two
17675 copies of the data and an O(n^2) reading algorithm. There is a new
17676 function BIO_new_mem_buf() which creates a read only memory BIO from
17677 an area of memory. Also modified the PKCS#7 routines to use read only
17682 * Bugfix: ssl23_get_client_hello did not work properly when called in
17683 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17684 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17685 but a retry condition occurred while trying to read the rest.
17689 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17690 NID_pkcs7_encrypted by default: this was wrong since this should almost
17691 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17692 the encrypted data type: this is a more sensible place to put it and it
17693 allows the PKCS#12 code to be tidied up that duplicated this
17698 * Changed obj_dat.pl script so it takes its input and output files on
17699 the command line. This should avoid shell escape redirection problems
17704 * Initial support for certificate extension requests, these are included
17705 in things like Xenroll certificate requests. Included functions to allow
17706 extensions to be obtained and added.
17710 * -crlf option to s_client and s_server for sending newlines as
17711 CRLF (as required by many protocols).
17715 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17717 * Install libRSAglue.a when OpenSSL is built with RSAref.
17719 *Ralf S. Engelschall*
17721 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17723 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17725 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17730 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17731 DH parameters/keys (q is lost during that conversion, but the resulting
17732 DH parameters contain its length).
17734 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17735 much faster than DH_generate_parameters (which creates parameters
17736 where `p = 2*q + 1`), and also the smaller q makes DH computations
17737 much more efficient (160-bit exponentiation instead of 1024-bit
17738 exponentiation); so this provides a convenient way to support DHE
17739 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17740 utter importance to use
17741 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17743 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17744 when such DH parameters are used, because otherwise small subgroup
17745 attacks may become possible!
17749 * Avoid memory leak in i2d_DHparams.
17753 * Allow the -k option to be used more than once in the enc program:
17754 this allows the same encrypted message to be read by multiple recipients.
17758 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17759 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17760 it will always use the numerical form of the OID, even if it has a short
17765 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17766 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17767 otherwise bn_mod_exp was called. In the case of hardware keys for example
17768 no private key components need be present and it might store extra data
17769 in the RSA structure, which cannot be accessed from bn_mod_exp.
17770 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17771 private key operations.
17775 * Added support for SPARC Linux.
17779 * pem_password_cb function type incompatibly changed from
17780 typedef int pem_password_cb(char *buf, int size, int rwflag);
17782 ....(char *buf, int size, int rwflag, void *userdata);
17783 so that applications can pass data to their callbacks:
17784 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17785 additional void * argument, which is just handed through whenever
17786 the password callback is called.
17788 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17790 New function SSL_CTX_set_default_passwd_cb_userdata.
17792 Compatibility note: As many C implementations push function arguments
17793 onto the stack in reverse order, the new library version is likely to
17794 interoperate with programs that have been compiled with the old
17795 pem_password_cb definition (PEM_whatever takes some data that
17796 happens to be on the stack as its last argument, and the callback
17797 just ignores this garbage); but there is no guarantee whatsoever that
17800 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17801 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17802 problems not only on Windows, but also on some Unix platforms.
17803 To avoid problematic command lines, these definitions are now in an
17804 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17805 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17809 * MIPS III/IV assembler module is reimplemented.
17813 * More DES library cleanups: remove references to srand/rand and
17814 delete an unused file.
17818 * Add support for the free Netwide assembler (NASM) under Win32,
17819 since not many people have MASM (ml) and it can be hard to obtain.
17820 This is currently experimental but it seems to work OK and pass all
17821 the tests. Check out INSTALL.W32 for info.
17825 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17826 without temporary keys kept an extra copy of the server key,
17827 and connections with temporary keys did not free everything in case
17832 * New function RSA_check_key and new openssl rsa option -check
17833 for verifying the consistency of RSA keys.
17835 *Ulf Moeller, Bodo Moeller*
17837 * Various changes to make Win32 compile work:
17838 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17839 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17840 comparison" warnings.
17841 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17845 * Add a debugging option to PKCS#5 v2 key generation function: when
17846 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17847 derived keys are printed to stderr.
17851 * Copy the flags in ASN1_STRING_dup().
17853 *Roman E. Pavlov <pre@mo.msk.ru>*
17855 * The x509 application mishandled signing requests containing DSA
17856 keys when the signing key was also DSA and the parameters didn't match.
17858 It was supposed to omit the parameters when they matched the signing key:
17859 the verifying software was then supposed to automatically use the CA's
17860 parameters if they were absent from the end user certificate.
17862 Omitting parameters is no longer recommended. The test was also
17863 the wrong way round! This was probably due to unusual behaviour in
17864 EVP_cmp_parameters() which returns 1 if the parameters match.
17865 This meant that parameters were omitted when they *didn't* match and
17866 the certificate was useless. Certificates signed with 'ca' didn't have
17869 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17871 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17872 The interface is as follows:
17873 Applications can use
17874 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17875 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17876 "off" is now the default.
17877 The library internally uses
17878 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17879 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17880 to disable memory-checking temporarily.
17882 Some inconsistent states that previously were possible (and were
17883 even the default) are now avoided.
17885 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17886 with each memory chunk allocated; this is occasionally more helpful
17887 than just having a counter.
17889 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17891 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17896 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17897 which largely parallels "options", but is for changing API behaviour,
17898 whereas "options" are about protocol behaviour.
17899 Initial "mode" flags are:
17901 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17902 a single record has been written.
17903 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17904 retries use the same buffer location.
17905 (But all of the contents must be
17910 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17913 * Fix problems with no-hmac etc.
17915 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17917 * New functions RSA_get_default_method(), RSA_set_method() and
17918 RSA_get_method(). These allows replacement of RSA_METHODs without having
17919 to mess around with the internals of an RSA structure.
17923 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17924 Also really enable memory leak checks in openssl.c and in some
17927 *Chad C. Mulligan, Bodo Moeller*
17929 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17930 up the length of negative integers. This has now been simplified to just
17931 store the length when it is first determined and use it later, rather
17932 than trying to keep track of where data is copied and updating it to
17934 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17936 * Add a new function PKCS7_signatureVerify. This allows the verification
17937 of a PKCS#7 signature but with the signing certificate passed to the
17938 function itself. This contrasts with PKCS7_dataVerify which assumes the
17939 certificate is present in the PKCS#7 structure. This isn't always the
17940 case: certificates can be omitted from a PKCS#7 structure and be
17941 distributed by "out of band" means (such as a certificate database).
17945 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17946 function prototypes in pem.h, also change util/mkdef.pl to add the
17947 necessary function names.
17951 * mk1mf.pl (used by Windows builds) did not properly read the
17952 options set by Configure in the top level Makefile, and Configure
17953 was not even able to write more than one option correctly.
17954 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17958 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17959 file to be loaded from a BIO or FILE pointer. The BIO version will
17960 for example allow memory BIOs to contain config info.
17964 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17965 Whoever hopes to achieve shared-library compatibility across versions
17966 must use this, not the compile-time macro.
17967 (Exercise 0.9.4: Which is the minimum library version required by
17969 Note: All this applies only to multi-threaded programs, others don't
17974 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17975 through a BIO pair triggered the default case, i.e.
17976 SSLerr(...,SSL_R_UNKNOWN_STATE).
17980 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17981 can use the SSL library even if none of the specific BIOs is
17986 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17987 for the encoded length.
17989 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17991 * Add initial documentation of the X509V3 functions.
17995 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17996 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17997 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17998 secure PKCS#8 private key format with a high iteration count.
18002 * Fix determination of Perl interpreter: A perl or perl5
18003 *directory* in $PATH was also accepted as the interpreter.
18005 *Ralf S. Engelschall*
18007 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18008 wrong with it but it was very old and did things like calling
18009 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18010 unusual formatting.
18014 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18015 to use the new extension code.
18019 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18020 with macros. This should make it easier to change their form, add extra
18021 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18026 * Add to configuration table a new entry that can specify an alternative
18027 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18028 according to Mark Crispin <MRC@Panda.COM>.
18032 * DES CBC did not update the IV. Weird.
18036 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18037 Changing the behaviour of the former might break existing programs --
18038 where IV updating is needed, des_ncbc_encrypt can be used.
18041 * When bntest is run from "make test" it drives bc to check its
18042 calculations, as well as internally checking them. If an internal check
18043 fails, it needs to cause bc to give a non-zero result or make test carries
18044 on without noticing the failure. Fixed.
18048 * DES library cleanups.
18052 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18053 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18054 ciphers. NOTE: although the key derivation function has been verified
18055 against some published test vectors it has not been extensively tested
18056 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18061 * Instead of "mkdir -p", which is not fully portable, use new
18062 Perl script "util/mkdir-p.pl".
18066 * Rewrite the way password based encryption (PBE) is handled. It used to
18067 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18068 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18069 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18070 the 'parameter' field of the AlgorithmIdentifier is passed to the
18071 underlying key generation function so it must do its own ASN1 parsing.
18072 This has also changed the EVP_PBE_CipherInit() function which now has a
18073 'parameter' argument instead of literal salt and iteration count values
18074 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18078 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18079 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18080 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18081 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18082 value was just used as a "magic string" and not used directly its
18083 value doesn't matter.
18087 * Introduce some semblance of const correctness to BN. Shame C doesn't
18092 * "linux-sparc64" configuration (ultrapenguin).
18094 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18095 "linux-sparc" configuration.
18097 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18099 * config now generates no-xxx options for missing ciphers.
18103 * Support the EBCDIC character set (work in progress).
18104 File ebcdic.c not yet included because it has a different license.
18106 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18108 * Support BS2000/OSD-POSIX.
18110 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18112 * Make callbacks for key generation use `void *` instead of `char *`.
18116 * Make S/MIME samples compile (not yet tested).
18120 * Additional typesafe stacks.
18124 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18128 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18130 * New configuration variant "sco5-gcc".
18132 * Updated some demos.
18134 *Sean O Riordain, Wade Scholine*
18136 * Add missing BIO_free at exit of pkcs12 application.
18140 * Fix memory leak in conf.c.
18144 * Updates for Win32 to assembler version of MD5.
18148 * Set #! path to perl in `apps/der_chop` to where we found it
18149 instead of using a fixed path.
18153 * SHA library changes for irix64-mips4-cc.
18157 * Improvements for VMS support.
18161 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18163 * Bignum library bug fix. IRIX 6 passes "make test" now!
18164 This also avoids the problems with SC4.2 and unpatched SC5.
18166 *Andy Polyakov <appro@fy.chalmers.se>*
18168 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18169 These are required because of the typesafe stack would otherwise break
18170 existing code. If old code used a structure member which used to be STACK
18171 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18172 sk_num or sk_value it would produce an error because the num, data members
18173 are not present in STACK_OF. Now it just produces a warning. sk_set
18174 replaces the old method of assigning a value to sk_value
18175 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18176 that does this will no longer work (and should use sk_set instead) but
18177 this could be regarded as a "questionable" behaviour anyway.
18181 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18182 correctly handle encrypted S/MIME data.
18186 * Change type of various DES function arguments from des_cblock
18187 (which means, in function argument declarations, pointer to char)
18188 to des_cblock * (meaning pointer to array with 8 char elements),
18189 which allows the compiler to do more typechecking; it was like
18190 that back in SSLeay, but with lots of ugly casts.
18192 Introduce new type const_des_cblock.
18196 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18197 problems: find RecipientInfo structure that matches recipient certificate
18198 and initialise the ASN1 structures properly based on passed cipher.
18202 * Belatedly make the BN tests actually check the results.
18206 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18207 to and from BNs: it was completely broken. New compilation option
18208 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18209 key elements as negative integers.
18213 * Reorganize and speed up MD5.
18215 *Andy Polyakov <appro@fy.chalmers.se>*
18219 *Richard Levitte <richard@levitte.org>*
18221 * New option -out to asn1parse to allow the parsed structure to be
18222 output to a file. This is most useful when combined with the -strparse
18223 option to examine the output of things like OCTET STRINGS.
18227 * Make SSL library a little more fool-proof by not requiring any longer
18228 that `SSL_set_{accept,connect}_state` be called before
18229 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18230 in many applications because usually everything *appeared* to work as
18231 intended anyway -- now it really works as intended).
18235 * Move openssl.cnf out of lib/.
18239 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18240 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18241 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18243 *Ralf S. Engelschall*
18245 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18246 handle PKCS#7 enveloped data properly.
18248 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18250 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18251 copying pointers. The cert_st handling is changed by this in
18252 various ways (and thus what used to be known as ctx->default_cert
18253 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18254 any longer when s->cert does not give us what we need).
18255 ssl_cert_instantiate becomes obsolete by this change.
18256 As soon as we've got the new code right (possibly it already is?),
18257 we have solved a couple of bugs of the earlier code where s->cert
18258 was used as if it could not have been shared with other SSL structures.
18260 Note that using the SSL API in certain dirty ways now will result
18261 in different behaviour than observed with earlier library versions:
18262 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18263 does not influence s as it used to.
18265 In order to clean up things more thoroughly, inside SSL_SESSION
18266 we don't use CERT any longer, but a new structure SESS_CERT
18267 that holds per-session data (if available); currently, this is
18268 the peer's certificate chain and, for clients, the server's certificate
18269 and temporary key. CERT holds only those values that can have
18270 meaningful defaults in an SSL_CTX.
18274 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18275 from the internal representation. Various PKCS#7 fixes: remove some
18276 evil casts and set the enc_dig_alg field properly based on the signing
18281 * Allow PKCS#12 password to be set from the command line or the
18282 environment. Let 'ca' get its config file name from the environment
18283 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18288 * Allow certificate policies extension to use an IA5STRING for the
18289 organization field. This is contrary to the PKIX definition but
18290 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18295 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18296 without disallowing inline assembler and the like for non-pedantic builds.
18300 * Support Borland C++ builder.
18302 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18308 * SHA-1 cleanups and performance enhancements.
18310 *Andy Polyakov <appro@fy.chalmers.se>*
18312 * Sparc v8plus assembler for the bignum library.
18314 *Andy Polyakov <appro@fy.chalmers.se>*
18316 * Accept any -xxx and +xxx compiler options in Configure.
18320 * Update HPUX configuration.
18324 * Add missing `sk_<type>_unshift()` function to safestack.h
18326 *Ralf S. Engelschall*
18328 * New function SSL_CTX_use_certificate_chain_file that sets the
18329 "extra_cert"s in addition to the certificate. (This makes sense
18330 only for "PEM" format files, as chains as a whole are not
18335 * Support verify_depth from the SSL API.
18336 x509_vfy.c had what can be considered an off-by-one-error:
18337 Its depth (which was not part of the external interface)
18338 was actually counting the number of certificates in a chain;
18339 now it really counts the depth.
18343 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18344 instead of X509err, which often resulted in confusing error
18345 messages since the error codes are not globally unique
18346 (e.g. an alleged error in ssl3_accept when a certificate
18347 didn't match the private key).
18349 * New function SSL_CTX_set_session_id_context that allows to set a default
18350 value (so that you don't need SSL_set_session_id_context for each
18351 connection using the SSL_CTX).
18355 * OAEP decoding bug fix.
18359 * Support INSTALL_PREFIX for package builders, as proposed by
18364 * New Configure options "threads" and "no-threads". For systems
18365 where the proper compiler options are known (currently Solaris
18366 and Linux), "threads" is the default.
18370 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18374 * Install various scripts to $(OPENSSLDIR)/misc, not to
18375 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18376 such as /usr/local/bin.
18380 * "make linux-shared" to build shared libraries.
18382 *Niels Poppe <niels@netbox.org>*
18384 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18388 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18389 extension adding in x509 utility.
18393 * Remove NOPROTO sections and error code comments.
18397 * Partial rewrite of the DEF file generator to now parse the ANSI
18402 * New Configure options --prefix=DIR and --openssldir=DIR.
18406 * Complete rewrite of the error code script(s). It is all now handled
18407 by one script at the top level which handles error code gathering,
18408 header rewriting and C source file generation. It should be much better
18409 than the old method: it now uses a modified version of Ulf's parser to
18410 read the ANSI prototypes in all header files (thus the old K&R definitions
18411 aren't needed for error creation any more) and do a better job of
18412 translating function codes into names. The old 'ASN1 error code embedded
18413 in a comment' is no longer necessary and it doesn't use .err files which
18414 have now been deleted. Also the error code call doesn't have to appear all
18415 on one line (which resulted in some large lines...).
18419 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18423 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18424 0 (which usually indicates a closed connection), but continue reading.
18428 * Fix some race conditions.
18432 * Add support for CRL distribution points extension. Add Certificate
18433 Policies and CRL distribution points documentation.
18437 * Move the autogenerated header file parts to crypto/opensslconf.h.
18441 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18442 8 of keying material. Merlin has also confirmed interop with this fix
18443 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18445 *Merlin Hughes <merlin@baltimore.ie>*
18447 * Fix lots of warnings.
18449 *Richard Levitte <levitte@stacken.kth.se>*
18451 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18452 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18454 *Richard Levitte <levitte@stacken.kth.se>*
18456 * Fix problems with sizeof(long) == 8.
18458 *Andy Polyakov <appro@fy.chalmers.se>*
18460 * Change functions to ANSI C.
18464 * Fix typos in error codes.
18466 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18468 * Remove defunct assembler files from Configure.
18472 * SPARC v8 assembler BIGNUM implementation.
18474 *Andy Polyakov <appro@fy.chalmers.se>*
18476 * Support for Certificate Policies extension: both print and set.
18477 Various additions to support the r2i method this uses.
18481 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18482 return a const string when you are expecting an allocated buffer.
18486 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18487 types DirectoryString and DisplayText.
18491 * Add code to allow r2i extensions to access the configuration database,
18492 add an LHASH database driver and add several ctx helper functions.
18496 * Fix an evil bug in bn_expand2() which caused various BN functions to
18497 fail when they extended the size of a BIGNUM.
18501 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18502 support typesafe stack.
18506 * Fix typo in SSL_[gs]et_options().
18508 *Nils Frostberg <nils@medcom.se>*
18510 * Delete various functions and files that belonged to the (now obsolete)
18511 old X509V3 handling code.
18515 * New Configure option "rsaref".
18519 * Don't auto-generate pem.h.
18523 * Introduce type-safe ASN.1 SETs.
18527 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18529 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18531 * Introduce type-safe STACKs. This will almost certainly break lots of code
18532 that links with OpenSSL (well at least cause lots of warnings), but fear
18533 not: the conversion is trivial, and it eliminates loads of evil casts. A
18534 few STACKed things have been converted already. Feel free to convert more.
18535 In the fullness of time, I'll do away with the STACK type altogether.
18539 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18540 specified in `<certfile>` by updating the entry in the index.txt file.
18541 This way one no longer has to edit the index.txt file manually for
18542 revoking a certificate. The -revoke option does the gory details now.
18544 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18546 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18547 `-text` option at all and this way the `-noout -text` combination was
18548 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18550 *Ralf S. Engelschall*
18552 * Make sure a corresponding plain text error message exists for the
18553 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18554 verify callback function determined that a certificate was revoked.
18556 *Ralf S. Engelschall*
18558 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18559 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18560 all available ciphers including rc5, which was forgotten until now.
18561 In order to let the testing shell script know which algorithms
18562 are available, a new (up to now undocumented) command
18563 `openssl list-cipher-commands` is used.
18567 * Bugfix: s_client occasionally would sleep in select() when
18568 it should have checked SSL_pending() first.
18572 * New functions DSA_do_sign and DSA_do_verify to provide access to
18573 the raw DSA values prior to ASN.1 encoding.
18577 * Tweaks to Configure
18579 *Niels Poppe <niels@netbox.org>*
18581 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18586 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18590 * New config option to avoid instructions that are illegal on the 80386.
18591 The default code is faster, but requires at least a 486.
18595 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18596 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18597 same as SSL2_VERSION anyway.
18601 * New "-showcerts" option for s_client.
18605 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18606 application. Various cleanups and fixes.
18610 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18611 modify error routines to work internally. Add error codes and PBE init
18612 to library startup routines.
18616 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18617 packing functions to asn1 and evp. Changed function names and error
18618 codes along the way.
18622 * PKCS12 integration: and so it begins... First of several patches to
18623 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18624 objects to objects.h
18628 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18629 and display support for Thawte strong extranet extension.
18633 * Add LinuxPPC support.
18635 *Jeff Dubrule <igor@pobox.org>*
18637 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18638 bn_div_words in alpha.s.
18640 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18642 * Make sure the RSA OAEP test is skipped under -DRSAref because
18643 OAEP isn't supported when OpenSSL is built with RSAref.
18645 *Ulf Moeller <ulf@fitug.de>*
18647 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18648 so they no longer are missing under -DNOPROTO.
18650 *Soren S. Jorvang <soren@t.dk>*
18652 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18654 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18655 doesn't work when the session is reused. Coming soon!
18659 * Fix a security hole, that allows sessions to be reused in the wrong
18660 context thus bypassing client cert protection! All software that uses
18661 client certs and session caches in multiple contexts NEEDS PATCHING to
18662 allow session reuse! A fuller solution is in the works.
18664 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18666 * Some more source tree cleanups (removed obsolete files
18667 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18668 permission on "config" script to be executable) and a fix for the INSTALL
18671 *Ulf Moeller <ulf@fitug.de>*
18673 * Remove some legacy and erroneous uses of malloc, free instead of
18676 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18678 * Make rsa_oaep_test return non-zero on error.
18680 *Ulf Moeller <ulf@fitug.de>*
18682 * Add support for native Solaris shared libraries. Configure
18683 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18684 if someone would make that last step automatic.
18686 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18688 * ctx_size was not built with the right compiler during "make links". Fixed.
18692 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18693 except NULL ciphers". This means the default cipher list will no longer
18694 enable NULL ciphers. They need to be specifically enabled e.g. with
18695 the string "DEFAULT:eNULL".
18699 * Fix to RSA private encryption routines: if p < q then it would
18700 occasionally produce an invalid result. This will only happen with
18701 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18705 * Be less restrictive and allow also `perl util/perlpath.pl
18706 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18707 because this way one can also use an interpreter named `perl5` (which is
18708 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18709 installed as `perl`).
18711 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18713 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18715 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18717 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18718 advapi32.lib to Win32 build and change the pem test comparison
18719 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18720 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18721 and crypto/des/ede_cbcm_enc.c.
18725 * DES quad checksum was broken on big-endian architectures. Fixed.
18729 * Comment out two functions in bio.h that aren't implemented. Fix up the
18730 Win32 test batch file so it (might) work again. The Win32 test batch file
18731 is horrible: I feel ill....
18735 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18736 in e_os.h. Audit of header files to check ANSI and non ANSI
18737 sections: 10 functions were absent from non ANSI section and not exported
18738 from Windows DLLs. Fixed up libeay.num for new functions.
18742 * Make `openssl version` output lines consistent.
18744 *Ralf S. Engelschall*
18746 * Fix Win32 symbol export lists for BIO functions: Added
18747 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18748 to ms/libeay{16,32}.def.
18750 *Ralf S. Engelschall*
18752 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18753 fine under Unix and passes some trivial tests I've now added. But the
18754 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18755 added to make sure no one expects that this stuff really works in the
18756 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18757 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18760 *Ralf S. Engelschall*
18762 * Fix the generation of two part addresses in perl.
18764 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18766 * Add config entry for Linux on MIPS.
18768 *John Tobey <jtobey@channel1.com>*
18770 * Make links whenever Configure is run, unless we are on Windoze.
18774 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18775 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18780 * Add a useful kludge to allow package maintainers to specify compiler and
18781 other platforms details on the command line without having to patch the
18782 Configure script every time: One now can use
18783 `perl Configure <id>:<details>`,
18784 i.e. platform ids are allowed to have details appended
18785 to them (separated by colons). This is treated as there would be a static
18786 pre-configured entry in Configure's %table under key `<id>` with value
18787 `<details>` and `perl Configure <id>` is called. So, when you want to
18788 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18789 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18790 now, which overrides the FreeBSD-elf entry on-the-fly.
18792 *Ralf S. Engelschall*
18794 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18798 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18799 on the `perl Configure ...` command line. This way one can compile
18800 OpenSSL libraries with Position Independent Code (PIC) which is needed
18801 for linking it into DSOs.
18803 *Ralf S. Engelschall*
18805 * Remarkably, export ciphers were totally broken and no-one had noticed!
18810 * Cleaned up the LICENSE document: The official contact for any license
18811 questions now is the OpenSSL core team under openssl-core@openssl.org.
18812 And add a paragraph about the dual-license situation to make sure people
18813 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18814 to the OpenSSL toolkit.
18816 *Ralf S. Engelschall*
18818 * General source tree makefile cleanups: Made `making xxx in yyy...`
18819 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18820 Additionally cleaned up the `make links` target: Remove unnecessary
18821 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18822 to speed processing and no longer clutter the display with confusing
18823 stuff. Instead only the actually done links are displayed.
18825 *Ralf S. Engelschall*
18827 * Permit null encryption ciphersuites, used for authentication only. It used
18828 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18829 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18834 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18835 signed attributes when verifying signatures (this would break them),
18836 the detached data encoding was wrong and public keys obtained using
18837 X509_get_pubkey() weren't freed.
18841 * Add text documentation for the BUFFER functions. Also added a work around
18842 to a Win95 console bug. This was triggered by the password read stuff: the
18843 last character typed gets carried over to the next fread(). If you were
18844 generating a new cert request using 'req' for example then the last
18845 character of the passphrase would be CR which would then enter the first
18850 * Added the new 'Includes OpenSSL Cryptography Software' button as
18851 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18852 button and can be used by applications based on OpenSSL to show the
18853 relationship to the OpenSSL project.
18855 *Ralf S. Engelschall*
18857 * Remove confusing variables in function signatures in files
18858 ssl/ssl_lib.c and ssl/ssl.h.
18860 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18862 * Don't install bss_file.c under PREFIX/include/
18864 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18866 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18867 functions that return function pointers and has support for NT specific
18868 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18869 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18870 unsigned to signed types: this was killing the Win32 compile.
18874 * Add new certificate file to stack functions,
18875 SSL_add_dir_cert_subjects_to_stack() and
18876 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18877 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18878 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18879 This means that Apache-SSL and similar packages don't have to mess around
18880 to add as many CAs as they want to the preferred list.
18884 * Experiment with doxygen documentation. Currently only partially applied to
18886 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18887 openssl.doxy as the configuration file.
18891 * Get rid of remaining C++-style comments which strict C compilers hate.
18893 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18895 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18896 compiled in by default: it has problems with large keys.
18900 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18901 DH private keys and/or callback functions which directly correspond to
18902 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18903 is needed for applications which have to configure certificates on a
18904 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18906 For the RSA certificate situation is makes no difference, but
18907 for the DSA certificate situation this fixes the "no shared cipher"
18908 problem where the OpenSSL cipher selection procedure failed because the
18909 temporary keys were not overtaken from the context and the API provided
18910 no way to reconfigure them.
18911 The new functions now let applications reconfigure the stuff and they
18912 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18913 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18914 non-public-API function ssl_cert_instantiate() is used as a helper
18915 function and also to reduce code redundancy inside ssl_rsa.c.
18917 *Ralf S. Engelschall*
18919 * Move s_server -dcert and -dkey options out of the undocumented feature
18920 area because they are useful for the DSA situation and should be
18921 recognized by the users.
18923 *Ralf S. Engelschall*
18925 * Fix the cipher decision scheme for export ciphers: the export bits are
18926 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18927 SSL_EXP_MASK. So, the original variable has to be used instead of the
18928 already masked variable.
18930 *Richard Levitte <levitte@stacken.kth.se>*
18932 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18934 *Richard Levitte <levitte@stacken.kth.se>*
18936 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18937 from `int` to `unsigned int` because it is a length and initialized by
18938 EVP_DigestFinal() which expects an `unsigned int *`.
18940 *Richard Levitte <levitte@stacken.kth.se>*
18942 * Don't hard-code path to Perl interpreter on shebang line of Configure
18943 script. Instead use the usual Shell->Perl transition trick.
18945 *Ralf S. Engelschall*
18947 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18948 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18949 -noout -modulus` as it's already the case for `openssl rsa -noout
18950 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18951 currently the public key is printed (a decision which was already done by
18952 `openssl dsa -modulus` in the past) which serves a similar purpose.
18953 Additionally the NO_RSA no longer completely removes the whole -modulus
18954 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18957 *Ralf S. Engelschall*
18959 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18960 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18962 *Arne Ansper <arne@ats.cyber.ee>*
18964 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18965 to be added. Now both 'req' and 'ca' can use new objects defined in the
18970 * Add cool BIO that does syslog (or event log on NT).
18972 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18974 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18975 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18976 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18977 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18981 * Add preliminary config info for new extension code.
18985 * Make RSA_NO_PADDING really use no padding.
18987 *Ulf Moeller <ulf@fitug.de>*
18989 * Generate errors when private/public key check is done.
18993 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18994 for some CRL extensions and new objects added.
18998 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18999 key usage extension and fuller support for authority key id.
19003 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19004 padding method for RSA, which is recommended for new applications in PKCS
19005 #1 v2.0 (RFC 2437, October 1998).
19006 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19007 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19008 against Bleichbacher's attack on RSA.
19009 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19012 * Updates to the new SSL compression code
19014 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19016 * Fix so that the version number in the master secret, when passed
19017 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19018 (because the server will not accept higher), that the version number
19019 is 0x03,0x01, not 0x03,0x00
19021 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19023 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19024 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19025 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19029 * Support for RAW extensions where an arbitrary extension can be
19030 created by including its DER encoding. See `apps/openssl.cnf` for
19035 * Make sure latest Perl versions don't interpret some generated C array
19036 code as Perl array code in the crypto/err/err_genc.pl script.
19038 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19040 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19041 not many people have the assembler. Various Win32 compilation fixes and
19042 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19043 build instructions.
19047 * Modify configure script 'Configure' to automatically create crypto/date.h
19048 file under Win32 and also build pem.h from pem.org. New script
19049 util/mkfiles.pl to create the MINFO file on environments that can't do a
19050 'make files': perl util/mkfiles.pl >MINFO should work.
19054 * Major rework of DES function declarations, in the pursuit of correctness
19055 and purity. As a result, many evil casts evaporated, and some weirdness,
19056 too. You may find this causes warnings in your code. Zapping your evil
19057 casts will probably fix them. Mostly.
19061 * Fix for a typo in asn1.h. Bug fix to object creation script
19062 obj_dat.pl. It considered a zero in an object definition to mean
19063 "end of object": none of the objects in objects.h have any zeros
19064 so it wasn't spotted.
19066 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19068 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19069 Masking (CBCM). In the absence of test vectors, the best I have been able
19070 to do is check that the decrypt undoes the encrypt, so far. Send me test
19071 vectors if you have them.
19075 * Correct calculation of key length for export ciphers (too much space was
19076 allocated for null ciphers). This has not been tested!
19080 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19081 message is now correct (it understands "crypto" and "ssl" on its
19082 command line). There is also now an "update" option. This will update
19083 the util/ssleay.num and util/libeay.num files with any new functions.
19085 perl util/mkdef.pl crypto ssl update
19086 it will update them.
19090 * Overhauled the Perl interface:
19091 - ported BN stuff to OpenSSL's different BN library
19092 - made the perl/ source tree CVS-aware
19093 - renamed the package from SSLeay to OpenSSL (the files still contain
19094 their history because I've copied them in the repository)
19095 - removed obsolete files (the test scripts will be replaced
19096 by better Test::Harness variants in the future)
19098 *Ralf S. Engelschall*
19100 * First cut for a very conservative source tree cleanup:
19101 1. merge various obsolete readme texts into doc/ssleay.txt
19102 where we collect the old documents and readme texts.
19103 2. remove the first part of files where I'm already sure that we no
19104 longer need them because of three reasons: either they are just temporary
19105 files which were left by Eric or they are preserved original files where
19106 I've verified that the diff is also available in the CVS via "cvs diff
19107 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19108 the crypto/md/ stuff).
19110 *Ralf S. Engelschall*
19112 * More extension code. Incomplete support for subject and issuer alt
19113 name, issuer and authority key id. Change the i2v function parameters
19114 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19115 what that's for :-) Fix to ASN1 macro which messed up
19116 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19120 * Preliminary support for ENUMERATED type. This is largely copied from the
19125 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19127 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19129 * Make sure `make rehash` target really finds the `openssl` program.
19131 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19133 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19134 like to hear about it if this slows down other processors.
19138 * Add CygWin32 platform information to Configure script.
19140 *Alan Batie <batie@aahz.jf.intel.com>*
19142 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19144 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19146 * New program nseq to manipulate netscape certificate sequences
19150 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19155 * Fixes to BN code. Previously the default was to define BN_RECURSION
19156 but the BN code had some problems that would cause failures when
19157 doing certificate verification and some other functions.
19159 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19161 * Add ASN1 and PEM code to support netscape certificate sequences.
19165 * Add ASN1 and PEM code to support netscape certificate sequences.
19169 * Add several PKIX and private extended key usage OIDs.
19173 * Modify the 'ca' program to handle the new extension code. Modify
19174 openssl.cnf for new extension format, add comments.
19178 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19179 and add a sample to openssl.cnf so req -x509 now adds appropriate
19184 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19185 error code, add initial support to X509_print() and x509 application.
19189 * Takes a deep breath and start adding X509 V3 extension support code. Add
19190 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19191 stuff is currently isolated and isn't even compiled yet.
19195 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19196 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19197 Removed the versions check from X509 routines when loading extensions:
19198 this allows certain broken certificates that don't set the version
19199 properly to be processed.
19203 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19204 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19205 can still be regenerated with "make depend".
19209 * Spelling mistake in C version of CAST-128.
19211 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19213 * Changes to the error generation code. The perl script err-code.pl
19214 now reads in the old error codes and retains the old numbers, only
19215 adding new ones if necessary. It also only changes the .err files if new
19216 codes are added. The makefiles have been modified to only insert errors
19217 when needed (to avoid needlessly modifying header files). This is done
19218 by only inserting errors if the .err file is newer than the auto generated
19219 C file. To rebuild all the error codes from scratch (the old behaviour)
19220 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19221 or delete all the .err files.
19225 * CAST-128 was incorrectly implemented for short keys. The C version has
19226 been fixed, but is untested. The assembler versions are also fixed, but
19227 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19228 to regenerate it if needed.
19229 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19230 Hagino <itojun@kame.net>*
19232 * File was opened incorrectly in randfile.c.
19234 *Ulf Möller <ulf@fitug.de>*
19236 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19237 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19238 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19239 al: it's just almost always a UTCTime. Note this patch adds new error
19240 codes so do a "make errors" if there are problems.
19244 * Correct Linux 1 recognition in config.
19246 *Ulf Möller <ulf@fitug.de>*
19248 * Remove pointless MD5 hash when using DSA keys in ca.
19250 *Anonymous <nobody@replay.com>*
19252 * Generate an error if given an empty string as a cert directory. Also
19253 generate an error if handed NULL (previously returned 0 to indicate an
19254 error, but didn't set one).
19256 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19258 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19262 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19263 parameters. This was causing a warning which killed off the Win32 compile.
19267 * Remove C++ style comments from crypto/bn/bn_local.h.
19269 *Neil Costigan <neil.costigan@celocom.com>*
19271 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19272 based on a text string, looking up short and long names and finally
19273 "dot" format. The "dot" format stuff didn't work. Added new function
19274 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19275 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19276 OID is not part of the table.
19280 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19281 X509_LOOKUP_by_alias().
19285 * Sort openssl functions by name.
19289 * Get the `gendsa` command working and add it to the `list` command. Remove
19290 encryption from sample DSA keys (in case anyone is interested the password
19295 * Make *all* `*_free` functions accept a NULL pointer.
19297 *Frans Heymans <fheymans@isaserver.be>*
19299 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19302 *Anonymous <nobody@replay.com>*
19304 * s_server should send the CAfile as acceptable CAs, not its own cert.
19306 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19308 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19310 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19312 * Temp key "for export" tests were wrong in s3_srvr.c.
19314 *Anonymous <nobody@replay.com>*
19316 * Add prototype for temp key callback functions
19317 SSL_CTX_set_tmp_{rsa,dh}_callback().
19321 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19322 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19326 * X509_name_add_entry() freed the wrong thing after an error.
19328 *Arne Ansper <arne@ats.cyber.ee>*
19330 * rsa_eay.c would attempt to free a NULL context.
19332 *Arne Ansper <arne@ats.cyber.ee>*
19334 * BIO_s_socket() had a broken should_retry() on Windoze.
19336 *Arne Ansper <arne@ats.cyber.ee>*
19338 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19340 *Arne Ansper <arne@ats.cyber.ee>*
19342 * Make sure the already existing X509_STORE->depth variable is initialized
19343 in X509_STORE_new(), but document the fact that this variable is still
19344 unused in the certificate verification process.
19346 *Ralf S. Engelschall*
19348 * Fix the various library and `apps/` files to free up pkeys obtained from
19349 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19353 * Fix reference counting in X509_PUBKEY_get(). This makes
19354 demos/maurice/example2.c work, amongst others, probably.
19356 *Steve Henson and Ben Laurie*
19358 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19359 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19360 are no longer created. This way we have a single and consistent command
19361 line interface `openssl <command>`, similar to `cvs <command>`.
19363 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19365 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19366 BIT STRING wrapper always have zero unused bits.
19370 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19374 * Make the top-level INSTALL documentation easier to understand.
19378 * Makefiles updated to exit if an error occurs in a sub-directory
19379 make (including if user presses ^C) [Paul Sutton]
19381 * Make Montgomery context stuff explicit in RSA data structure.
19385 * Fix build order of pem and err to allow for generated pem.h.
19389 * Fix renumbering bug in X509_NAME_delete_entry().
19393 * Enhanced the err-ins.pl script so it makes the error library number
19394 global and can add a library name. This is needed for external ASN1 and
19395 other error libraries.
19399 * Fixed sk_insert which never worked properly.
19403 * Fix ASN1 macros so they can handle indefinite length constructed
19404 EXPLICIT tags. Some non standard certificates use these: they can now
19409 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19410 into a single doc/ssleay.txt bundle. This way the information is still
19411 preserved but no longer messes up this directory. Now it's new room for
19412 the new set of documentation files.
19414 *Ralf S. Engelschall*
19416 * SETs were incorrectly DER encoded. This was a major pain, because they
19417 shared code with SEQUENCEs, which aren't coded the same. This means that
19418 almost everything to do with SETs or SEQUENCEs has either changed name or
19419 number of arguments.
19421 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19423 * Fix test data to work with the above.
19427 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19428 was already fixed by Eric for 0.9.1 it seems.
19430 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19432 * Autodetect FreeBSD3.
19436 * Fix various bugs in Configure. This affects the following platforms:
19440 unixware-2.0-pentium
19445 * Eliminate generated files from CVS. Reorder tests to regenerate files
19446 before they are needed.
19450 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19454 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19456 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19457 changed SSLeay to OpenSSL in version strings.
19459 *Ralf S. Engelschall*
19461 * Some fixups to the top-level documents.
19465 * Fixed the nasty bug where rsaref.h was not found under compile-time
19466 because the symlink to include/ was missing.
19468 *Ralf S. Engelschall*
19470 * Incorporated the popular no-RSA/DSA-only patches
19471 which allow to compile a RSA-free SSLeay.
19473 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19475 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19476 when "ssleay" is still not found.
19478 *Ralf S. Engelschall*
19480 * Added more platforms to Configure: Cray T3E, HPUX 11,
19482 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19484 * Updated the README file.
19486 *Ralf S. Engelschall*
19488 * Added various .cvsignore files in the CVS repository subdirs
19489 to make a "cvs update" really silent.
19491 *Ralf S. Engelschall*
19493 * Recompiled the error-definition header files and added
19494 missing symbols to the Win32 linker tables.
19496 *Ralf S. Engelschall*
19498 * Cleaned up the top-level documents;
19499 o new files: CHANGES and LICENSE
19500 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19501 o merged COPYRIGHT into LICENSE
19502 o removed obsolete TODO file
19503 o renamed MICROSOFT to INSTALL.W32
19505 *Ralf S. Engelschall*
19507 * Removed dummy files from the 0.9.1b source tree:
19508 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19509 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19510 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19511 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19512 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19514 *Ralf S. Engelschall*
19516 * Added various platform portability fixes.
19520 * The Genesis of the OpenSSL rpject:
19521 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19522 Young and Tim J. Hudson created while they were working for C2Net until
19525 *The OpenSSL Project*
19527 ### Changes between 0.9.0b and 0.9.1b [not released]
19529 * Updated a few CA certificates under certs/
19533 * Changed some BIGNUM api stuff.
19537 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19538 DGUX x86, Linux Alpha, etc.
19542 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19543 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19548 * Add -strparse option to asn1pars program which parses nested
19551 *Dr Stephen Henson <shenson@bigfoot.com>*
19553 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19557 * DSA fix for "ca" program.
19561 * Added "-genkey" option to "dsaparam" program.
19565 * Added RIPE MD160 (rmd160) message digest.
19569 * Added -a (all) option to "ssleay version" command.
19573 * Added PLATFORM define which is the id given to Configure.
19577 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19581 * Extended the ASN.1 parser routines.
19585 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19589 * Added a BN_CTX to the BN library.
19593 * Fixed the weak key values in DES library
19597 * Changed API in EVP library for cipher aliases.
19601 * Added support for RC2/64bit cipher.
19605 * Converted the lhash library to the crypto/mem.c functions.
19609 * Added more recognized ASN.1 object ids.
19613 * Added more RSA padding checks for SSL/TLS.
19617 * Added BIO proxy/filter functionality.
19621 * Added extra_certs to SSL_CTX which can be used
19622 send extra CA certificates to the client in the CA cert chain sending
19623 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19627 * Now Fortezza is denied in the authentication phase because
19628 this is key exchange mechanism is not supported by SSLeay at all.
19632 * Additional PKCS1 checks.
19636 * Support the string "TLSv1" for all TLS v1 ciphers.
19640 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19641 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19645 * Fixed a few memory leaks.
19649 * Fixed various code and comment typos.
19653 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19654 bytes sent in the client random.
19656 *Edward Bishop <ebishop@spyglass.com>*
19660 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19661 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19662 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19663 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19664 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19665 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19666 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19667 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19668 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19669 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19670 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19671 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19672 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19673 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19674 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19675 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19676 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19677 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19678 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19679 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19680 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19681 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19682 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19683 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19684 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19685 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19686 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19687 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19688 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19689 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19690 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19691 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19692 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19693 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19694 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19695 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19696 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19697 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19698 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19699 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19700 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19701 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19702 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19703 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19704 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19705 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19706 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19707 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19708 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19709 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19710 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19711 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19712 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19713 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19714 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19715 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19716 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19717 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19718 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19719 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19720 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19721 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19722 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19723 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19724 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19725 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19726 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19727 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19728 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19729 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19730 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19731 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19732 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19733 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19734 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19735 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19736 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19737 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19738 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19739 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19740 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19741 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19742 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19743 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19744 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19745 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19746 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19747 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19748 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19749 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19750 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19751 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19752 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19753 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19754 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19755 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19756 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19757 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19758 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19759 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19760 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19761 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19762 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19763 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19764 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19765 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19766 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19767 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19768 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19769 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19770 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19771 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19772 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19773 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19774 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19775 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19776 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19777 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19778 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19779 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19780 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19781 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19782 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19783 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19784 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19785 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19786 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19787 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19788 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19789 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19790 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19791 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19792 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19793 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19794 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19795 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19796 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19797 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19798 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19799 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19800 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19801 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19802 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19803 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19804 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19805 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19806 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19807 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19808 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19809 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19810 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19811 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19812 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19813 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19814 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19815 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19816 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19817 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19818 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19819 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19820 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19821 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655