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 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
131 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
132 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
133 X.509 version 3 if the certificate information includes X.509 extensions.
137 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
138 such as adding a trace facility for debugging certificate chain building.
142 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
143 in particular supporting requests for central key generation, generalized
144 polling, and various types of genm/genp exchanges defined in CMP Updates.
148 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
149 like correcting the TLS and proxy support and adding tracing for debugging.
153 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
157 * Fixed and extended `util/check-format.pl` for checking adherence to the
158 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
159 The checks are meanwhile more complete and yield fewer false positives.
163 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
164 sending and receiving multiple messages in a single call. An implementation
165 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
169 * The `SSL_CERT_PATH` and `SSL_CERT_URI` environment variables are introduced.
170 `SSL_CERT_URI` can be used to specify a URI for a root certificate store. The
171 `SSL_CERT_PATH` environment variable specifies a delimiter-separated list of
172 paths which are searched for root certificates.
174 The existing `SSL_CERT_DIR` environment variable is deprecated.
175 `SSL_CERT_DIR` was previously used to specify either a delimiter-separated
176 list of paths or an URI, which is ambiguous. Setting `SSL_CERT_PATH` causes
177 `SSL_CERT_DIR` to be ignored for the purposes of determining root certificate
178 directories, and setting `SSL_CERT_URI` causes `SSL_CERT_DIR` to be ignored
179 for the purposes of determining root certificate stores.
183 * Support for loading root certificates from the Windows certificate store
184 has been added. The support is in the form of a store which recognises the
185 URI string of `org.openssl.winstore://`. This store is enabled by default and
186 can be disabled using the new compile-time option `no-winstore`.
190 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
191 kernel versions that support KTLS have a known bug in CCM processing. That
192 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
193 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
198 * Zerocopy KTLS sendfile() support on Linux.
202 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
203 decryption as a protection against Bleichenbacher-like attacks.
204 The RSA decryption API will now return a randomly generated deterministic
205 message instead of an error in case it detects an error when checking
206 padding during PKCS#1 v1.5 decryption. This is a general protection against
207 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
209 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
210 on the RSA decryption context.
217 ### Changes between 3.0 and 3.1.0 [xx XXX xxxx]
219 * The FIPS provider includes a few non-approved algorithms for
220 backward compatibility purposes and the "fips=yes" property query
221 must be used for all algorithm fetches to ensure FIPS compliance.
223 The algorithms that are included but not approved are Triple DES and EdDSA.
227 * Added support for KMAC in KBKDF.
231 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
232 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
233 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
234 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
235 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
236 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
237 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
238 for legacy EC and SM2 keys is also changed similarly to honor the
239 equivalent conversion format flag as specified in the underlying
240 `EC_KEY` object being exported to a provider, when this function is
241 called through `EVP_PKEY_export()`.
245 * RNDR and RNDRRS support in provider functions to provide
246 random number generation for Arm CPUs (aarch64).
250 * s_client and s_server apps now explicitly say when the TLS version
251 does not include the renegotiation mechanism. This avoids confusion
252 between that scenario versus when the TLS version includes secure
253 renegotiation but the peer lacks support for it.
257 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
259 *Tomasz Kantecki, Andrey Matyukov*
261 * The various OBJ_* functions have been made thread safe.
265 * Parallel dual-prime 1536/2048-bit modular exponentiation for
266 AVX512_IFMA capable processors.
268 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
270 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
271 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
272 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
273 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
274 `OPENSSL_NO_DEPRECATED_3_1`.
276 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
277 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
278 definitions for these functions regardless of whether
279 `OPENSSL_NO_DEPRECATED_3_1` is defined.
281 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
282 functions regardless of whether they are using them. It is recommended that
283 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
287 * When generating safe-prime DH parameters set the recommended private key
288 length equivalent to minimum key lengths as in RFC 7919.
292 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
293 maximum size that is smaller or equal to the digest length to comply with
294 FIPS 186-4 section 5. This is implemented by a new option
295 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
296 `rsa_pss_saltlen` parameter, which is now the default. Signature
297 verification is not affected by this change and continues to work as before.
304 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
305 listed here are only a brief description.
306 The migration guide contains more detailed information related to new features,
307 breaking changes, and mappings for the large list of deprecated functions.
309 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
311 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
313 * Fixed two buffer overflows in punycode decoding functions.
315 A buffer overrun can be triggered in X.509 certificate verification,
316 specifically in name constraint checking. Note that this occurs after
317 certificate chain signature verification and requires either a CA to
318 have signed the malicious certificate or for the application to continue
319 certificate verification despite failure to construct a path to a trusted
322 In a TLS client, this can be triggered by connecting to a malicious
323 server. In a TLS server, this can be triggered if the server requests
324 client authentication and a malicious client connects.
326 An attacker can craft a malicious email address to overflow
327 an arbitrary number of bytes containing the `.` character (decimal 46)
328 on the stack. This buffer overflow could result in a crash (causing a
332 An attacker can craft a malicious email address to overflow four
333 attacker-controlled bytes on the stack. This buffer overflow could
334 result in a crash (causing a denial of service) or potentially remote code
335 execution depending on stack layout for any given platform/compiler.
340 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
341 parameters in OpenSSL code.
342 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
343 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
344 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
345 Using these invalid names may cause algorithms to use slower methods
346 that ignore the CRT parameters.
350 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
355 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
356 data to be signed before signing the certificate.
360 * Added RIPEMD160 to the default provider.
364 * Ensured that the key share group sent or accepted for the key exchange
365 is allowed for the protocol version.
369 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
371 * OpenSSL supports creating a custom cipher via the legacy
372 EVP_CIPHER_meth_new() function and associated function calls. This function
373 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
374 to use the new provider mechanism in order to implement custom ciphers.
376 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
377 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
378 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
379 and decryption initialisation functions). Instead of using the custom cipher
380 directly it incorrectly tries to fetch an equivalent cipher from the
381 available providers. An equivalent cipher is found based on the NID passed to
382 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
383 given cipher. However it is possible for an application to incorrectly pass
384 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
385 is used in this way the OpenSSL encryption/decryption initialisation function
386 will match the NULL cipher as being equivalent and will fetch this from the
387 available providers. This will succeed if the default provider has been
388 loaded (or if a third party provider has been loaded that offers this
389 cipher). Using the NULL cipher means that the plaintext is emitted as the
392 Applications are only affected by this issue if they call
393 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
394 encryption/decryption initialisation function. Applications that only use
395 SSL/TLS are not impacted by this issue.
400 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
405 * Fixed the linux-mips64 Configure target which was missing the
406 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
411 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
416 * Correctly handle a retransmitted ClientHello in DTLS
420 * Fixed detection of ktls support in cross-compile environment on Linux
424 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
429 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
430 report correct results in some cases
434 * Fix UWP builds by defining VirtualLock
438 * For known safe primes use the minimum key length according to RFC 7919.
439 Longer private key sizes unnecessarily raise the cycles needed to compute the
440 shared secret without any increase of the real security. This fixes a
441 regression from 1.1.1 where these shorter keys were generated for the known
446 * Added the loongarch64 target
450 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
451 only passed to the FIPS provider and not to the default or legacy provider.
455 * Fixed reported performance degradation on aarch64. Restored the
456 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
457 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
458 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
459 The new algorithm is still used for 32 bit targets.
463 * Added a missing header for memcmp that caused compilation failure on some
468 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
470 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
471 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
472 This issue makes the RSA implementation with 2048 bit private keys
473 incorrect on such machines and memory corruption will happen during
474 the computation. As a consequence of the memory corruption an attacker
475 may be able to trigger a remote code execution on the machine performing
478 SSL/TLS servers or other servers using 2048 bit RSA private keys running
479 on machines supporting AVX512IFMA instructions of the X86_64 architecture
480 are affected by this issue.
485 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
486 implementation would not encrypt the entirety of the data under some
487 circumstances. This could reveal sixteen bytes of data that was
488 preexisting in the memory that wasn't written. In the special case of
489 "in place" encryption, sixteen bytes of the plaintext would be revealed.
491 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
492 they are both unaffected.
495 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
497 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
499 * In addition to the c_rehash shell command injection identified in
500 CVE-2022-1292, further bugs where the c_rehash script does not
501 properly sanitise shell metacharacters to prevent command injection have been
504 When the CVE-2022-1292 was fixed it was not discovered that there
505 are other places in the script where the file names of certificates
506 being hashed were possibly passed to a command executed through the shell.
508 This script is distributed by some operating systems in a manner where
509 it is automatically executed. On such operating systems, an attacker
510 could execute arbitrary commands with the privileges of the script.
512 Use of the c_rehash script is considered obsolete and should be replaced
513 by the OpenSSL rehash command line tool.
516 *Daniel Fiala, Tomáš Mráz*
518 * Case insensitive string comparison no longer uses locales. It has instead
519 been directly implemented.
523 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
525 * Case insensitive string comparison is reimplemented via new locale-agnostic
526 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
527 comparison. The previous implementation had problems when the Turkish locale
532 * Fixed a bug in the c_rehash script which was not properly sanitising shell
533 metacharacters to prevent command injection. This script is distributed by
534 some operating systems in a manner where it is automatically executed. On
535 such operating systems, an attacker could execute arbitrary commands with the
536 privileges of the script.
538 Use of the c_rehash script is considered obsolete and should be replaced
539 by the OpenSSL rehash command line tool.
544 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
545 certificate on an OCSP response. The bug caused the function in the case
546 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
547 response (meaning a successful verification) even in the case where the
548 response signing certificate fails to verify.
550 It is anticipated that most users of `OCSP_basic_verify` will not use the
551 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
552 a negative value (indicating a fatal error) in the case of a certificate
553 verification failure. The normal expected return value in this case would be
556 This issue also impacts the command line OpenSSL "ocsp" application. When
557 verifying an ocsp response with the "-no_cert_checks" option the command line
558 application will report that the verification is successful even though it
559 has in fact failed. In this case the incorrect successful response will also
560 be accompanied by error messages showing the failure and contradicting the
561 apparently successful result.
566 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
567 AAD data as the MAC key. This made the MAC key trivially predictable.
569 An attacker could exploit this issue by performing a man-in-the-middle attack
570 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
571 that the modified data would still pass the MAC integrity check.
573 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
574 endpoint will always be rejected by the recipient and the connection will
575 fail at that point. Many application protocols require data to be sent from
576 the client to the server first. Therefore, in such a case, only an OpenSSL
577 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
579 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
580 sent in both directions. In this case both clients and servers could be
581 affected, regardless of the application protocol.
583 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
584 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
585 the handshake when using this ciphersuite.
587 The confidentiality of data is not impacted by this issue, i.e. an attacker
588 cannot decrypt data that has been encrypted using this ciphersuite - they can
591 In order for this attack to work both endpoints must legitimately negotiate
592 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
593 OpenSSL 3.0, and is not available within the default provider or the default
594 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
595 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
596 following must have occurred:
598 1) OpenSSL must have been compiled with the (non-default) compile time option
599 enable-weak-ssl-ciphers
601 2) OpenSSL must have had the legacy provider explicitly loaded (either
602 through application code or via configuration)
604 3) The ciphersuite must have been explicitly added to the ciphersuite list
606 4) The libssl security level must have been set to 0 (default is 1)
608 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
610 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
611 others that both endpoints have in common
616 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
617 occupied by the removed hash table entries.
619 This function is used when decoding certificates or keys. If a long lived
620 process periodically decodes certificates or keys its memory usage will
621 expand without bounds and the process might be terminated by the operating
622 system causing a denial of service. Also traversing the empty hash table
623 entries will take increasingly more time.
625 Typically such long lived processes might be TLS clients or TLS servers
626 configured to accept client certificate authentication.
629 *Hugo Landau, Aliaksei Levin*
631 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
632 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
633 statistics are no longer supported. For compatibility, these statistics are
634 still listed in the output but are now always reported as zero.
638 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
640 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
641 for non-prime moduli.
643 Internally this function is used when parsing certificates that contain
644 elliptic curve public keys in compressed form or explicit elliptic curve
645 parameters with a base point encoded in compressed form.
647 It is possible to trigger the infinite loop by crafting a certificate that
648 has invalid explicit curve parameters.
650 Since certificate parsing happens prior to verification of the certificate
651 signature, any process that parses an externally supplied certificate may thus
652 be subject to a denial of service attack. The infinite loop can also be
653 reached when parsing crafted private keys as they can contain explicit
654 elliptic curve parameters.
656 Thus vulnerable situations include:
658 - TLS clients consuming server certificates
659 - TLS servers consuming client certificates
660 - Hosting providers taking certificates or private keys from customers
661 - Certificate authorities parsing certification requests from subscribers
662 - Anything else which parses ASN.1 elliptic curve parameters
664 Also any other applications that use the BN_mod_sqrt() where the attacker
665 can control the parameter values are vulnerable to this DoS issue.
670 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
671 to the list of ciphersuites providing Perfect Forward Secrecy as
672 required by SECLEVEL >= 3.
674 *Dmitry Belyavskiy, Nicola Tuveri*
676 * Made the AES constant time code for no-asm configurations
677 optional due to the resulting 95% performance degradation.
678 The AES constant time code can be enabled, for no assembly
679 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
683 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
688 * The negative return value handling of the certificate verification callback
689 was reverted. The replacement is to set the verification retry state with
690 the SSL_set_retry_verify() function.
694 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
696 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
697 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
698 verify a certificate supplied by a server. That function may return a
699 negative return value to indicate an internal error (for example out of
700 memory). Such a negative return value is mishandled by OpenSSL and will cause
701 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
702 success and a subsequent call to SSL_get_error() to return the value
703 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
704 returned by OpenSSL if the application has previously called
705 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
706 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
707 totally unexpected and applications may not behave correctly as a result. The
708 exact behaviour will depend on the application but it could result in
709 crashes, infinite loops or other similar incorrect responses.
711 This issue is made more serious in combination with a separate bug in OpenSSL
712 3.0 that will cause X509_verify_cert() to indicate an internal error when
713 processing a certificate chain. This will occur where a certificate does not
714 include the Subject Alternative Name extension but where a Certificate
715 Authority has enforced name constraints. This issue can occur even with valid
721 * Corrected a few file name and file reference bugs in the build,
722 installation and setup scripts, which lead to installation verification
723 failures. Slightly enhanced the installation verification script.
727 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
732 * Fixed PVK encoder to properly query for the passphrase.
736 * Multiple fixes in the OSSL_HTTP API functions.
740 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
741 OSSL_PARAM_INTEGER data type and return error on negative numbers
742 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
743 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
747 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
751 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
755 * Multiple threading fixes.
759 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
763 * Allow fetching an operation from the provider that owns an unexportable key
764 as a fallback if that is still allowed by the property query.
768 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
770 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
775 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
776 S390X capability vector to zero. This simplifies testing of different code
777 paths on S390X architecture.
781 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
782 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
783 SP 800-38D". The communication will fail at this point.
787 * The EC_GROUP_clear_free() function is deprecated as there is nothing
788 confidential in EC_GROUP data.
792 * The byte order mark (BOM) character is ignored if encountered at the
793 beginning of a PEM-formatted file.
797 * Added CMS support for the Russian GOST algorithms.
801 * Due to move of the implementation of cryptographic operations
802 to the providers, validation of various operation parameters can
803 be postponed until the actual operation is executed where previously
804 it happened immediately when an operation parameter was set.
806 For example when setting an unsupported curve with
807 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
808 fail but later keygen operations with the EVP_PKEY_CTX will fail.
810 *OpenSSL team members and many third party contributors*
812 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
813 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
814 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
815 instead to retrieve these algorithms from a provider.
819 * On build targets where the multilib postfix is set in the build
820 configuration the libdir directory was changing based on whether
821 the lib directory with the multilib postfix exists on the system
822 or not. This unpredictable behavior was removed and eventual
823 multilib postfix is now always added to the default libdir. Use
824 `--libdir=lib` to override the libdir if adding the postfix is
829 * The triple DES key wrap functionality now conforms to RFC 3217 but is
830 no longer interoperable with OpenSSL 1.1.1.
834 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
835 function codes, this function can only cause problems for calling
840 * Add a configurable flag to output date formats as ISO 8601. Does not
841 change the default date format.
845 * Version of MSVC earlier than 1300 could get link warnings, which could
846 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
847 Support for this flag has been removed.
851 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
852 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
853 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
854 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
855 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
859 * The signatures of the functions to get and set options on SSL and
860 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
861 Some source code changes may be required.
865 * The public definitions of conf_method_st and conf_st have been
866 deprecated. They will be made opaque in a future release.
868 *Rich Salz and Tomáš Mráz*
870 * Client-initiated renegotiation is disabled by default. To allow it, use
871 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
872 flag, or the "ClientRenegotiation" config parameter as appropriate.
876 * Add "abspath" and "includedir" pragma's to config files, to prevent,
877 or modify relative pathname inclusion.
881 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
882 validated. Please consult the README-FIPS and
883 README-PROVIDERS files, as well as the migration guide.
885 *OpenSSL team members and many third party contributors*
887 * For the key types DH and DHX the allowed settable parameters are now different.
891 * The openssl commands that read keys, certificates, and CRLs now
892 automatically detect the PEM or DER format of the input files.
894 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
896 * Added enhanced PKCS#12 APIs which accept a library context.
900 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
904 * Added support for Kernel TLS (KTLS).
906 *Boris Pismenny, John Baldwin and Andrew Gallatin*
908 * Support for RFC 5746 secure renegotiation is now required by default for
909 SSL or TLS connections to succeed.
913 * The signature of the `copy` functional parameter of the
914 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
915 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
916 the signature of the `pub_decode` functional parameter of the
917 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
918 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
922 * The error return values from some control calls (ctrl) have changed.
926 * A public key check is now performed during EVP_PKEY_derive_set_peer().
930 * Many functions in the EVP_ namespace that are getters of values from
931 implementations or contexts were renamed to include get or get0 in their
932 names. Old names are provided as macro aliases for compatibility and
937 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
938 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
939 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
944 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
947 * The output from the command line applications may have minor
952 * The output from numerous "printing" may have minor changes.
956 * Windows thread synchronization uses read/write primitives (SRWLock) when
957 supported by the OS, otherwise CriticalSection continues to be used.
961 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
962 work on read only BIO source/sinks that do not support these functions.
963 This allows piping or redirection of a file BIO using stdin to be buffered
964 into memory. This is used internally in OSSL_DECODER_from_bio().
968 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
969 this function would return one of the values OSSL_STORE_INFO_NAME,
970 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
971 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
972 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
973 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
974 using this function should be amended to handle the changed return value.
978 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
979 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
980 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
981 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
982 contains more than one certificate identifier: This means that all
983 certificates referenced there MUST be part of the validation chain.
987 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
988 RC5, DESX and DES have been moved to the legacy provider.
992 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
993 RIPEMD-160 have been moved to the legacy provider.
997 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1002 * The deprecated functions EVP_PKEY_get0_RSA(),
1003 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1004 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1005 well as the similarly named "get1" functions behave differently in
1010 * A number of functions handling low-level keys or engines were deprecated
1011 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1012 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1013 EVP_PKEY_get0_siphash().
1017 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1018 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1019 will need to load the legacy crypto provider. This includes these PBE
1020 algorithms which use this KDF:
1021 - NID_pbeWithMD2AndDES_CBC
1022 - NID_pbeWithMD5AndDES_CBC
1023 - NID_pbeWithSHA1AndRC2_CBC
1024 - NID_pbeWithMD2AndRC2_CBC
1025 - NID_pbeWithMD5AndRC2_CBC
1026 - NID_pbeWithSHA1AndDES_CBC
1030 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1031 BIO_debug_callback() functions.
1035 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1036 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1040 * The RAND_METHOD APIs have been deprecated.
1044 * The SRP APIs have been deprecated.
1048 * Add a compile time option to prevent the caching of provider fetched
1049 algorithms. This is enabled by including the no-cached-fetch option
1050 at configuration time.
1054 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1055 count of PKCS12_DEFAULT_ITER.
1057 *Tomáš Mráz and Sahana Prasad*
1059 * The openssl speed command does not use low-level API calls anymore.
1063 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1066 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1068 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1072 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1073 providers may supply their own group implementations (using either the "key
1074 exchange" or the "key encapsulation" methods) which will automatically be
1075 detected and used by libssl.
1077 *Matt Caswell, Nicola Tuveri*
1079 * The undocumented function X509_certificate_type() has been deprecated;
1083 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1087 * Removed RSA padding mode for SSLv23 (which was only used for
1088 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1089 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1094 * Deprecated the obsolete X9.31 RSA key generation related functions.
1096 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1097 is not allowed to return a value > 1, this is no more taken as failure.
1099 *Viktor Dukhovni and David von Oheimb*
1101 * Deprecated the obsolete X9.31 RSA key generation related functions
1102 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1103 BN_X931_generate_prime_ex().
1107 * The default key generation method for the regular 2-prime RSA keys was
1108 changed to the FIPS 186-4 B.3.6 method.
1112 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1116 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1120 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1121 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1123 *Rich Salz, Richard Levitte, and David von Oheimb*
1125 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1129 * Deprecated `OCSP_parse_url()`.
1133 * Validation of SM2 keys has been separated from the validation of regular EC
1138 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1139 switches: a validation failure triggers an early exit, returning a failure
1140 exit status to the parent process.
1144 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1145 to ignore unknown ciphers.
1149 * The `-cipher-commands` and `-digest-commands` options
1150 of the command line utility `list` have been deprecated.
1151 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1155 * Added convenience functions for generating asymmetric key pairs:
1156 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1157 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1161 * All of the low-level EC_KEY functions have been deprecated.
1163 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1165 * Deprecated all the libcrypto and libssl error string loading
1170 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1171 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1176 * The `-crypt` option to the `passwd` command line tool has been removed.
1180 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1185 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1189 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1190 EVP_PKEY_get1_tls_encodedpoint().
1194 * The security callback, which can be customised by application code, supports
1195 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1196 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1200 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1201 interface. Their functionality remains unchanged.
1205 * Added new option for 'openssl list', '-providers', which will display the
1206 list of loaded providers, their names, version and status. It optionally
1207 displays their gettable parameters.
1211 * Removed EVP_PKEY_set_alias_type().
1215 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1216 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1220 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1221 parameters are still checked for type safety at compile time via helper
1226 * Remove the RAND_DRBG API
1228 *Paul Dale and Matthias St. Pierre*
1230 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1231 as well as actual hostnames.
1235 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1236 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1237 conversely, silently ignore DTLS protocol version bounds when configuring
1238 TLS-based contexts. The commands can be repeated to set bounds of both
1239 types. The same applies with the corresponding "min_protocol" and
1240 "max_protocol" command-line switches, in case some application uses both TLS
1243 SSL_CTX instances that are created for a fixed protocol version (e.g.
1244 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1245 attempts to apply bounds to these protocol versions would result in an
1246 error. Now only the "version-flexible" SSL_CTX instances are subject to
1247 limits in configuration files in command-line options.
1251 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1256 * Reworked the recorded ERR codes to make better space for system errors.
1257 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1258 given code is a system error (true) or an OpenSSL error (false).
1262 * Reworked the test perl framework to better allow parallel testing.
1264 *Nicola Tuveri and David von Oheimb*
1266 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1267 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1271 * 'Configure' has been changed to figure out the configuration target if
1272 none is given on the command line. Consequently, the 'config' script is
1273 now only a mere wrapper. All documentation is changed to only mention
1276 *Rich Salz and Richard Levitte*
1278 * Added a library context `OSSL_LIB_CTX` that applications as well as
1279 other libraries can use to form a separate context within which
1280 libcrypto operations are performed.
1284 * Added various `_ex` functions to the OpenSSL API that support using
1285 a non-default `OSSL_LIB_CTX`.
1289 * Handshake now fails if Extended Master Secret extension is dropped
1294 * Dropped interactive mode from the `openssl` program.
1298 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1300 *David von Oheimb and Shane Lontis*
1302 * Deprecated `EC_METHOD_get_field_type()`.
1306 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1307 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1308 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1312 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1316 * Add CAdES-BES signature verification support, mostly derived
1317 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1319 *Filipe Raimundo da Silva*
1321 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1325 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1326 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1330 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1334 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1335 EC_KEY_precompute_mult().
1339 * Deprecated EC_POINTs_mul().
1343 * Removed FIPS_mode() and FIPS_mode_set().
1347 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1351 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1352 EC_POINT_get_Jprojective_coordinates_GFp().
1356 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1357 arrays to be more easily constructed via a series of utility functions.
1358 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1359 the various push functions and finally convert to a passable OSSL_PARAM
1360 array using OSSL_PARAM_BLD_to_param().
1364 * The security strength of SHA1 and MD5 based signatures in TLS has been
1369 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1370 contain a provider side internal key.
1374 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1378 * Project text documents not yet having a proper file name extension
1379 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1380 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1384 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1385 have been converted to Markdown with the goal to produce documents
1386 which not only look pretty when viewed online in the browser, but
1387 remain well readable inside a plain text editor.
1389 To achieve this goal, a 'minimalistic' Markdown style has been applied
1390 which avoids formatting elements that interfere too much with the
1391 reading flow in the text file. For example, it
1393 * avoids [ATX headings][] and uses [setext headings][] instead
1394 (which works for `<h1>` and `<h2>` headings only).
1395 * avoids [inline links][] and uses [reference links][] instead.
1396 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1398 [ATX headings]: https://github.github.com/gfm/#atx-headings
1399 [setext headings]: https://github.github.com/gfm/#setext-headings
1400 [inline links]: https://github.github.com/gfm/#inline-link
1401 [reference links]: https://github.github.com/gfm/#reference-link
1402 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1403 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1405 *Matthias St. Pierre*
1407 * The test suite is changed to preserve results of each test recipe.
1408 A new directory test-runs/ with subdirectories named like the
1409 test recipes are created in the build tree for this purpose.
1413 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1414 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1415 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1417 *David von Oheimb, Martin Peylo*
1419 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1420 It supports arbitrary request and response content types, GET redirection,
1421 TLS, connections via HTTP(S) proxies, connections and exchange via
1422 user-defined BIOs (allowing implicit connections), persistent connections,
1423 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1424 The legacy OCSP-focused (and only partly documented) API
1425 is retained for backward compatibility, while most of it is deprecated.
1429 * Added `util/check-format.pl`, a tool for checking adherence to the
1430 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1431 The checks performed are incomplete and yield some false positives.
1432 Still the tool should be useful for detecting most typical glitches.
1436 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1437 If domain name resolution yields multiple IP addresses all of them are tried
1438 after `connect()` failures.
1442 * All of the low-level RSA functions have been deprecated.
1446 * X509 certificates signed using SHA1 are no longer allowed at security
1451 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1452 modified to use PKEY APIs. These commands are now in maintenance mode
1453 and no new features will be added to them.
1457 * The command line utility rsautl has been deprecated.
1461 * The command line utilities genrsa and rsa have been modified to use PKEY
1462 APIs. They now write PKCS#8 keys by default. These commands are now in
1463 maintenance mode and no new features will be added to them.
1467 * All of the low-level DH functions have been deprecated.
1469 *Paul Dale and Matt Caswell*
1471 * All of the low-level DSA functions have been deprecated.
1475 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1476 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1480 * Deprecated low-level ECDH and ECDSA functions.
1484 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1488 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1489 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1490 a new formulation to include all the things it can be used for,
1491 as well as words of caution.
1495 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1499 * All of the low-level HMAC functions have been deprecated.
1501 *Paul Dale and David von Oheimb*
1503 * Over two thousand fixes were made to the documentation, including:
1504 - Common options (such as -rand/-writerand, TLS version control, etc)
1505 were refactored and point to newly-enhanced descriptions in openssl.pod.
1506 - Added style conformance for all options (with help from Richard Levitte),
1507 documented all reported missing options, added a CI build to check
1508 that all options are documented and that no unimplemented options
1510 - Documented some internals, such as all use of environment variables.
1511 - Addressed all internal broken L<> references.
1515 * All of the low-level CMAC functions have been deprecated.
1519 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1520 functions have been deprecated.
1522 *Paul Dale and David von Oheimb*
1524 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1525 set of functions. The documentation mentioned negative values for some
1526 errors, but this was never the case, so the mention of negative values
1529 Code that followed the documentation and thereby check with something
1530 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1534 * All of the low-level cipher functions have been deprecated.
1536 *Matt Caswell and Paul Dale*
1538 * Removed include/openssl/opensslconf.h.in and replaced it with
1539 include/openssl/configuration.h.in, which differs in not including
1540 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1541 was added to include both.
1543 This allows internal hacks where one might need to modify the set
1544 of configured macros, for example this if deprecated symbols are
1545 still supposed to be available internally:
1547 #include <openssl/configuration.h>
1549 #undef OPENSSL_NO_DEPRECATED
1550 #define OPENSSL_SUPPRESS_DEPRECATED
1552 #include <openssl/macros.h>
1554 This should not be used by applications that use the exported
1555 symbols, as that will lead to linking errors.
1559 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1560 used in exponentiation with 512-bit moduli. No EC algorithms are
1561 affected. Analysis suggests that attacks against 2-prime RSA1024,
1562 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1563 difficult to perform and are not believed likely. Attacks against DH512
1564 are considered just feasible. However, for an attack the target would
1565 have to re-use the DH512 private key, which is not recommended anyway.
1566 Also applications directly using the low-level API BN_mod_exp may be
1567 affected if they use BN_FLG_CONSTTIME.
1572 * Most memory-debug features have been deprecated, and the functionality
1573 replaced with no-ops.
1577 * Added documentation for the STACK API.
1581 * Introduced a new method type and API, OSSL_ENCODER, to represent
1582 generic encoders. These do the same sort of job that PEM writers
1583 and d2i functions do, but with support for methods supplied by
1584 providers, and the possibility for providers to support other
1589 * Introduced a new method type and API, OSSL_DECODER, to represent
1590 generic decoders. These do the same sort of job that PEM readers
1591 and i2d functions do, but with support for methods supplied by
1592 providers, and the possibility for providers to support other
1597 * Added a .pragma directive to the syntax of configuration files, to
1598 allow varying behavior in a supported and predictable manner.
1599 Currently added pragma:
1603 This allows dollar signs to be a keyword character unless it's
1604 followed by a opening brace or parenthesis. This is useful for
1605 platforms where dollar signs are commonly used in names, such as
1606 volume names and system directory names on VMS.
1610 * Added functionality to create an EVP_PKEY from user data.
1614 * Change the interpretation of the '--api' configuration option to
1615 mean that this is a desired API compatibility level with no
1616 further meaning. The previous interpretation, that this would
1617 also mean to remove all deprecated symbols up to and including
1618 the given version, no requires that 'no-deprecated' is also used
1619 in the configuration.
1621 When building applications, the desired API compatibility level
1622 can be set with the OPENSSL_API_COMPAT macro like before. For
1623 API compatibility version below 3.0, the old style numerical
1624 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1625 For version 3.0 and on, the value is expected to be the decimal
1626 value calculated from the major and minor version like this:
1628 MAJOR * 10000 + MINOR * 100
1632 -DOPENSSL_API_COMPAT=30000 For 3.0
1633 -DOPENSSL_API_COMPAT=30200 For 3.2
1635 To hide declarations that are deprecated up to and including the
1636 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1637 given when building the application as well.
1641 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1642 access to certificate and CRL stores via URIs and OSSL_STORE
1645 This adds the following functions:
1647 - X509_LOOKUP_store()
1648 - X509_STORE_load_file()
1649 - X509_STORE_load_path()
1650 - X509_STORE_load_store()
1651 - SSL_add_store_cert_subjects_to_stack()
1652 - SSL_CTX_set_default_verify_store()
1653 - SSL_CTX_load_verify_file()
1654 - SSL_CTX_load_verify_dir()
1655 - SSL_CTX_load_verify_store()
1659 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1660 The presence of this system service is determined at run-time.
1664 * Added functionality to create an EVP_PKEY context based on data
1665 for methods from providers. This takes an algorithm name and a
1666 property query string and simply stores them, with the intent
1667 that any operation that uses this context will use those strings
1668 to fetch the needed methods implicitly, thereby making the port
1669 of application written for pre-3.0 OpenSSL easier.
1673 * The undocumented function NCONF_WIN32() has been deprecated; for
1674 conversion details see the HISTORY section of doc/man5/config.pod
1678 * Introduced the new functions EVP_DigestSignInit_ex() and
1679 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1680 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1681 pages for further details.
1685 * Over two thousand fixes were made to the documentation, including:
1686 adding missing command flags, better style conformance, documentation
1689 *Rich Salz, Richard Levitte*
1691 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1692 X25519, X448, Ed25519 and Ed448.
1696 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1701 * Deprecated the public definition of `ERR_STATE` as well as the function
1702 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1707 * Added ERR functionality to give callers access to the stored function
1708 names that have replaced the older function code based functions.
1710 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1711 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1712 ERR_peek_error_all() and ERR_peek_last_error_all().
1714 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1715 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1716 ERR_func_error_string().
1720 * Extended testing to be verbose for failing tests only. The make variables
1721 VERBOSE_FAILURE or VF can be used to enable this:
1723 $ make VF=1 test # Unix
1724 $ mms /macro=(VF=1) test ! OpenVMS
1725 $ nmake VF=1 test # Windows
1729 * Added the `-copy_extensions` option to the `x509` command for use with
1730 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1731 all extensions in the request are copied to the certificate or vice versa.
1733 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1735 * Added the `-copy_extensions` option to the `req` command for use with
1736 `-x509`. When given with the `copy` or `copyall` argument,
1737 all extensions in the certification request are copied to the certificate.
1741 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1742 they generate are by default RFC 5280 compliant in the following sense:
1743 There is a subjectKeyIdentifier extension with a hash value of the public key
1744 and for not self-signed certs there is an authorityKeyIdentifier extension
1745 with a keyIdentifier field or issuer information identifying the signing key.
1746 This is done unless some configuration overrides the new default behavior,
1747 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1751 * Added several checks to `X509_verify_cert()` according to requirements in
1752 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1753 (which may be done by using the CLI option `-x509_strict`):
1754 * The basicConstraints of CA certificates must be marked critical.
1755 * CA certificates must explicitly include the keyUsage extension.
1756 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1757 * The issuer name of any certificate must not be empty.
1758 * The subject name of CA certs, certs with keyUsage crlSign,
1759 and certs without subjectAlternativeName must not be empty.
1760 * If a subjectAlternativeName extension is given it must not be empty.
1761 * The signatureAlgorithm field and the cert signature must be consistent.
1762 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1763 must not be marked critical.
1764 * The authorityKeyIdentifier must be given for X.509v3 certs
1765 unless they are self-signed.
1766 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1770 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1771 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1775 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1776 used even when parsing explicit parameters, when loading a encoded key
1777 or calling `EC_GROUP_new_from_ecpkparameters()`/
1778 `EC_GROUP_new_from_ecparameters()`.
1779 This prevents bypass of security hardening and performance gains,
1780 especially for curves with specialized EC_METHODs.
1781 By default, if a key encoded with explicit parameters is loaded and later
1782 encoded, the output is still encoded with explicit parameters, even if
1783 internally a "named" EC_GROUP is used for computation.
1787 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1788 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1789 NULL. After this change, only the cofactor parameter can be NULL. It also
1790 does some minimal sanity checks on the passed order.
1795 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1796 An attack is simple, if the first CMS_recipientInfo is valid but the
1797 second CMS_recipientInfo is chosen ciphertext. If the second
1798 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1799 encryption key will be replaced by garbage, and the message cannot be
1800 decoded, but if the RSA decryption fails, the correct encryption key is
1801 used and the recipient will not notice the attack.
1802 As a work around for this potential attack the length of the decrypted
1803 key must be equal to the cipher default key length, in case the
1804 certificate is not given and all recipientInfo are tried out.
1805 The old behaviour can be re-enabled in the CMS code by setting the
1806 CMS_DEBUG_DECRYPT flag.
1810 * Early start up entropy quality from the DEVRANDOM seed source has been
1811 improved for older Linux systems. The RAND subsystem will wait for
1812 /dev/random to be producing output before seeding from /dev/urandom.
1813 The seeded state is stored for future library initialisations using
1814 a system global shared memory segment. The shared memory identifier
1815 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1816 the desired value. The default identifier is 114.
1820 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1821 when primes for RSA keys are computed.
1822 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1823 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1824 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1825 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1826 This avoids possible fingerprinting of newly generated RSA modules.
1830 * Correct the extended master secret constant on EBCDIC systems. Without this
1831 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1832 negotiate EMS will fail. Unfortunately this also means that TLS connections
1833 between EBCDIC systems with this fix, and EBCDIC systems without this
1834 fix will fail if they negotiate EMS.
1838 * Changed the library initialisation so that the config file is now loaded
1839 by default. This was already the case for libssl. It now occurs for both
1840 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1841 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1845 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1846 where the former acts as a replacement for `ERR_put_error()`, and the
1847 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1848 `ERR_raise_data()` adds more flexibility by taking a format string and
1849 an arbitrary number of arguments following it, to be processed with
1854 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1855 to check if a named provider is loaded and available. When called, it
1856 will also activate all fallback providers if such are still present.
1860 * Enforce a minimum DH modulus size of 512 bits.
1864 * Changed DH parameters to generate the order q subgroup instead of 2q.
1865 Previously generated DH parameters are still accepted by DH_check
1866 but DH_generate_key works around that by clearing bit 0 of the
1867 private key for those. This avoids leaking bit 0 of the private key.
1871 * Significantly reduce secure memory usage by the randomness pools.
1875 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1880 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1881 algorithms. An implementation of a key exchange algorithm can be obtained
1882 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1883 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1884 the older EVP_PKEY_derive_init() function. See the man pages for the new
1885 functions for further details.
1889 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1893 * Removed the function names from error messages and deprecated the
1898 * Removed NextStep support and the macro OPENSSL_UNISTD
1902 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1903 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1904 Also removed "export var as function" capability; we do not export
1905 variables, only functions.
1909 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1910 an error and 1 indicating success. In previous versions of OpenSSL this
1911 was a void type. If a key was set longer than the maximum possible this
1916 * Support SM2 signing and verification schemes with X509 certificate.
1920 * Use SHA256 as the default digest for TS query in the `ts` app.
1924 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1928 * Default cipher lists/suites are now available via a function, the
1929 #defines are deprecated.
1933 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1934 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1935 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1939 * Join the directories crypto/x509 and crypto/x509v3
1943 * Added command 'openssl kdf' that uses the EVP_KDF API.
1947 * Added command 'openssl mac' that uses the EVP_MAC API.
1951 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1952 as default directories. Also added the command 'openssl info'
1953 for scripting purposes.
1957 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1962 * Add prediction resistance to the DRBG reseeding process.
1966 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1967 mandated by IEEE Std 1619-2018.
1971 * Added newline escaping functionality to a filename when using openssl dgst.
1972 This output format is to replicate the output format found in the `*sum`
1973 checksum programs. This aims to preserve backward compatibility.
1975 *Matt Eaton, Richard Levitte, and Paul Dale*
1977 * Removed the heartbeat message in DTLS feature, as it has very
1978 little usage and doesn't seem to fulfill a valuable purpose.
1979 The configuration option is now deprecated.
1983 * Changed the output of 'openssl {digestname} < file' to display the
1984 digest name in its output.
1988 * Added a new generic trace API which provides support for enabling
1989 instrumentation through trace output.
1991 *Richard Levitte & Matthias St. Pierre*
1993 * Added build tests for C++. These are generated files that only do one
1994 thing, to include one public OpenSSL head file each. This tests that
1995 the public header files can be usefully included in a C++ application.
1997 This test isn't enabled by default. It can be enabled with the option
1998 'enable-buildtest-c++'.
2002 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2006 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2010 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2014 * Added KMAC to EVP_MAC.
2018 * Added property based algorithm implementation selection framework to
2023 * Added SCA hardening for modular field inversion in EC_GROUP through
2024 a new dedicated field_inv() pointer in EC_METHOD.
2025 This also addresses a leakage affecting conversions from projective
2026 to affine coordinates.
2028 *Billy Bob Brumley, Nicola Tuveri*
2030 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2031 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2032 those algorithms that were already supported through the EVP_PKEY API
2033 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2034 and scrypt are now wrappers that call EVP_KDF.
2038 * Build devcrypto engine as a dynamic engine.
2040 *Eneas U de Queiroz*
2042 * Add keyed BLAKE2 to EVP_MAC.
2046 * Fix a bug in the computation of the endpoint-pair shared secret used
2047 by DTLS over SCTP. This breaks interoperability with older versions
2048 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2049 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2050 interoperability with such broken implementations. However, enabling
2051 this switch breaks interoperability with correct implementations.
2053 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2054 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2058 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2062 * Changed the license to the Apache License v2.0.
2066 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2068 - Major releases (indicated by incrementing the MAJOR release number)
2069 may introduce incompatible API/ABI changes.
2070 - Minor releases (indicated by incrementing the MINOR release number)
2071 may introduce new features but retain API/ABI compatibility.
2072 - Patch releases (indicated by incrementing the PATCH number)
2073 are intended for bug fixes and other improvements of existing
2074 features only (like improving performance or adding documentation)
2075 and retain API/ABI compatibility.
2079 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2083 * Remove the 'dist' target and add a tarball building script. The
2084 'dist' target has fallen out of use, and it shouldn't be
2085 necessary to configure just to create a source distribution.
2089 * Recreate the OS390-Unix config target. It no longer relies on a
2090 special script like it did for OpenSSL pre-1.1.0.
2094 * Instead of having the source directories listed in Configure, add
2095 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2100 * Add GMAC to EVP_MAC.
2104 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2108 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2109 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2110 to facilitate the continued use of MACs through raw private keys in
2111 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2115 * Deprecate ECDH_KDF_X9_62().
2119 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2120 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2121 are retained for backwards compatibility.
2125 * AES-XTS mode now enforces that its two keys are different to mitigate
2126 the attacked described in "Efficient Instantiations of Tweakable
2127 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2128 Details of this attack can be obtained from:
2129 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2133 * Rename the object files, i.e. give them other names than in previous
2134 versions. Their names now include the name of the final product, as
2135 well as its type mnemonic (bin, lib, shlib).
2139 * Added new option for 'openssl list', '-objects', which will display the
2140 list of built in objects, i.e. OIDs with names.
2144 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2145 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2150 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2151 improves application performance by removing data copies and providing
2152 applications with zero-copy system calls such as sendfile and splice.
2156 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2160 * `PKCS12_parse` now maintains the order of the parsed certificates
2161 when outputting them via `*ca` (rather than reversing it).
2165 * Deprecated pthread fork support methods.
2169 * Added support for FFDHE key exchange in TLS 1.3.
2173 * Added a new concept for OpenSSL plugability: providers. This
2174 functionality is designed to replace the ENGINE API and ENGINE
2175 implementations, and to be much more dynamic, allowing provider
2176 authors to introduce new algorithms among other things, as long as
2177 there's an API that supports the algorithm type.
2179 With this concept comes a new core API for interaction between
2180 libcrypto and provider implementations. Public libcrypto functions
2181 that want to use providers do so through this core API.
2183 The main documentation for this core API is found in
2184 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2185 refer to other manuals describing the API specific for supported
2186 algorithm types (also called operations).
2193 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2195 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2197 * Avoid loading of a dynamic engine twice.
2201 * Prioritise DANE TLSA issuer certs over peer certs
2205 * Fixed random API for MacOS prior to 10.12
2207 These MacOS versions don't support the CommonCrypto APIs
2211 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2213 * Fixed an SM2 Decryption Buffer Overflow.
2215 In order to decrypt SM2 encrypted data an application is expected to
2216 call the API function EVP_PKEY_decrypt(). Typically an application will
2217 call this function twice. The first time, on entry, the "out" parameter
2218 can be NULL and, on exit, the "outlen" parameter is populated with the
2219 buffer size required to hold the decrypted plaintext. The application
2220 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2221 again, but this time passing a non-NULL value for the "out" parameter.
2223 A bug in the implementation of the SM2 decryption code means that the
2224 calculation of the buffer size required to hold the plaintext returned
2225 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2226 size required by the second call. This can lead to a buffer overflow
2227 when EVP_PKEY_decrypt() is called by the application a second time with
2228 a buffer that is too small.
2230 A malicious attacker who is able present SM2 content for decryption to
2231 an application could cause attacker chosen data to overflow the buffer
2232 by up to a maximum of 62 bytes altering the contents of other data held
2233 after the buffer, possibly changing application behaviour or causing
2234 the application to crash. The location of the buffer is application
2235 dependent but is typically heap allocated.
2240 * Fixed various read buffer overruns processing ASN.1 strings
2242 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2243 structure which contains a buffer holding the string data and a field
2244 holding the buffer length. This contrasts with normal C strings which
2245 are represented as a buffer for the string data which is terminated
2246 with a NUL (0) byte.
2248 Although not a strict requirement, ASN.1 strings that are parsed using
2249 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2250 well as any string whose value has been set with the ASN1_STRING_set()
2251 function will additionally NUL terminate the byte array in the
2252 ASN1_STRING structure.
2254 However, it is possible for applications to directly construct valid
2255 ASN1_STRING structures which do not NUL terminate the byte array by
2256 directly setting the "data" and "length" fields in the ASN1_STRING
2257 array. This can also happen by using the ASN1_STRING_set0() function.
2259 Numerous OpenSSL functions that print ASN.1 data have been found to
2260 assume that the ASN1_STRING byte array will be NUL terminated, even
2261 though this is not guaranteed for strings that have been directly
2262 constructed. Where an application requests an ASN.1 structure to be
2263 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2264 been directly constructed by the application without NUL terminating
2265 the "data" field, then a read buffer overrun can occur.
2267 The same thing can also occur during name constraints processing
2268 of certificates (for example if a certificate has been directly
2269 constructed by the application instead of loading it via the OpenSSL
2270 parsing functions, and the certificate contains non NUL terminated
2271 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2272 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2274 If a malicious actor can cause an application to directly construct an
2275 ASN1_STRING and then process it through one of the affected OpenSSL
2276 functions then this issue could be hit. This might result in a crash
2277 (causing a Denial of Service attack). It could also result in the
2278 disclosure of private memory contents (such as private keys, or
2279 sensitive plaintext).
2284 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2286 * Fixed a problem with verifying a certificate chain when using the
2287 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2288 the certificates present in a certificate chain. It is not set by default.
2290 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2291 the chain that have explicitly encoded elliptic curve parameters was added
2292 as an additional strict check.
2294 An error in the implementation of this check meant that the result of a
2295 previous check to confirm that certificates in the chain are valid CA
2296 certificates was overwritten. This effectively bypasses the check
2297 that non-CA certificates must not be able to issue other certificates.
2299 If a "purpose" has been configured then there is a subsequent opportunity
2300 for checks that the certificate is a valid CA. All of the named "purpose"
2301 values implemented in libcrypto perform this check. Therefore, where
2302 a purpose is set the certificate chain will still be rejected even when the
2303 strict flag has been used. A purpose is set by default in libssl client and
2304 server certificate verification routines, but it can be overridden or
2305 removed by an application.
2307 In order to be affected, an application must explicitly set the
2308 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2309 for the certificate verification or, in the case of TLS client or server
2310 applications, override the default purpose.
2315 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2316 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2317 renegotiation ClientHello omits the signature_algorithms extension (where it
2318 was present in the initial ClientHello), but includes a
2319 signature_algorithms_cert extension then a NULL pointer dereference will
2320 result, leading to a crash and a denial of service attack.
2322 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2323 (which is the default configuration). OpenSSL TLS clients are not impacted by
2327 *Peter Kästle and Samuel Sapalski*
2329 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2331 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2332 create a unique hash value based on the issuer and serial number data
2333 contained within an X509 certificate. However, it was failing to correctly
2334 handle any errors that may occur while parsing the issuer field (which might
2335 occur if the issuer field is maliciously constructed). This may subsequently
2336 result in a NULL pointer deref and a crash leading to a potential denial of
2342 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2343 padding mode to correctly check for rollback attacks. This is considered a
2344 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2349 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2350 functions. Previously they could overflow the output length argument in some
2351 cases where the input length is close to the maximum permissible length for
2352 an integer on the platform. In such cases the return value from the function
2353 call would be 1 (indicating success), but the output length value would be
2354 negative. This could cause applications to behave incorrectly or crash.
2359 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2360 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2361 could be exploited in a side channel attack to recover the password. Since
2362 the attack is local host only this is outside of the current OpenSSL
2363 threat model and therefore no CVE is assigned.
2365 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2370 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2372 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2373 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2374 If an attacker can control both items being compared then this could lead
2375 to a possible denial of service attack. OpenSSL itself uses the
2376 GENERAL_NAME_cmp function for two purposes:
2377 1) Comparing CRL distribution point names between an available CRL and a
2378 CRL distribution point embedded in an X509 certificate
2379 2) When verifying that a timestamp response token signer matches the
2380 timestamp authority name (exposed via the API functions
2381 TS_RESP_verify_response and TS_RESP_verify_token)
2386 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2388 * Certificates with explicit curve parameters are now disallowed in
2389 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2393 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2394 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2395 conversely, silently ignore DTLS protocol version bounds when configuring
2396 TLS-based contexts. The commands can be repeated to set bounds of both
2397 types. The same applies with the corresponding "min_protocol" and
2398 "max_protocol" command-line switches, in case some application uses both TLS
2401 SSL_CTX instances that are created for a fixed protocol version (e.g.
2402 TLSv1_server_method()) also silently ignore version bounds. Previously
2403 attempts to apply bounds to these protocol versions would result in an
2404 error. Now only the "version-flexible" SSL_CTX instances are subject to
2405 limits in configuration files in command-line options.
2409 * Handshake now fails if Extended Master Secret extension is dropped
2414 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2416 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2418 * Fixed segmentation fault in SSL_check_chain()
2419 Server or client applications that call the SSL_check_chain() function
2420 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2421 dereference as a result of incorrect handling of the
2422 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2423 or unrecognised signature algorithm is received from the peer. This could
2424 be exploited by a malicious peer in a Denial of Service attack.
2429 * Added AES consttime code for no-asm configurations
2430 an optional constant time support for AES was added
2431 when building openssl for no-asm.
2432 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2433 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2434 At this time this feature is by default disabled.
2435 It will be enabled by default in 3.0.
2439 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2441 * Revert the change of EOF detection while reading in libssl to avoid
2442 regressions in applications depending on the current way of reporting
2443 the EOF. As the existing method is not fully accurate the change to
2444 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2445 branch and will be present in the 3.0 release.
2449 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2450 when primes for RSA keys are computed.
2451 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2452 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2453 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2454 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2455 This avoids possible fingerprinting of newly generated RSA modules.
2459 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2461 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2462 while reading in libssl then we would report an error back to the
2463 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2464 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2465 therefore give a hint as to what went wrong.
2469 * Check that ed25519 and ed448 are allowed by the security level. Previously
2470 signature algorithms not using an MD were not being checked that they were
2471 allowed by the security level.
2475 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2476 was not quite right. The behaviour was not consistent between resumption
2477 and normal handshakes, and also not quite consistent with historical
2478 behaviour. The behaviour in various scenarios has been clarified and
2479 it has been updated to make it match historical behaviour as closely as
2484 * *[VMS only]* The header files that the VMS compilers include automatically,
2485 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2486 that the C++ compiler doesn't understand. This is a shortcoming in the
2487 compiler, but can be worked around with `__cplusplus` guards.
2489 C++ applications that use OpenSSL libraries must be compiled using the
2490 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2491 functions. Otherwise, only functions with symbols of less than 31
2492 characters can be used, as the linker will not be able to successfully
2493 resolve symbols with longer names.
2497 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2498 The presence of this system service is determined at run-time.
2502 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2507 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2509 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2510 number generator (RNG). This was intended to include protection in the
2511 event of a fork() system call in order to ensure that the parent and child
2512 processes did not share the same RNG state. However, this protection was not
2513 being used in the default case.
2515 A partial mitigation for this issue is that the output from a high
2516 precision timer is mixed into the RNG state so the likelihood of a parent
2517 and child process sharing state is significantly reduced.
2519 If an application already calls OPENSSL_init_crypto() explicitly using
2520 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2523 *Matthias St. Pierre*
2525 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2526 used even when parsing explicit parameters, when loading a encoded key
2527 or calling `EC_GROUP_new_from_ecpkparameters()`/
2528 `EC_GROUP_new_from_ecparameters()`.
2529 This prevents bypass of security hardening and performance gains,
2530 especially for curves with specialized EC_METHODs.
2531 By default, if a key encoded with explicit parameters is loaded and later
2532 encoded, the output is still encoded with explicit parameters, even if
2533 internally a "named" EC_GROUP is used for computation.
2537 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2538 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2539 NULL. After this change, only the cofactor parameter can be NULL. It also
2540 does some minimal sanity checks on the passed order.
2545 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2546 An attack is simple, if the first CMS_recipientInfo is valid but the
2547 second CMS_recipientInfo is chosen ciphertext. If the second
2548 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2549 encryption key will be replaced by garbage, and the message cannot be
2550 decoded, but if the RSA decryption fails, the correct encryption key is
2551 used and the recipient will not notice the attack.
2552 As a work around for this potential attack the length of the decrypted
2553 key must be equal to the cipher default key length, in case the
2554 certificate is not given and all recipientInfo are tried out.
2555 The old behaviour can be re-enabled in the CMS code by setting the
2556 CMS_DEBUG_DECRYPT flag.
2561 * Early start up entropy quality from the DEVRANDOM seed source has been
2562 improved for older Linux systems. The RAND subsystem will wait for
2563 /dev/random to be producing output before seeding from /dev/urandom.
2564 The seeded state is stored for future library initialisations using
2565 a system global shared memory segment. The shared memory identifier
2566 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2567 the desired value. The default identifier is 114.
2571 * Correct the extended master secret constant on EBCDIC systems. Without this
2572 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2573 negotiate EMS will fail. Unfortunately this also means that TLS connections
2574 between EBCDIC systems with this fix, and EBCDIC systems without this
2575 fix will fail if they negotiate EMS.
2579 * Use Windows installation paths in the mingw builds
2581 Mingw isn't a POSIX environment per se, which means that Windows
2582 paths should be used for installation.
2587 * Changed DH_check to accept parameters with order q and 2q subgroups.
2588 With order 2q subgroups the bit 0 of the private key is not secret
2589 but DH_generate_key works around that by clearing bit 0 of the
2590 private key for those. This avoids leaking bit 0 of the private key.
2594 * Significantly reduce secure memory usage by the randomness pools.
2598 * Revert the DEVRANDOM_WAIT feature for Linux systems
2600 The DEVRANDOM_WAIT feature added a select() call to wait for the
2601 /dev/random device to become readable before reading from the
2602 /dev/urandom device.
2604 It turned out that this change had negative side effects on
2605 performance which were not acceptable. After some discussion it
2606 was decided to revert this feature and leave it up to the OS
2607 resp. the platform maintainer to ensure a proper initialization
2608 during early boot time.
2610 *Matthias St. Pierre*
2612 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2614 * Add build tests for C++. These are generated files that only do one
2615 thing, to include one public OpenSSL head file each. This tests that
2616 the public header files can be usefully included in a C++ application.
2618 This test isn't enabled by default. It can be enabled with the option
2619 'enable-buildtest-c++'.
2623 * Enable SHA3 pre-hashing for ECDSA and DSA.
2627 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2628 This changes the size when using the `genpkey` command when no size is given.
2629 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2630 generation commands to use 2048 bits by default.
2634 * Reorganize the manual pages to consistently have RETURN VALUES,
2635 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2636 util/fix-doc-nits accordingly.
2638 *Paul Yang, Joshua Lock*
2640 * Add the missing accessor EVP_PKEY_get0_engine()
2644 * Have commands like `s_client` and `s_server` output the signature scheme
2645 along with other cipher suite parameters when debugging.
2649 * Make OPENSSL_config() error agnostic again.
2653 * Do the error handling in RSA decryption constant time.
2657 * Prevent over long nonces in ChaCha20-Poly1305.
2659 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2660 for every encryption operation. RFC 7539 specifies that the nonce value
2661 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2662 and front pads the nonce with 0 bytes if it is less than 12
2663 bytes. However it also incorrectly allows a nonce to be set of up to 16
2664 bytes. In this case only the last 12 bytes are significant and any
2665 additional leading bytes are ignored.
2667 It is a requirement of using this cipher that nonce values are
2668 unique. Messages encrypted using a reused nonce value are susceptible to
2669 serious confidentiality and integrity attacks. If an application changes
2670 the default nonce length to be longer than 12 bytes and then makes a
2671 change to the leading bytes of the nonce expecting the new value to be a
2672 new unique nonce then such an application could inadvertently encrypt
2673 messages with a reused nonce.
2675 Additionally the ignored bytes in a long nonce are not covered by the
2676 integrity guarantee of this cipher. Any application that relies on the
2677 integrity of these ignored leading bytes of a long nonce may be further
2678 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2679 is safe because no such use sets such a long nonce value. However user
2680 applications that use this cipher directly and set a non-default nonce
2681 length to be longer than 12 bytes may be vulnerable.
2683 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2689 * Add DEVRANDOM_WAIT feature for Linux systems
2691 On older Linux systems where the getrandom() system call is not available,
2692 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2693 Contrary to getrandom(), the /dev/urandom device will not block during
2694 early boot when the kernel CSPRNG has not been seeded yet.
2696 To mitigate this known weakness, use select() to wait for /dev/random to
2697 become readable before reading from /dev/urandom.
2699 * Ensure that SM2 only uses SM3 as digest algorithm
2703 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2705 * Change the info callback signals for the start and end of a post-handshake
2706 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2707 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2708 confused by this and assume that a TLSv1.2 renegotiation has started. This
2709 can break KeyUpdate handling. Instead we no longer signal the start and end
2710 of a post handshake message exchange (although the messages themselves are
2711 still signalled). This could break some applications that were expecting
2712 the old signals. However without this KeyUpdate is not usable for many
2717 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2719 * Timing vulnerability in DSA signature generation
2721 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2722 timing side channel attack. An attacker could use variations in the signing
2723 algorithm to recover the private key.
2725 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2730 * Timing vulnerability in ECDSA signature generation
2732 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2733 timing side channel attack. An attacker could use variations in the signing
2734 algorithm to recover the private key.
2736 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2741 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2742 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2743 of two gigabytes and the error handling improved.
2745 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2746 categorized as a normal bug, not a security issue, because the DRBG reseeds
2747 automatically and is fully functional even without additional randomness
2748 provided by the application.
2750 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2752 * Add a new ClientHello callback. Provides a callback interface that gives
2753 the application the ability to adjust the nascent SSL object at the
2754 earliest stage of ClientHello processing, immediately after extensions have
2755 been collected but before they have been processed. In particular, this
2756 callback can adjust the supported TLS versions in response to the contents
2761 * Add SM2 base algorithm support.
2765 * s390x assembly pack: add (improved) hardware-support for the following
2766 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2767 aes-cfb/cfb8, aes-ecb.
2771 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2772 parameter is no longer accepted, as it leads to a corrupt table. NULL
2773 pem_str is reserved for alias entries only.
2777 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2778 step for prime curves. The new implementation is based on formulae from
2779 differential addition-and-doubling in homogeneous projective coordinates
2780 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2781 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2782 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2783 to work in projective coordinates.
2785 *Billy Bob Brumley, Nicola Tuveri*
2787 * Change generating and checking of primes so that the error rate of not
2788 being prime depends on the intended use based on the size of the input.
2789 For larger primes this will result in more rounds of Miller-Rabin.
2790 The maximal error rate for primes with more than 1080 bits is lowered
2793 *Kurt Roeckx, Annie Yousar*
2795 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2799 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2800 moving between systems, and to avoid confusion when a Windows build is
2801 done with mingw vs with MSVC. For POSIX installs, there's still a
2802 symlink or copy named 'tsget' to avoid that confusion as well.
2806 * Revert blinding in ECDSA sign and instead make problematic addition
2807 length-invariant. Switch even to fixed-length Montgomery multiplication.
2811 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2812 step for binary curves. The new implementation is based on formulae from
2813 differential addition-and-doubling in mixed Lopez-Dahab projective
2814 coordinates, modified to independently blind the operands.
2816 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2818 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2819 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2820 EC_METHODs to implement their own specialized "ladder step", to take
2821 advantage of more favorable coordinate systems or more efficient
2822 differential addition-and-doubling algorithms.
2824 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2826 * Modified the random device based seed sources to keep the relevant
2827 file descriptors open rather than reopening them on each access.
2828 This allows such sources to operate in a chroot() jail without
2829 the associated device nodes being available. This behaviour can be
2830 controlled using RAND_keep_random_devices_open().
2834 * Numerous side-channel attack mitigations have been applied. This may have
2835 performance impacts for some algorithms for the benefit of improved
2836 security. Specific changes are noted in this change log by their respective
2841 * AIX shared library support overhaul. Switch to AIX "natural" way of
2842 handling shared libraries, which means collecting shared objects of
2843 different versions and bitnesses in one common archive. This allows to
2844 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2845 doesn't affect the way 3rd party applications are linked, only how
2846 multi-version installation is managed.
2850 * Make ec_group_do_inverse_ord() more robust and available to other
2851 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2852 mitigations are applied to the fallback BN_mod_inverse().
2853 When using this function rather than BN_mod_inverse() directly, new
2854 EC cryptosystem implementations are then safer-by-default.
2858 * Add coordinate blinding for EC_POINT and implement projective
2859 coordinate blinding for generic prime curves as a countermeasure to
2860 chosen point SCA attacks.
2862 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2864 * Add blinding to ECDSA and DSA signatures to protect against side channel
2865 attacks discovered by Keegan Ryan (NCC Group).
2869 * Enforce checking in the `pkeyutl` command to ensure that the input
2870 length does not exceed the maximum supported digest length when performing
2871 a sign, verify or verifyrecover operation.
2875 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2876 I/O in combination with something like select() or poll() will hang. This
2877 can be turned off again using SSL_CTX_clear_mode().
2878 Many applications do not properly handle non-application data records, and
2879 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2880 around the problems in those applications, but can also break some.
2881 It's recommended to read the manpages about SSL_read(), SSL_write(),
2882 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2883 SSL_CTX_set_read_ahead() again.
2887 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2888 now allow empty (zero character) pass phrases.
2892 * Apply blinding to binary field modular inversion and remove patent
2893 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2897 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2898 binary and prime elliptic curves.
2902 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2903 constant time fixed point multiplication.
2907 * Revise elliptic curve scalar multiplication with timing attack
2908 defenses: ec_wNAF_mul redirects to a constant time implementation
2909 when computing fixed point and variable point multiplication (which
2910 in OpenSSL are mostly used with secret scalars in keygen, sign,
2911 ECDH derive operations).
2912 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2915 * Updated CONTRIBUTING
2919 * Updated DRBG / RAND to request nonce and additional low entropy
2920 randomness from the system.
2922 *Matthias St. Pierre*
2924 * Updated 'openssl rehash' to use OpenSSL consistent default.
2928 * Moved the load of the ssl_conf module to libcrypto, which helps
2929 loading engines that libssl uses before libssl is initialised.
2933 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2937 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2939 *Ingo Schwarze, Rich Salz*
2941 * Added output of accepting IP address and port for 'openssl s_server'
2945 * Added a new API for TLSv1.3 ciphersuites:
2946 SSL_CTX_set_ciphersuites()
2947 SSL_set_ciphersuites()
2951 * Memory allocation failures consistently add an error to the error
2956 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2957 in libcrypto when run as setuid/setgid.
2961 * Load any config file by default when libssl is used.
2965 * Added new public header file <openssl/rand_drbg.h> and documentation
2966 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2968 *Matthias St. Pierre*
2970 * QNX support removed (cannot find contributors to get their approval
2971 for the license change).
2975 * TLSv1.3 replay protection for early data has been implemented. See the
2976 SSL_read_early_data() man page for further details.
2980 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2981 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2982 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2983 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2984 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2985 configuration has been separated out. See the ciphers man page or the
2986 SSL_CTX_set_ciphersuites() man page for more information.
2990 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2991 in responder mode now supports the new "-multi" option, which
2992 spawns the specified number of child processes to handle OCSP
2993 requests. The "-timeout" option now also limits the OCSP
2994 responder's patience to wait to receive the full client request
2995 on a newly accepted connection. Child processes are respawned
2996 as needed, and the CA index file is automatically reloaded
2997 when changed. This makes it possible to run the "ocsp" responder
2998 as a long-running service, making the OpenSSL CA somewhat more
2999 feature-complete. In this mode, most diagnostic messages logged
3000 after entering the event loop are logged via syslog(3) rather than
3005 * Added support for X448 and Ed448. Heavily based on original work by
3010 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3011 objects loaded. This adds the functions OSSL_STORE_expect() and
3012 OSSL_STORE_find() as well as needed tools to construct searches and
3013 get the search data out of them.
3017 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3018 version of OpenSSL should review their configuration settings to ensure
3019 that they are still appropriate for TLSv1.3. For further information see:
3020 <https://wiki.openssl.org/index.php/TLS1.3>
3024 * Grand redesign of the OpenSSL random generator
3026 The default RAND method now utilizes an AES-CTR DRBG according to
3027 NIST standard SP 800-90Ar1. The new random generator is essentially
3028 a port of the default random generator from the OpenSSL FIPS 2.0
3029 object module. It is a hybrid deterministic random bit generator
3030 using an AES-CTR bit stream and which seeds and reseeds itself
3031 automatically using trusted system entropy sources.
3033 Some of its new features are:
3034 - Support for multiple DRBG instances with seed chaining.
3035 - The default RAND method makes use of a DRBG.
3036 - There is a public and private DRBG instance.
3037 - The DRBG instances are fork-safe.
3038 - Keep all global DRBG instances on the secure heap if it is enabled.
3039 - The public and private DRBG instance are per thread for lock free
3042 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3044 * Changed Configure so it only says what it does and doesn't dump
3045 so much data. Instead, ./configdata.pm should be used as a script
3046 to display all sorts of configuration data.
3050 * Added processing of "make variables" to Configure.
3054 * Added SHA512/224 and SHA512/256 algorithm support.
3058 * The last traces of Netware support, first removed in 1.1.0, have
3063 * Get rid of Makefile.shared, and in the process, make the processing
3064 of certain files (rc.obj, or the .def/.map/.opt files produced from
3065 the ordinal files) more visible and hopefully easier to trace and
3066 debug (or make silent).
3070 * Make it possible to have environment variable assignments as
3071 arguments to config / Configure.
3075 * Add multi-prime RSA (RFC 8017) support.
3079 * Add SM3 implemented according to GB/T 32905-2016
3080 *Jack Lloyd <jack.lloyd@ribose.com>,*
3081 *Ronald Tse <ronald.tse@ribose.com>,*
3082 *Erick Borsboom <erick.borsboom@ribose.com>*
3084 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3085 as documented in RFC6066.
3086 Based on a patch from Tomasz Moń
3088 *Filipe Raimundo da Silva*
3090 * Add SM4 implemented according to GB/T 32907-2016.
3091 *Jack Lloyd <jack.lloyd@ribose.com>,*
3092 *Ronald Tse <ronald.tse@ribose.com>,*
3093 *Erick Borsboom <erick.borsboom@ribose.com>*
3095 * Reimplement -newreq-nodes and ERR_error_string_n; the
3096 original author does not agree with the license change.
3100 * Add ARIA AEAD TLS support.
3104 * Some macro definitions to support VS6 have been removed. Visual
3105 Studio 6 has not worked since 1.1.0
3109 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3110 without clearing the errors.
3114 * Add "atfork" functions. If building on a system that without
3115 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3116 requirements. The RAND facility now uses/requires this.
3124 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3125 not possible to disable entirely. However, it's still possible to
3126 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3129 To disable, configure with 'no-ui-console'. 'no-ui' is still
3130 possible to use as an alias. Check at compile time with the
3131 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3132 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3136 * Add a STORE module, which implements a uniform and URI based reader of
3137 stores that can contain keys, certificates, CRLs and numerous other
3138 objects. The main API is loosely based on a few stdio functions,
3139 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3140 OSSL_STORE_error and OSSL_STORE_close.
3141 The implementation uses backends called "loaders" to implement arbitrary
3142 URI schemes. There is one built in "loader" for the 'file' scheme.
3146 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3147 then adjusted to work on FreeBSD 8.4 as well.
3148 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3149 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3153 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3154 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3155 error code calls like this:
3157 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3159 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3160 that can be encoded in C. For the foreseeable future, this will only
3163 *Richard Levitte and Tim Hudson*
3165 * Removed BSD cryptodev engine.
3169 * Add a build target 'build_all_generated', to build all generated files
3170 and only that. This can be used to prepare everything that requires
3171 things like perl for a system that lacks perl and then move everything
3172 to that system and do the rest of the build there.
3176 * In the UI interface, make it possible to duplicate the user data. This
3177 can be used by engines that need to retain the data for a longer time
3178 than just the call where this user data is passed.
3182 * Ignore the '-named_curve auto' value for compatibility of applications
3185 *Tomáš Mráz <tmraz@fedoraproject.org>*
3187 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3188 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3189 alerts across multiple records (some of which could be empty). In practice
3190 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3191 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3192 support this at all. Supporting it adds significant complexity to the
3193 record layer, and its removal is unlikely to cause interoperability
3198 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3199 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3200 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3205 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3206 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3208 *Richard Levitte, Andy Polyakov*
3210 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3215 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3216 platform rather than 'mingw'.
3220 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3221 success if they are asked to add an object which already exists
3222 in the store. This change cascades to other functions which load
3223 certificates and CRLs.
3227 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3228 facilitate stack unwinding even from assembly subroutines.
3232 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3233 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3237 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3238 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3239 which is the minimum version we support.
3243 * Certificate time validation (X509_cmp_time) enforces stricter
3244 compliance with RFC 5280. Fractional seconds and timezone offsets
3245 are no longer allowed.
3249 * Add support for ARIA
3253 * s_client will now send the Server Name Indication (SNI) extension by
3254 default unless the new "-noservername" option is used. The server name is
3255 based on the host provided to the "-connect" option unless overridden by
3256 using "-servername".
3260 * Add support for SipHash
3264 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3265 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3266 prevent issues where no progress is being made and the peer continually
3267 sends unrecognised record types, using up resources processing them.
3271 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3272 using the algorithm defined in
3273 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3277 * Heartbeat support has been removed; the ABI is changed for now.
3279 *Richard Levitte, Rich Salz*
3281 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3285 * The RSA "null" method, which was partially supported to avoid patent
3286 issues, has been replaced to always returns NULL.
3293 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3295 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3296 used even when parsing explicit parameters, when loading a encoded key
3297 or calling `EC_GROUP_new_from_ecpkparameters()`/
3298 `EC_GROUP_new_from_ecparameters()`.
3299 This prevents bypass of security hardening and performance gains,
3300 especially for curves with specialized EC_METHODs.
3301 By default, if a key encoded with explicit parameters is loaded and later
3302 encoded, the output is still encoded with explicit parameters, even if
3303 internally a "named" EC_GROUP is used for computation.
3307 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3308 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3309 NULL. After this change, only the cofactor parameter can be NULL. It also
3310 does some minimal sanity checks on the passed order.
3315 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3316 An attack is simple, if the first CMS_recipientInfo is valid but the
3317 second CMS_recipientInfo is chosen ciphertext. If the second
3318 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3319 encryption key will be replaced by garbage, and the message cannot be
3320 decoded, but if the RSA decryption fails, the correct encryption key is
3321 used and the recipient will not notice the attack.
3322 As a work around for this potential attack the length of the decrypted
3323 key must be equal to the cipher default key length, in case the
3324 certificate is not given and all recipientInfo are tried out.
3325 The old behaviour can be re-enabled in the CMS code by setting the
3326 CMS_DEBUG_DECRYPT flag.
3331 * Use Windows installation paths in the mingw builds
3333 Mingw isn't a POSIX environment per se, which means that Windows
3334 paths should be used for installation.
3339 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3341 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3342 This changes the size when using the `genpkey` command when no size is given.
3343 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3344 generation commands to use 2048 bits by default.
3348 * Prevent over long nonces in ChaCha20-Poly1305.
3350 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3351 for every encryption operation. RFC 7539 specifies that the nonce value
3352 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3353 and front pads the nonce with 0 bytes if it is less than 12
3354 bytes. However it also incorrectly allows a nonce to be set of up to 16
3355 bytes. In this case only the last 12 bytes are significant and any
3356 additional leading bytes are ignored.
3358 It is a requirement of using this cipher that nonce values are
3359 unique. Messages encrypted using a reused nonce value are susceptible to
3360 serious confidentiality and integrity attacks. If an application changes
3361 the default nonce length to be longer than 12 bytes and then makes a
3362 change to the leading bytes of the nonce expecting the new value to be a
3363 new unique nonce then such an application could inadvertently encrypt
3364 messages with a reused nonce.
3366 Additionally the ignored bytes in a long nonce are not covered by the
3367 integrity guarantee of this cipher. Any application that relies on the
3368 integrity of these ignored leading bytes of a long nonce may be further
3369 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3370 is safe because no such use sets such a long nonce value. However user
3371 applications that use this cipher directly and set a non-default nonce
3372 length to be longer than 12 bytes may be vulnerable.
3374 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3380 * Added SCA hardening for modular field inversion in EC_GROUP through
3381 a new dedicated field_inv() pointer in EC_METHOD.
3382 This also addresses a leakage affecting conversions from projective
3383 to affine coordinates.
3385 *Billy Bob Brumley, Nicola Tuveri*
3387 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3388 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3392 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3396 * Remove the 'dist' target and add a tarball building script. The
3397 'dist' target has fallen out of use, and it shouldn't be
3398 necessary to configure just to create a source distribution.
3402 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3404 * Timing vulnerability in DSA signature generation
3406 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3407 timing side channel attack. An attacker could use variations in the signing
3408 algorithm to recover the private key.
3410 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3415 * Timing vulnerability in ECDSA signature generation
3417 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3418 timing side channel attack. An attacker could use variations in the signing
3419 algorithm to recover the private key.
3421 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3426 * Add coordinate blinding for EC_POINT and implement projective
3427 coordinate blinding for generic prime curves as a countermeasure to
3428 chosen point SCA attacks.
3430 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3432 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3434 * Client DoS due to large DH parameter
3436 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3437 malicious server can send a very large prime value to the client. This will
3438 cause the client to spend an unreasonably long period of time generating a
3439 key for this prime resulting in a hang until the client has finished. This
3440 could be exploited in a Denial Of Service attack.
3442 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3447 * Cache timing vulnerability in RSA Key Generation
3449 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3450 a cache timing side channel attack. An attacker with sufficient access to
3451 mount cache timing attacks during the RSA key generation process could
3452 recover the private key.
3454 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3455 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3460 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3461 parameter is no longer accepted, as it leads to a corrupt table. NULL
3462 pem_str is reserved for alias entries only.
3466 * Revert blinding in ECDSA sign and instead make problematic addition
3467 length-invariant. Switch even to fixed-length Montgomery multiplication.
3471 * Change generating and checking of primes so that the error rate of not
3472 being prime depends on the intended use based on the size of the input.
3473 For larger primes this will result in more rounds of Miller-Rabin.
3474 The maximal error rate for primes with more than 1080 bits is lowered
3477 *Kurt Roeckx, Annie Yousar*
3479 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3483 * Add blinding to ECDSA and DSA signatures to protect against side channel
3484 attacks discovered by Keegan Ryan (NCC Group).
3488 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3489 now allow empty (zero character) pass phrases.
3493 * Certificate time validation (X509_cmp_time) enforces stricter
3494 compliance with RFC 5280. Fractional seconds and timezone offsets
3495 are no longer allowed.
3499 * Fixed a text canonicalisation bug in CMS
3501 Where a CMS detached signature is used with text content the text goes
3502 through a canonicalisation process first prior to signing or verifying a
3503 signature. This process strips trailing space at the end of lines, converts
3504 line terminators to CRLF and removes additional trailing line terminators
3505 at the end of a file. A bug in the canonicalisation process meant that
3506 some characters, such as form-feed, were incorrectly treated as whitespace
3507 and removed. This is contrary to the specification (RFC5485). This fix
3508 could mean that detached text data signed with an earlier version of
3509 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3510 signed with a fixed OpenSSL may fail to verify with an earlier version of
3511 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3512 and use the "-binary" flag (for the "cms" command line application) or set
3513 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3517 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3519 * Constructed ASN.1 types with a recursive definition could exceed the stack
3521 Constructed ASN.1 types with a recursive definition (such as can be found
3522 in PKCS7) could eventually exceed the stack given malicious input with
3523 excessive recursion. This could result in a Denial Of Service attack. There
3524 are no such structures used within SSL/TLS that come from untrusted sources
3525 so this is considered safe.
3527 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3533 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3535 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3536 effectively reduced to only comparing the least significant bit of each
3537 byte. This allows an attacker to forge messages that would be considered as
3538 authenticated in an amount of tries lower than that guaranteed by the
3539 security claims of the scheme. The module can only be compiled by the
3540 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3542 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3548 * Add a build target 'build_all_generated', to build all generated files
3549 and only that. This can be used to prepare everything that requires
3550 things like perl for a system that lacks perl and then move everything
3551 to that system and do the rest of the build there.
3555 * Backport SSL_OP_NO_RENGOTIATION
3557 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3558 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3559 changes this is no longer possible in 1.1.0. Therefore, the new
3560 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3561 1.1.0 to provide equivalent functionality.
3563 Note that if an application built against 1.1.0h headers (or above) is run
3564 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3565 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3569 * Removed the OS390-Unix config target. It relied on a script that doesn't
3574 * rsaz_1024_mul_avx2 overflow bug on x86_64
3576 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3577 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3578 Analysis suggests that attacks against RSA and DSA as a result of this
3579 defect would be very difficult to perform and are not believed likely.
3580 Attacks against DH1024 are considered just feasible, because most of the
3581 work necessary to deduce information about a private key may be performed
3582 offline. The amount of resources required for such an attack would be
3583 significant. However, for an attack on TLS to be meaningful, the server
3584 would have to share the DH1024 private key among multiple clients, which is
3585 no longer an option since CVE-2016-0701.
3587 This only affects processors that support the AVX2 but not ADX extensions
3588 like Intel Haswell (4th generation).
3590 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3591 was originally found via the OSS-Fuzz project.
3596 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3598 * bn_sqrx8x_internal carry bug on x86_64
3600 There is a carry propagating bug in the x86_64 Montgomery squaring
3601 procedure. No EC algorithms are affected. Analysis suggests that attacks
3602 against RSA and DSA as a result of this defect would be very difficult to
3603 perform and are not believed likely. Attacks against DH are considered just
3604 feasible (although very difficult) because most of the work necessary to
3605 deduce information about a private key may be performed offline. The amount
3606 of resources required for such an attack would be very significant and
3607 likely only accessible to a limited number of attackers. An attacker would
3608 additionally need online access to an unpatched system using the target
3609 private key in a scenario with persistent DH parameters and a private
3610 key that is shared between multiple clients.
3612 This only affects processors that support the BMI1, BMI2 and ADX extensions
3613 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3615 This issue was reported to OpenSSL by the OSS-Fuzz project.
3620 * Malformed X.509 IPAddressFamily could cause OOB read
3622 If an X.509 certificate has a malformed IPAddressFamily extension,
3623 OpenSSL could do a one-byte buffer overread. The most likely result
3624 would be an erroneous display of the certificate in text format.
3626 This issue was reported to OpenSSL by the OSS-Fuzz project.
3631 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3633 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3634 platform rather than 'mingw'.
3638 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3639 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3640 which is the minimum version we support.
3644 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3646 * Encrypt-Then-Mac renegotiation crash
3648 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3649 negotiated where it was not in the original handshake (or vice-versa) then
3650 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3651 and servers are affected.
3653 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3658 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3660 * Truncated packet could crash via OOB read
3662 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3663 cipher is being used, then a truncated packet can cause that host to
3664 perform an out-of-bounds read, usually resulting in a crash.
3666 This issue was reported to OpenSSL by Robert Święcki of Google.
3671 * Bad (EC)DHE parameters cause a client crash
3673 If a malicious server supplies bad parameters for a DHE or ECDHE key
3674 exchange then this can result in the client attempting to dereference a
3675 NULL pointer leading to a client crash. This could be exploited in a Denial
3678 This issue was reported to OpenSSL by Guido Vranken.
3683 * BN_mod_exp may produce incorrect results on x86_64
3685 There is a carry propagating bug in the x86_64 Montgomery squaring
3686 procedure. No EC algorithms are affected. Analysis suggests that attacks
3687 against RSA and DSA as a result of this defect would be very difficult to
3688 perform and are not believed likely. Attacks against DH are considered just
3689 feasible (although very difficult) because most of the work necessary to
3690 deduce information about a private key may be performed offline. The amount
3691 of resources required for such an attack would be very significant and
3692 likely only accessible to a limited number of attackers. An attacker would
3693 additionally need online access to an unpatched system using the target
3694 private key in a scenario with persistent DH parameters and a private
3695 key that is shared between multiple clients. For example this can occur by
3696 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3697 similar to CVE-2015-3193 but must be treated as a separate problem.
3699 This issue was reported to OpenSSL by the OSS-Fuzz project.
3704 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3706 * ChaCha20/Poly1305 heap-buffer-overflow
3708 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3709 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3710 crash. This issue is not considered to be exploitable beyond a DoS.
3712 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3717 * CMS Null dereference
3719 Applications parsing invalid CMS structures can crash with a NULL pointer
3720 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3721 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3722 structure callback if an attempt is made to free certain invalid encodings.
3723 Only CHOICE structures using a callback which do not handle NULL value are
3726 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3731 * Montgomery multiplication may produce incorrect results
3733 There is a carry propagating bug in the Broadwell-specific Montgomery
3734 multiplication procedure that handles input lengths divisible by, but
3735 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3736 and DH private keys are impossible. This is because the subroutine in
3737 question is not used in operations with the private key itself and an input
3738 of the attacker's direct choice. Otherwise the bug can manifest itself as
3739 transient authentication and key negotiation failures or reproducible
3740 erroneous outcome of public-key operations with specially crafted input.
3741 Among EC algorithms only Brainpool P-512 curves are affected and one
3742 presumably can attack ECDH key negotiation. Impact was not analyzed in
3743 detail, because pre-requisites for attack are considered unlikely. Namely
3744 multiple clients have to choose the curve in question and the server has to
3745 share the private key among them, neither of which is default behaviour.
3746 Even then only clients that chose the curve will be affected.
3748 This issue was publicly reported as transient failures and was not
3749 initially recognized as a security issue. Thanks to Richard Morgan for
3750 providing reproducible case.
3755 * Removed automatic addition of RPATH in shared libraries and executables,
3756 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3760 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3762 * Fix Use After Free for large message sizes
3764 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3765 message larger than approx 16k is received then the underlying buffer to
3766 store the incoming message is reallocated and moved. Unfortunately a
3767 dangling pointer to the old location is left which results in an attempt to
3768 write to the previously freed location. This is likely to result in a
3769 crash, however it could potentially lead to execution of arbitrary code.
3771 This issue only affects OpenSSL 1.1.0a.
3773 This issue was reported to OpenSSL by Robert Święcki.
3778 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3780 * OCSP Status Request extension unbounded memory growth
3782 A malicious client can send an excessively large OCSP Status Request
3783 extension. If that client continually requests renegotiation, sending a
3784 large OCSP Status Request extension each time, then there will be unbounded
3785 memory growth on the server. This will eventually lead to a Denial Of
3786 Service attack through memory exhaustion. Servers with a default
3787 configuration are vulnerable even if they do not support OCSP. Builds using
3788 the "no-ocsp" build time option are not affected.
3790 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3795 * SSL_peek() hang on empty record
3797 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3798 sends an empty record. This could be exploited by a malicious peer in a
3799 Denial Of Service attack.
3801 This issue was reported to OpenSSL by Alex Gaynor.
3806 * Excessive allocation of memory in tls_get_message_header() and
3807 dtls1_preprocess_fragment()
3809 A (D)TLS message includes 3 bytes for its length in the header for the
3810 message. This would allow for messages up to 16Mb in length. Messages of
3811 this length are excessive and OpenSSL includes a check to ensure that a
3812 peer is sending reasonably sized messages in order to avoid too much memory
3813 being consumed to service a connection. A flaw in the logic of version
3814 1.1.0 means that memory for the message is allocated too early, prior to
3815 the excessive message length check. Due to way memory is allocated in
3816 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3817 to service a connection. This could lead to a Denial of Service through
3818 memory exhaustion. However, the excessive message length check still takes
3819 place, and this would cause the connection to immediately fail. Assuming
3820 that the application calls SSL_free() on the failed connection in a timely
3821 manner then the 21Mb of allocated memory will then be immediately freed
3822 again. Therefore, the excessive memory allocation will be transitory in
3823 nature. This then means that there is only a security impact if:
3825 1) The application does not call SSL_free() in a timely manner in the event
3826 that the connection fails
3828 2) The application is working in a constrained environment where there is
3829 very little free memory
3831 3) The attacker initiates multiple connection attempts such that there are
3832 multiple connections in a state where memory has been allocated for the
3833 connection; SSL_free() has not yet been called; and there is insufficient
3834 memory to service the multiple requests.
3836 Except in the instance of (1) above any Denial Of Service is likely to be
3837 transitory because as soon as the connection fails the memory is
3838 subsequently freed again in the SSL_free() call. However there is an
3839 increased risk during this period of application crashes due to the lack of
3840 memory - which would then mean a more serious Denial of Service.
3842 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3843 (CVE-2016-6307 and CVE-2016-6308)
3847 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3848 had to be removed. Primary reason is that vendor assembler can't
3849 assemble our modules with -KPIC flag. As result it, assembly
3850 support, was not even available as option. But its lack means
3851 lack of side-channel resistant code, which is incompatible with
3852 security by todays standards. Fortunately gcc is readily available
3853 prepackaged option, which we firmly point at...
3857 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3859 * Windows command-line tool supports UTF-8 opt-in option for arguments
3860 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3861 (to any value) allows Windows user to access PKCS#12 file generated
3862 with Windows CryptoAPI and protected with non-ASCII password, as well
3863 as files generated under UTF-8 locale on Linux also protected with
3868 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3869 have been disabled by default and removed from DEFAULT, just like RC4.
3870 See the RC4 item below to re-enable both.
3874 * The method for finding the storage location for the Windows RAND seed file
3875 has changed. First we check %RANDFILE%. If that is not set then we check
3876 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3877 all else fails we fall back to C:\.
3881 * The EVP_EncryptUpdate() function has had its return type changed from void
3882 to int. A return of 0 indicates and error while a return of 1 indicates
3887 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3888 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3889 off the constant time implementation for RSA, DSA and DH have been made
3890 no-ops and deprecated.
3894 * Windows RAND implementation was simplified to only get entropy by
3895 calling CryptGenRandom(). Various other RAND-related tickets
3898 *Joseph Wylie Yandle, Rich Salz*
3900 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3901 and `OPENSSL_LH_`, respectively. The old names are available
3902 with API compatibility. They new names are now completely documented.
3906 * Unify TYPE_up_ref(obj) methods signature.
3907 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3908 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3909 int (instead of void) like all others TYPE_up_ref() methods.
3910 So now these methods also check the return value of CRYPTO_atomic_add(),
3911 and the validity of object reference counter.
3913 *fdasilvayy@gmail.com*
3915 * With Windows Visual Studio builds, the .pdb files are installed
3916 alongside the installed libraries and executables. For a static
3917 library installation, ossl_static.pdb is the associate compiler
3918 generated .pdb file to be used when linking programs.
3922 * Remove openssl.spec. Packaging files belong with the packagers.
3926 * Automatic Darwin/OSX configuration has had a refresh, it will now
3927 recognise x86_64 architectures automatically. You can still decide
3928 to build for a different bitness with the environment variable
3929 KERNEL_BITS (can be 32 or 64), for example:
3931 KERNEL_BITS=32 ./config
3935 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3936 256 bit AES and HMAC with SHA256.
3940 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3944 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3948 * To enable users to have their own config files and build file templates,
3949 Configure looks in the directory indicated by the environment variable
3950 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3951 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3952 name and is used as is.
3956 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3957 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3958 X509_CERT_FILE_CTX was removed.
3962 * "shared" builds are now the default. To create only static libraries use
3963 the "no-shared" Configure option.
3967 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3968 All of these option have not worked for some while and are fundamental
3973 * Make various cleanup routines no-ops and mark them as deprecated. Most
3974 global cleanup functions are no longer required because they are handled
3975 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3976 Explicitly de-initing can cause problems (e.g. where a library that uses
3977 OpenSSL de-inits, but an application is still using it). The affected
3978 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3979 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3980 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3981 COMP_zlib_cleanup().
3985 * --strict-warnings no longer enables runtime debugging options
3986 such as REF_DEBUG. Instead, debug options are automatically
3987 enabled with '--debug' builds.
3989 *Andy Polyakov, Emilia Käsper*
3991 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3992 have been moved out of the public header files. New functions for managing
3993 these have been added.
3997 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3998 objects have been moved out of the public header files. New
3999 functions for managing these have been added.
4003 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4004 have been moved out of the public header files. New functions for managing
4005 these have been added.
4009 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4010 moved out of the public header files. New functions for managing these
4015 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4019 * Removed the mk1mf build scripts.
4023 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4024 it is always safe to #include a header now.
4028 * Removed the aged BC-32 config and all its supporting scripts
4032 * Removed support for Ultrix, Netware, and OS/2.
4036 * Add support for HKDF.
4038 *Alessandro Ghedini*
4040 * Add support for blake2b and blake2s
4044 * Added support for "pipelining". Ciphers that have the
4045 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4046 encryptions/decryptions simultaneously. There are currently no built-in
4047 ciphers with this property but the expectation is that engines will be able
4048 to offer it to significantly improve throughput. Support has been extended
4049 into libssl so that multiple records for a single connection can be
4050 processed in one go (for >=TLS 1.1).
4054 * Added the AFALG engine. This is an async capable engine which is able to
4055 offload work to the Linux kernel. In this initial version it only supports
4056 AES128-CBC. The kernel must be version 4.1.0 or greater.
4060 * OpenSSL now uses a new threading API. It is no longer necessary to
4061 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4062 are two supported threading models: pthreads and windows threads. It is
4063 also possible to configure OpenSSL at compile time for "no-threads". The
4064 old threading API should no longer be used. The functions have been
4065 replaced with "no-op" compatibility macros.
4067 *Alessandro Ghedini, Matt Caswell*
4069 * Modify behavior of ALPN to invoke callback after SNI/servername
4070 callback, such that updates to the SSL_CTX affect ALPN.
4074 * Add SSL_CIPHER queries for authentication and key-exchange.
4078 * Changes to the DEFAULT cipherlist:
4079 - Prefer (EC)DHE handshakes over plain RSA.
4080 - Prefer AEAD ciphers over legacy ciphers.
4081 - Prefer ECDSA over RSA when both certificates are available.
4082 - Prefer TLSv1.2 ciphers/PRF.
4083 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4088 * Change the ECC default curve list to be this, in order: x25519,
4089 secp256r1, secp521r1, secp384r1.
4093 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4094 disabled by default. They can be re-enabled using the
4095 enable-weak-ssl-ciphers option to Configure.
4099 * If the server has ALPN configured, but supports no protocols that the
4100 client advertises, send a fatal "no_application_protocol" alert.
4101 This behaviour is SHALL in RFC 7301, though it isn't universally
4102 implemented by other servers.
4106 * Add X25519 support.
4107 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4108 for public and private key encoding using the format documented in
4109 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4110 key generation and key derivation.
4112 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4117 * Deprecate SRP_VBASE_get_by_user.
4118 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4119 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4120 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4121 seed, even if the seed is configured.
4123 Users should use SRP_VBASE_get1_by_user instead. Note that in
4124 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4125 also that even though configuring the SRP seed attempts to hide
4126 invalid usernames by continuing the handshake with fake
4127 credentials, this behaviour is not constant time and no strong
4128 guarantees are made that the handshake is indistinguishable from
4129 that of a valid user.
4133 * Configuration change; it's now possible to build dynamic engines
4134 without having to build shared libraries and vice versa. This
4135 only applies to the engines in `engines/`, those in `crypto/engine/`
4136 will always be built into libcrypto (i.e. "static").
4138 Building dynamic engines is enabled by default; to disable, use
4139 the configuration option "disable-dynamic-engine".
4141 The only requirements for building dynamic engines are the
4142 presence of the DSO module and building with position independent
4143 code, so they will also automatically be disabled if configuring
4144 with "disable-dso" or "disable-pic".
4146 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4147 are also taken away from openssl/opensslconf.h, as they are
4152 * Configuration change; if there is a known flag to compile
4153 position independent code, it will always be applied on the
4154 libcrypto and libssl object files, and never on the application
4155 object files. This means other libraries that use routines from
4156 libcrypto / libssl can be made into shared libraries regardless
4157 of how OpenSSL was configured.
4159 If this isn't desirable, the configuration options "disable-pic"
4160 or "no-pic" can be used to disable the use of PIC. This will
4161 also disable building shared libraries and dynamic engines.
4165 * Removed JPAKE code. It was experimental and has no wide use.
4169 * The INSTALL_PREFIX Makefile variable has been renamed to
4170 DESTDIR. That makes for less confusion on what this variable
4171 is for. Also, the configuration option --install_prefix is
4176 * Heartbeat for TLS has been removed and is disabled by default
4177 for DTLS; configure with enable-heartbeats. Code that uses the
4178 old #define's might need to be updated.
4180 *Emilia Käsper, Rich Salz*
4182 * Rename REF_CHECK to REF_DEBUG.
4186 * New "unified" build system
4188 The "unified" build system is aimed to be a common system for all
4189 platforms we support. With it comes new support for VMS.
4191 This system builds supports building in a different directory tree
4192 than the source tree. It produces one Makefile (for unix family
4193 or lookalikes), or one descrip.mms (for VMS).
4195 The source of information to make the Makefile / descrip.mms is
4196 small files called 'build.info', holding the necessary
4197 information for each directory with source to compile, and a
4198 template in Configurations, like unix-Makefile.tmpl or
4201 With this change, the library names were also renamed on Windows
4202 and on VMS. They now have names that are closer to the standard
4203 on Unix, and include the major version number, and in certain
4204 cases, the architecture they are built for. See "Notes on shared
4205 libraries" in INSTALL.
4207 We rely heavily on the perl module Text::Template.
4211 * Added support for auto-initialisation and de-initialisation of the library.
4212 OpenSSL no longer requires explicit init or deinit routines to be called,
4213 except in certain circumstances. See the OPENSSL_init_crypto() and
4214 OPENSSL_init_ssl() man pages for further information.
4218 * The arguments to the DTLSv1_listen function have changed. Specifically the
4219 "peer" argument is now expected to be a BIO_ADDR object.
4221 * Rewrite of BIO networking library. The BIO library lacked consistent
4222 support of IPv6, and adding it required some more extensive
4223 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4224 which hold all types of addresses and chains of address information.
4225 It also introduces a new API, with functions like BIO_socket,
4226 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4227 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4228 have been adapted accordingly.
4232 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4237 * CRIME protection: disable compression by default, even if OpenSSL is
4238 compiled with zlib enabled. Applications can still enable compression
4239 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4240 using the SSL_CONF library to configure compression.
4244 * The signature of the session callback configured with
4245 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4246 was explicitly marked as `const unsigned char*` instead of
4251 * Always DPURIFY. Remove the use of uninitialized memory in the
4252 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4256 * Removed many obsolete configuration items, including
4257 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4258 MD2_CHAR, MD2_INT, MD2_LONG
4260 IDEA_SHORT, IDEA_LONG
4261 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4263 *Rich Salz, with advice from Andy Polyakov*
4265 * Many BN internals have been moved to an internal header file.
4267 *Rich Salz with help from Andy Polyakov*
4269 * Configuration and writing out the results from it has changed.
4270 Files such as Makefile include/openssl/opensslconf.h and are now
4271 produced through general templates, such as Makefile.in and
4272 crypto/opensslconf.h.in and some help from the perl module
4275 Also, the center of configuration information is no longer
4276 Makefile. Instead, Configure produces a perl module in
4277 configdata.pm which holds most of the config data (in the hash
4278 table %config), the target data that comes from the target
4279 configuration in one of the `Configurations/*.conf` files (in
4284 * To clarify their intended purposes, the Configure options
4285 --prefix and --openssldir change their semantics, and become more
4286 straightforward and less interdependent.
4288 --prefix shall be used exclusively to give the location INSTALLTOP
4289 where programs, scripts, libraries, include files and manuals are
4290 going to be installed. The default is now /usr/local.
4292 --openssldir shall be used exclusively to give the default
4293 location OPENSSLDIR where certificates, private keys, CRLs are
4294 managed. This is also where the default openssl.cnf gets
4296 If the directory given with this option is a relative path, the
4297 values of both the --prefix value and the --openssldir value will
4298 be combined to become OPENSSLDIR.
4299 The default for --openssldir is INSTALLTOP/ssl.
4301 Anyone who uses --openssldir to specify where OpenSSL is to be
4302 installed MUST change to use --prefix instead.
4306 * The GOST engine was out of date and therefore it has been removed. An up
4307 to date GOST engine is now being maintained in an external repository.
4308 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4309 support for GOST ciphersuites (these are only activated if a GOST engine
4314 * EGD is no longer supported by default; use enable-egd when
4317 *Ben Kaduk and Rich Salz*
4319 * The distribution now has Makefile.in files, which are used to
4320 create Makefile's when Configure is run. *Configure must be run
4321 before trying to build now.*
4325 * The return value for SSL_CIPHER_description() for error conditions
4330 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4332 Obtaining and performing DNSSEC validation of TLSA records is
4333 the application's responsibility. The application provides
4334 the TLSA records of its choice to OpenSSL, and these are then
4335 used to authenticate the peer.
4337 The TLSA records need not even come from DNS. They can, for
4338 example, be used to implement local end-entity certificate or
4339 trust-anchor "pinning", where the "pin" data takes the form
4340 of TLSA records, which can augment or replace verification
4341 based on the usual WebPKI public certification authorities.
4345 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4346 continues to support deprecated interfaces in default builds.
4347 However, applications are strongly advised to compile their
4348 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4349 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4350 or the 1.1.0 releases.
4352 In environments in which all applications have been ported to
4353 not use any deprecated interfaces OpenSSL's Configure script
4354 should be used with the --api=1.1.0 option to entirely remove
4355 support for the deprecated features from the library and
4356 unconditionally disable them in the installed headers.
4357 Essentially the same effect can be achieved with the "no-deprecated"
4358 argument to Configure, except that this will always restrict
4359 the build to just the latest API, rather than a fixed API
4362 As applications are ported to future revisions of the API,
4363 they should update their compile-time OPENSSL_API_COMPAT define
4364 accordingly, but in most cases should be able to continue to
4365 compile with later releases.
4367 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4368 0x10000000L and 0x00908000L, respectively. However those
4369 versions did not support the OPENSSL_API_COMPAT feature, and
4370 so applications are not typically tested for explicit support
4371 of just the undeprecated features of either release.
4375 * Add support for setting the minimum and maximum supported protocol.
4376 It can bet set via the SSL_set_min_proto_version() and
4377 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4378 MaxProtocol. It's recommended to use the new APIs to disable
4379 protocols instead of disabling individual protocols using
4380 SSL_set_options() or SSL_CONF's Protocol. This change also
4381 removes support for disabling TLS 1.2 in the OpenSSL TLS
4382 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4386 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4390 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4391 and integrates ECDSA and ECDH functionality into EC. Implementations can
4392 now redirect key generation and no longer need to convert to or from
4395 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4396 include the ec.h header file instead.
4400 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4401 ciphers who are no longer supported and drops support the ephemeral RSA key
4402 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4406 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4407 opaque. For HMAC_CTX, the following constructors and destructors
4410 HMAC_CTX *HMAC_CTX_new(void);
4411 void HMAC_CTX_free(HMAC_CTX *ctx);
4413 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4414 destroy such methods has been added. See EVP_MD_meth_new(3) and
4415 EVP_CIPHER_meth_new(3) for documentation.
4418 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4419 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4420 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4421 an already created structure.
4422 2) For consistency with the majority of our object creators and
4423 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4424 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4425 for deprecated builds.
4429 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4430 cryptographic operations to be performed asynchronously as long as an
4431 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4432 further details. Libssl has also had this capability integrated with the
4433 introduction of the new mode SSL_MODE_ASYNC and associated error
4434 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4435 pages. This work was developed in partnership with Intel Corp.
4439 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4440 always enabled now. If you want to disable the support you should
4441 exclude it using the list of supported ciphers. This also means that the
4442 "-no_ecdhe" option has been removed from s_server.
4446 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4447 SSL_{CTX_}set1_curves() which can set a list.
4451 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4452 curve you want to support using SSL_{CTX_}set1_curves().
4456 * State machine rewrite. The state machine code has been significantly
4457 refactored in order to remove much duplication of code and solve issues
4458 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4459 further details). This change does have some associated API changes.
4460 Notably the SSL_state() function has been removed and replaced by
4461 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4462 SSL_set_state() has been removed altogether. The previous handshake states
4463 defined in ssl.h and ssl3.h have also been removed.
4467 * All instances of the string "ssleay" in the public API were replaced
4468 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4469 Some error codes related to internal RSA_eay API's were renamed.
4473 * The demo files in crypto/threads were moved to demo/threads.
4477 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4480 *Matt Caswell, Rich Salz*
4482 * New ASN.1 embed macro.
4484 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4485 structure is not allocated: it is part of the parent. That is instead of
4493 This reduces memory fragmentation and make it impossible to accidentally
4494 set a mandatory field to NULL.
4496 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4497 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4498 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4503 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4507 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4508 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4509 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4510 DES and RC4 ciphersuites.
4514 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4515 This changes the decoding behaviour for some invalid messages,
4516 though the change is mostly in the more lenient direction, and
4517 legacy behaviour is preserved as much as possible.
4521 * Fix no-stdio build.
4522 *David Woodhouse <David.Woodhouse@intel.com> and also*
4523 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4525 * New testing framework
4526 The testing framework has been largely rewritten and is now using
4527 perl and the perl modules Test::Harness and an extended variant of
4528 Test::More called OpenSSL::Test to do its work. All test scripts in
4529 test/ have been rewritten into test recipes, and all direct calls to
4530 executables in test/Makefile have become individual recipes using the
4531 simplified testing OpenSSL::Test::Simple.
4533 For documentation on our testing modules, do:
4535 perldoc test/testlib/OpenSSL/Test/Simple.pm
4536 perldoc test/testlib/OpenSSL/Test.pm
4540 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4541 are used; the latter aborts on memory leaks (usually checked on exit).
4542 Some undocumented "set malloc, etc., hooks" functions were removed
4543 and others were changed. All are now documented.
4547 * In DSA_generate_parameters_ex, if the provided seed is too short,
4550 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4552 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4553 from RFC4279, RFC4785, RFC5487, RFC5489.
4555 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4556 original RSA_PSK patch.
4560 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4561 era flag was never set throughout the codebase (only read). Also removed
4562 SSL3_FLAGS_POP_BUFFER which was only used if
4563 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4567 * Changed the default name options in the "ca", "crl", "req" and "x509"
4568 to be "oneline" instead of "compat".
4572 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4573 not aware of clients that still exhibit this bug, and the workaround
4574 hasn't been working properly for a while.
4578 * The return type of BIO_number_read() and BIO_number_written() as well as
4579 the corresponding num_read and num_write members in the BIO structure has
4580 changed from unsigned long to uint64_t. On platforms where an unsigned
4581 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4586 * Given the pervasive nature of TLS extensions it is inadvisable to run
4587 OpenSSL without support for them. It also means that maintaining
4588 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4589 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4593 * Removed support for the two export grade static DH ciphersuites
4594 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4595 were newly added (along with a number of other static DH ciphersuites) to
4596 1.0.2. However the two export ones have *never* worked since they were
4597 introduced. It seems strange in any case to be adding new export
4598 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4602 * Version negotiation has been rewritten. In particular SSLv23_method(),
4603 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4604 and turned into macros which simply call the new preferred function names
4605 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4606 should use the new names instead. Also as part of this change the ssl23.h
4607 header file has been removed.
4611 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4612 code and the associated standard is no longer considered fit-for-purpose.
4616 * RT2547 was closed. When generating a private key, try to make the
4617 output file readable only by the owner. This behavior change might
4618 be noticeable when interacting with other software.
4620 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4625 * Added HTTP GET support to the ocsp command.
4629 * Changed default digest for the dgst and enc commands from MD5 to
4634 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4638 * Added support for TLS extended master secret from
4639 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4640 initial patch which was a great help during development.
4644 * All libssl internal structures have been removed from the public header
4645 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4646 now redundant). Users should not attempt to access internal structures
4647 directly. Instead they should use the provided API functions.
4651 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4652 Access to deprecated functions can be re-enabled by running config with
4653 "enable-deprecated". In addition applications wishing to use deprecated
4654 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4655 will, by default, disable some transitive includes that previously existed
4656 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4660 * Added support for OCB mode. OpenSSL has been granted a patent license
4661 compatible with the OpenSSL license for use of OCB. Details are available
4662 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4663 for OCB can be removed by calling config with no-ocb.
4667 * SSLv2 support has been removed. It still supports receiving an SSLv2
4668 compatible client hello.
4672 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4673 done while fixing the error code for the key-too-small case.
4675 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4677 * CA.sh has been removed; use CA.pl instead.
4681 * Removed old DES API.
4685 * Remove various unsupported platforms:
4691 Sinix/ReliantUNIX RM400
4696 16-bit platforms such as WIN16
4700 * Clean up OPENSSL_NO_xxx #define's
4701 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4702 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4703 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4704 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4705 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4706 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4707 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4708 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4709 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4710 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4714 * Cleaned up dead code
4715 Remove all but one '#ifdef undef' which is to be looked at.
4719 * Clean up calling of xxx_free routines.
4720 Just like free(), fix most of the xxx_free routines to accept
4721 NULL. Remove the non-null checks from callers. Save much code.
4725 * Add secure heap for storage of private keys (when possible).
4726 Add BIO_s_secmem(), CBIGNUM, etc.
4727 Contributed by Akamai Technologies under our Corporate CLA.
4731 * Experimental support for a new, fast, unbiased prime candidate generator,
4732 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4734 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4736 * New output format NSS in the sess_id command line tool. This allows
4737 exporting the session id and the master key in NSS keylog format.
4739 *Martin Kaiser <martin@kaiser.cx>*
4741 * Harmonize version and its documentation. -f flag is used to display
4744 *mancha <mancha1@zoho.com>*
4746 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4747 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4749 *mancha <mancha1@zoho.com>*
4751 * Fix some double frees. These are not thought to be exploitable.
4753 *mancha <mancha1@zoho.com>*
4755 * A missing bounds check in the handling of the TLS heartbeat extension
4756 can be used to reveal up to 64k of memory to a connected client or
4759 Thanks for Neel Mehta of Google Security for discovering this bug and to
4760 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4761 preparing the fix ([CVE-2014-0160])
4763 *Adam Langley, Bodo Moeller*
4765 * Fix for the attack described in the paper "Recovering OpenSSL
4766 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4767 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4768 <http://eprint.iacr.org/2014/140>
4770 Thanks to Yuval Yarom and Naomi Benger for discovering this
4771 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4773 *Yuval Yarom and Naomi Benger*
4775 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4776 this fixes a limitation in previous versions of OpenSSL.
4780 * Experimental encrypt-then-mac support.
4782 Experimental support for encrypt then mac from
4783 draft-gutmann-tls-encrypt-then-mac-02.txt
4785 To enable it set the appropriate extension number (0x42 for the test
4786 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4788 For non-compliant peers (i.e. just about everything) this should have no
4791 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4795 * Add EVP support for key wrapping algorithms, to avoid problems with
4796 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4797 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4798 algorithms and include tests cases.
4802 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4807 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4808 MGF1 digest and OAEP label.
4812 * Make openssl verify return errors.
4814 *Chris Palmer <palmer@google.com> and Ben Laurie*
4816 * New function ASN1_TIME_diff to calculate the difference between two
4817 ASN1_TIME structures or one structure and the current time.
4821 * Update fips_test_suite to support multiple command line options. New
4822 test to induce all self test errors in sequence and check expected
4827 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4828 sign or verify all in one operation.
4832 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4833 test programs and fips_test_suite. Includes functionality to parse
4834 the minimal script output of fipsalgest.pl directly.
4838 * Add authorisation parameter to FIPS_module_mode_set().
4842 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4846 * Use separate DRBG fields for internal and external flags. New function
4847 FIPS_drbg_health_check() to perform on demand health checking. Add
4848 generation tests to fips_test_suite with reduced health check interval to
4849 demonstrate periodic health checking. Add "nodh" option to
4850 fips_test_suite to skip very slow DH test.
4854 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4859 * More extensive health check for DRBG checking many more failure modes.
4860 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4861 combination: call this in fips_test_suite.
4865 * Add support for canonical generation of DSA parameter 'g'. See
4868 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4869 POST to handle HMAC cases.
4873 * Add functions FIPS_module_version() and FIPS_module_version_text()
4874 to return numerical and string versions of the FIPS module number.
4878 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4879 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4880 outside the validated module in the FIPS capable OpenSSL.
4884 * Minor change to DRBG entropy callback semantics. In some cases
4885 there is no multiple of the block length between min_len and
4886 max_len. Allow the callback to return more than max_len bytes
4887 of entropy but discard any extra: it is the callback's responsibility
4888 to ensure that the extra data discarded does not impact the
4889 requested amount of entropy.
4893 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4894 information in FIPS186-3, SP800-57 and SP800-131A.
4898 * CCM support via EVP. Interface is very similar to GCM case except we
4899 must supply all data in one chunk (i.e. no update, final) and the
4900 message length must be supplied if AAD is used. Add algorithm test
4905 * Initial version of POST overhaul. Add POST callback to allow the status
4906 of POST to be monitored and/or failures induced. Modify fips_test_suite
4907 to use callback. Always run all selftests even if one fails.
4911 * XTS support including algorithm test driver in the fips_gcmtest program.
4912 Note: this does increase the maximum key length from 32 to 64 bytes but
4913 there should be no binary compatibility issues as existing applications
4914 will never use XTS mode.
4918 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4919 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4920 performs algorithm blocking for unapproved PRNG types. Also do not
4921 set PRNG type in FIPS_mode_set(): leave this to the application.
4922 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4923 the standard OpenSSL PRNG: set additional data to a date time vector.
4927 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4928 This shouldn't present any incompatibility problems because applications
4929 shouldn't be using these directly and any that are will need to rethink
4930 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4934 * Extensive self tests and health checking required by SP800-90 DRBG.
4935 Remove strength parameter from FIPS_drbg_instantiate and always
4936 instantiate at maximum supported strength.
4940 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4944 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4948 * New function DH_compute_key_padded() to compute a DH key and pad with
4949 leading zeroes if needed: this complies with SP800-56A et al.
4953 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4954 anything, incomplete, subject to change and largely untested at present.
4958 * Modify fipscanisteronly build option to only build the necessary object
4959 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4963 * Add experimental option FIPSSYMS to give all symbols in
4964 fipscanister.o and FIPS or fips prefix. This will avoid
4965 conflicts with future versions of OpenSSL. Add perl script
4966 util/fipsas.pl to preprocess assembly language source files
4967 and rename any affected symbols.
4971 * Add selftest checks and algorithm block of non-fips algorithms in
4972 FIPS mode. Remove DES2 from selftests.
4976 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4977 return internal method without any ENGINE dependencies. Add new
4978 tiny fips sign and verify functions.
4982 * New build option no-ec2m to disable characteristic 2 code.
4986 * New build option "fipscanisteronly". This only builds fipscanister.o
4987 and (currently) associated fips utilities. Uses the file Makefile.fips
4988 instead of Makefile.org as the prototype.
4992 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4993 Update fips_gcmtest to use IV generator.
4997 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4998 setting output buffer to NULL. The `*Final` function must be
4999 called although it will not retrieve any additional data. The tag
5000 can be set or retrieved with a ctrl. The IV length is by default 12
5001 bytes (96 bits) but can be set to an alternative value. If the IV
5002 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5007 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5008 underlying do_cipher function handles all cipher semantics itself
5009 including padding and finalisation. This is useful if (for example)
5010 an ENGINE cipher handles block padding itself. The behaviour of
5011 do_cipher is subtly changed if this flag is set: the return value
5012 is the number of characters written to the output buffer (zero is
5013 no longer an error code) or a negative error code. Also if the
5014 input buffer is NULL and length 0 finalisation should be performed.
5018 * If a candidate issuer certificate is already part of the constructed
5019 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5023 * Improve forward-security support: add functions
5025 void SSL_CTX_set_not_resumable_session_callback(
5026 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5027 void SSL_set_not_resumable_session_callback(
5028 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5030 for use by SSL/TLS servers; the callback function will be called whenever a
5031 new session is created, and gets to decide whether the session may be
5032 cached to make it resumable (return 0) or not (return 1). (As by the
5033 SSL/TLS protocol specifications, the session_id sent by the server will be
5034 empty to indicate that the session is not resumable; also, the server will
5035 not generate RFC 4507 (RFC 5077) session tickets.)
5037 A simple reasonable callback implementation is to return is_forward_secure.
5038 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5039 by the SSL/TLS server library, indicating whether it can provide forward
5042 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5044 * New -verify_name option in command line utilities to set verification
5049 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5050 Add CMAC pkey methods.
5054 * Experimental renegotiation in s_server -www mode. If the client
5055 browses /reneg connection is renegotiated. If /renegcert it is
5056 renegotiated requesting a certificate.
5060 * Add an "external" session cache for debugging purposes to s_server. This
5061 should help trace issues which normally are only apparent in deployed
5062 multi-process servers.
5066 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5067 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5068 BIO_set_cipher() and some obscure PEM functions were changed so they
5069 can now return an error. The RAND changes required a change to the
5070 RAND_METHOD structure.
5074 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5075 a gcc attribute to warn if the result of a function is ignored. This
5076 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5077 whose return value is often ignored.
5081 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5082 These allow SCTs (signed certificate timestamps) to be requested and
5083 validated when establishing a connection.
5085 *Rob Percival <robpercival@google.com>*
5090 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5092 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5093 used even when parsing explicit parameters, when loading a encoded key
5094 or calling `EC_GROUP_new_from_ecpkparameters()`/
5095 `EC_GROUP_new_from_ecparameters()`.
5096 This prevents bypass of security hardening and performance gains,
5097 especially for curves with specialized EC_METHODs.
5098 By default, if a key encoded with explicit parameters is loaded and later
5099 encoded, the output is still encoded with explicit parameters, even if
5100 internally a "named" EC_GROUP is used for computation.
5104 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5105 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5106 NULL. After this change, only the cofactor parameter can be NULL. It also
5107 does some minimal sanity checks on the passed order.
5112 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5113 An attack is simple, if the first CMS_recipientInfo is valid but the
5114 second CMS_recipientInfo is chosen ciphertext. If the second
5115 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5116 encryption key will be replaced by garbage, and the message cannot be
5117 decoded, but if the RSA decryption fails, the correct encryption key is
5118 used and the recipient will not notice the attack.
5119 As a work around for this potential attack the length of the decrypted
5120 key must be equal to the cipher default key length, in case the
5121 certificate is not given and all recipientInfo are tried out.
5122 The old behaviour can be re-enabled in the CMS code by setting the
5123 CMS_DEBUG_DECRYPT flag.
5128 * Document issue with installation paths in diverse Windows builds
5130 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5131 binaries and run-time config file.
5136 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5138 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5139 This changes the size when using the `genpkey` command when no size is given.
5140 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5141 generation commands to use 2048 bits by default.
5145 * Add FIPS support for Android Arm 64-bit
5147 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5148 Module in Version 2.0.10. For some reason, the corresponding target
5149 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5150 built with FIPS support on Android Arm 64-bit. This omission has been
5153 *Matthias St. Pierre*
5155 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5157 * 0-byte record padding oracle
5159 If an application encounters a fatal protocol error and then calls
5160 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5161 then OpenSSL can respond differently to the calling application if a 0 byte
5162 record is received with invalid padding compared to if a 0 byte record is
5163 received with an invalid MAC. If the application then behaves differently
5164 based on that in a way that is detectable to the remote peer, then this
5165 amounts to a padding oracle that could be used to decrypt data.
5167 In order for this to be exploitable "non-stitched" ciphersuites must be in
5168 use. Stitched ciphersuites are optimised implementations of certain
5169 commonly used ciphersuites. Also the application must call SSL_shutdown()
5170 twice even if a protocol error has occurred (applications should not do
5171 this but some do anyway).
5173 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5174 Aviram, with additional investigation by Steven Collison and Andrew
5175 Hourselt. It was reported to OpenSSL on 10th December 2018.
5180 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5184 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5186 * Microarchitecture timing vulnerability in ECC scalar multiplication
5188 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5189 shown to be vulnerable to a microarchitecture timing side channel attack.
5190 An attacker with sufficient access to mount local timing attacks during
5191 ECDSA signature generation could recover the private key.
5193 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5194 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5200 * Timing vulnerability in DSA signature generation
5202 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5203 timing side channel attack. An attacker could use variations in the signing
5204 algorithm to recover the private key.
5206 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5211 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5212 Module, accidentally introduced while backporting security fixes from the
5213 development branch and hindering the use of ECC in FIPS mode.
5217 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5219 * Client DoS due to large DH parameter
5221 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5222 malicious server can send a very large prime value to the client. This will
5223 cause the client to spend an unreasonably long period of time generating a
5224 key for this prime resulting in a hang until the client has finished. This
5225 could be exploited in a Denial Of Service attack.
5227 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5232 * Cache timing vulnerability in RSA Key Generation
5234 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5235 a cache timing side channel attack. An attacker with sufficient access to
5236 mount cache timing attacks during the RSA key generation process could
5237 recover the private key.
5239 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5240 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5245 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5246 parameter is no longer accepted, as it leads to a corrupt table. NULL
5247 pem_str is reserved for alias entries only.
5251 * Revert blinding in ECDSA sign and instead make problematic addition
5252 length-invariant. Switch even to fixed-length Montgomery multiplication.
5256 * Change generating and checking of primes so that the error rate of not
5257 being prime depends on the intended use based on the size of the input.
5258 For larger primes this will result in more rounds of Miller-Rabin.
5259 The maximal error rate for primes with more than 1080 bits is lowered
5262 *Kurt Roeckx, Annie Yousar*
5264 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5268 * Add blinding to ECDSA and DSA signatures to protect against side channel
5269 attacks discovered by Keegan Ryan (NCC Group).
5273 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5274 now allow empty (zero character) pass phrases.
5278 * Certificate time validation (X509_cmp_time) enforces stricter
5279 compliance with RFC 5280. Fractional seconds and timezone offsets
5280 are no longer allowed.
5284 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5286 * Constructed ASN.1 types with a recursive definition could exceed the stack
5288 Constructed ASN.1 types with a recursive definition (such as can be found
5289 in PKCS7) could eventually exceed the stack given malicious input with
5290 excessive recursion. This could result in a Denial Of Service attack. There
5291 are no such structures used within SSL/TLS that come from untrusted sources
5292 so this is considered safe.
5294 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5300 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5302 * Read/write after SSL object in error state
5304 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5305 mechanism. The intent was that if a fatal error occurred during a handshake
5306 then OpenSSL would move into the error state and would immediately fail if
5307 you attempted to continue the handshake. This works as designed for the
5308 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5309 SSL_connect()), however due to a bug it does not work correctly if
5310 SSL_read() or SSL_write() is called directly. In that scenario, if the
5311 handshake fails then a fatal error will be returned in the initial function
5312 call. If SSL_read()/SSL_write() is subsequently called by the application
5313 for the same SSL object then it will succeed and the data is passed without
5314 being decrypted/encrypted directly from the SSL/TLS record layer.
5316 In order to exploit this issue an application bug would have to be present
5317 that resulted in a call to SSL_read()/SSL_write() being issued after having
5318 already received a fatal error.
5320 This issue was reported to OpenSSL by David Benjamin (Google).
5325 * rsaz_1024_mul_avx2 overflow bug on x86_64
5327 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5328 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5329 Analysis suggests that attacks against RSA and DSA as a result of this
5330 defect would be very difficult to perform and are not believed likely.
5331 Attacks against DH1024 are considered just feasible, because most of the
5332 work necessary to deduce information about a private key may be performed
5333 offline. The amount of resources required for such an attack would be
5334 significant. However, for an attack on TLS to be meaningful, the server
5335 would have to share the DH1024 private key among multiple clients, which is
5336 no longer an option since CVE-2016-0701.
5338 This only affects processors that support the AVX2 but not ADX extensions
5339 like Intel Haswell (4th generation).
5341 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5342 was originally found via the OSS-Fuzz project.
5347 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5349 * bn_sqrx8x_internal carry bug on x86_64
5351 There is a carry propagating bug in the x86_64 Montgomery squaring
5352 procedure. No EC algorithms are affected. Analysis suggests that attacks
5353 against RSA and DSA as a result of this defect would be very difficult to
5354 perform and are not believed likely. Attacks against DH are considered just
5355 feasible (although very difficult) because most of the work necessary to
5356 deduce information about a private key may be performed offline. The amount
5357 of resources required for such an attack would be very significant and
5358 likely only accessible to a limited number of attackers. An attacker would
5359 additionally need online access to an unpatched system using the target
5360 private key in a scenario with persistent DH parameters and a private
5361 key that is shared between multiple clients.
5363 This only affects processors that support the BMI1, BMI2 and ADX extensions
5364 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5366 This issue was reported to OpenSSL by the OSS-Fuzz project.
5371 * Malformed X.509 IPAddressFamily could cause OOB read
5373 If an X.509 certificate has a malformed IPAddressFamily extension,
5374 OpenSSL could do a one-byte buffer overread. The most likely result
5375 would be an erroneous display of the certificate in text format.
5377 This issue was reported to OpenSSL by the OSS-Fuzz project.
5381 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5383 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5384 platform rather than 'mingw'.
5388 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5390 * Truncated packet could crash via OOB read
5392 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5393 cipher is being used, then a truncated packet can cause that host to
5394 perform an out-of-bounds read, usually resulting in a crash.
5396 This issue was reported to OpenSSL by Robert Święcki of Google.
5401 * BN_mod_exp may produce incorrect results on x86_64
5403 There is a carry propagating bug in the x86_64 Montgomery squaring
5404 procedure. No EC algorithms are affected. Analysis suggests that attacks
5405 against RSA and DSA as a result of this defect would be very difficult to
5406 perform and are not believed likely. Attacks against DH are considered just
5407 feasible (although very difficult) because most of the work necessary to
5408 deduce information about a private key may be performed offline. The amount
5409 of resources required for such an attack would be very significant and
5410 likely only accessible to a limited number of attackers. An attacker would
5411 additionally need online access to an unpatched system using the target
5412 private key in a scenario with persistent DH parameters and a private
5413 key that is shared between multiple clients. For example this can occur by
5414 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5415 similar to CVE-2015-3193 but must be treated as a separate problem.
5417 This issue was reported to OpenSSL by the OSS-Fuzz project.
5422 * Montgomery multiplication may produce incorrect results
5424 There is a carry propagating bug in the Broadwell-specific Montgomery
5425 multiplication procedure that handles input lengths divisible by, but
5426 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5427 and DH private keys are impossible. This is because the subroutine in
5428 question is not used in operations with the private key itself and an input
5429 of the attacker's direct choice. Otherwise the bug can manifest itself as
5430 transient authentication and key negotiation failures or reproducible
5431 erroneous outcome of public-key operations with specially crafted input.
5432 Among EC algorithms only Brainpool P-512 curves are affected and one
5433 presumably can attack ECDH key negotiation. Impact was not analyzed in
5434 detail, because pre-requisites for attack are considered unlikely. Namely
5435 multiple clients have to choose the curve in question and the server has to
5436 share the private key among them, neither of which is default behaviour.
5437 Even then only clients that chose the curve will be affected.
5439 This issue was publicly reported as transient failures and was not
5440 initially recognized as a security issue. Thanks to Richard Morgan for
5441 providing reproducible case.
5446 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5447 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5448 prevent issues where no progress is being made and the peer continually
5449 sends unrecognised record types, using up resources processing them.
5453 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5455 * Missing CRL sanity check
5457 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5458 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5459 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5461 This issue only affects the OpenSSL 1.0.2i
5466 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5468 * OCSP Status Request extension unbounded memory growth
5470 A malicious client can send an excessively large OCSP Status Request
5471 extension. If that client continually requests renegotiation, sending a
5472 large OCSP Status Request extension each time, then there will be unbounded
5473 memory growth on the server. This will eventually lead to a Denial Of
5474 Service attack through memory exhaustion. Servers with a default
5475 configuration are vulnerable even if they do not support OCSP. Builds using
5476 the "no-ocsp" build time option are not affected.
5478 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5483 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5486 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5492 * OOB write in MDC2_Update()
5494 An overflow can occur in MDC2_Update() either if called directly or
5495 through the EVP_DigestUpdate() function using MDC2. If an attacker
5496 is able to supply very large amounts of input data after a previous
5497 call to EVP_EncryptUpdate() with a partial block then a length check
5498 can overflow resulting in a heap corruption.
5500 The amount of data needed is comparable to SIZE_MAX which is impractical
5503 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5508 * Malformed SHA512 ticket DoS
5510 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5511 DoS attack where a malformed ticket will result in an OOB read which will
5514 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5515 a custom server callback and ticket lookup mechanism.
5517 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5522 * OOB write in BN_bn2dec()
5524 The function BN_bn2dec() does not check the return value of BN_div_word().
5525 This can cause an OOB write if an application uses this function with an
5526 overly large BIGNUM. This could be a problem if an overly large certificate
5527 or CRL is printed out from an untrusted source. TLS is not affected because
5528 record limits will reject an oversized certificate before it is parsed.
5530 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5535 * OOB read in TS_OBJ_print_bio()
5537 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5538 the total length the OID text representation would use and not the amount
5539 of data written. This will result in OOB reads when large OIDs are
5542 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5547 * Pointer arithmetic undefined behaviour
5549 Avoid some undefined pointer arithmetic
5551 A common idiom in the codebase is to check limits in the following manner:
5554 Where "p" points to some malloc'd data of SIZE bytes and
5557 "len" here could be from some externally supplied data (e.g. from a TLS
5560 The rules of C pointer arithmetic are such that "p + len" is only well
5561 defined where len <= SIZE. Therefore the above idiom is actually
5562 undefined behaviour.
5564 For example this could cause problems if some malloc implementation
5565 provides an address for "p" such that "p + len" actually overflows for
5566 values of len that are too big and therefore p + len < limit.
5568 This issue was reported to OpenSSL by Guido Vranken
5573 * Constant time flag not preserved in DSA signing
5575 Operations in the DSA signing algorithm should run in constant time in
5576 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5577 implementation means that a non-constant time codepath is followed for
5578 certain operations. This has been demonstrated through a cache-timing
5579 attack to be sufficient for an attacker to recover the private DSA key.
5581 This issue was reported by César Pereida (Aalto University), Billy Brumley
5582 (Tampere University of Technology), and Yuval Yarom (The University of
5583 Adelaide and NICTA).
5588 * DTLS buffered message DoS
5590 In a DTLS connection where handshake messages are delivered out-of-order
5591 those messages that OpenSSL is not yet ready to process will be buffered
5592 for later use. Under certain circumstances, a flaw in the logic means that
5593 those messages do not get removed from the buffer even though the handshake
5594 has been completed. An attacker could force up to approx. 15 messages to
5595 remain in the buffer when they are no longer required. These messages will
5596 be cleared when the DTLS connection is closed. The default maximum size for
5597 a message is 100k. Therefore, the attacker could force an additional 1500k
5598 to be consumed per connection. By opening many simultaneous connections an
5599 attacker could cause a DoS attack through memory exhaustion.
5601 This issue was reported to OpenSSL by Quan Luo.
5606 * DTLS replay protection DoS
5608 A flaw in the DTLS replay attack protection mechanism means that records
5609 that arrive for future epochs update the replay protection "window" before
5610 the MAC for the record has been validated. This could be exploited by an
5611 attacker by sending a record for the next epoch (which does not have to
5612 decrypt or have a valid MAC), with a very large sequence number. This means
5613 that all subsequent legitimate packets are dropped causing a denial of
5614 service for a specific DTLS connection.
5616 This issue was reported to OpenSSL by the OCAP audit team.
5621 * Certificate message OOB reads
5623 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5624 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5625 theoretical DoS risk but this has not been observed in practice on common
5628 The messages affected are client certificate, client certificate request
5629 and server certificate. As a result the attack can only be performed
5630 against a client or a server which enables client authentication.
5632 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5637 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5639 * Prevent padding oracle in AES-NI CBC MAC check
5641 A MITM attacker can use a padding oracle attack to decrypt traffic
5642 when the connection uses an AES CBC cipher and the server support
5645 This issue was introduced as part of the fix for Lucky 13 padding
5646 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5647 constant time by making sure that always the same bytes are read and
5648 compared against either the MAC or padding bytes. But it no longer
5649 checked that there was enough data to have both the MAC and padding
5652 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5656 * Fix EVP_EncodeUpdate overflow
5658 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5659 Base64 encoding of binary data. If an attacker is able to supply very large
5660 amounts of input data then a length check can overflow resulting in a heap
5663 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5664 the `PEM_write_bio*` family of functions. These are mainly used within the
5665 OpenSSL command line applications, so any application which processes data
5666 from an untrusted source and outputs it as a PEM file should be considered
5667 vulnerable to this issue. User applications that call these APIs directly
5668 with large amounts of untrusted data may also be vulnerable.
5670 This issue was reported by Guido Vranken.
5675 * Fix EVP_EncryptUpdate overflow
5677 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5678 is able to supply very large amounts of input data after a previous call to
5679 EVP_EncryptUpdate() with a partial block then a length check can overflow
5680 resulting in a heap corruption. Following an analysis of all OpenSSL
5681 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5682 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5683 the first called function after an EVP_EncryptInit(), and therefore that
5684 specific call must be safe. The second form is where the length passed to
5685 EVP_EncryptUpdate() can be seen from the code to be some small value and
5686 therefore there is no possibility of an overflow. Since all instances are
5687 one of these two forms, it is believed that there can be no overflows in
5688 internal code due to this problem. It should be noted that
5689 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5690 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5691 of these calls have also been analysed too and it is believed there are no
5692 instances in internal usage where an overflow could occur.
5694 This issue was reported by Guido Vranken.
5699 * Prevent ASN.1 BIO excessive memory allocation
5701 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5702 a short invalid encoding can cause allocation of large amounts of memory
5703 potentially consuming excessive resources or exhausting memory.
5705 Any application parsing untrusted data through d2i BIO functions is
5706 affected. The memory based functions such as d2i_X509() are *not* affected.
5707 Since the memory based functions are used by the TLS library, TLS
5708 applications are not affected.
5710 This issue was reported by Brian Carpenter.
5717 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5718 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5719 in arbitrary stack data being returned in the buffer.
5721 This issue was reported by Guido Vranken.
5726 * Modify behavior of ALPN to invoke callback after SNI/servername
5727 callback, such that updates to the SSL_CTX affect ALPN.
5731 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5736 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5737 methods are enabled and ssl2 is disabled the methods return NULL.
5741 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5743 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5744 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5745 provide any "EXPORT" or "LOW" strength ciphers.
5749 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5750 is by default disabled at build-time. Builds that are not configured with
5751 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5752 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5753 will need to explicitly call either of:
5755 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5757 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5759 as appropriate. Even if either of those is used, or the application
5760 explicitly uses the version-specific SSLv2_method() or its client and
5761 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5762 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5763 ciphers, and SSLv2 56-bit DES are no longer available.
5768 * Fix a double-free in DSA code
5770 A double free bug was discovered when OpenSSL parses malformed DSA private
5771 keys and could lead to a DoS attack or memory corruption for applications
5772 that receive DSA private keys from untrusted sources. This scenario is
5775 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5781 * Disable SRP fake user seed to address a server memory leak.
5783 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5785 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5786 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5787 was changed to ignore the "fake user" SRP seed, even if the seed
5790 Users should use SRP_VBASE_get1_by_user instead. Note that in
5791 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5792 also that even though configuring the SRP seed attempts to hide
5793 invalid usernames by continuing the handshake with fake
5794 credentials, this behaviour is not constant time and no strong
5795 guarantees are made that the handshake is indistinguishable from
5796 that of a valid user.
5801 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5803 In the BN_hex2bn function the number of hex digits is calculated using an
5804 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5805 large values of `i` this can result in `bn_expand` not allocating any
5806 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5807 field as NULL leading to a subsequent NULL ptr deref. For very large values
5808 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5809 In this case memory is allocated to the internal BIGNUM data field, but it
5810 is insufficiently sized leading to heap corruption. A similar issue exists
5811 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5812 is ever called by user applications with very large untrusted hex/dec data.
5813 This is anticipated to be a rare occurrence.
5815 All OpenSSL internal usage of these functions use data that is not expected
5816 to be untrusted, e.g. config file data or application command line
5817 arguments. If user developed applications generate config file data based
5818 on untrusted data then it is possible that this could also lead to security
5819 consequences. This is also anticipated to be rare.
5821 This issue was reported to OpenSSL by Guido Vranken.
5826 * Fix memory issues in `BIO_*printf` functions
5828 The internal `fmtstr` function used in processing a "%s" format string in
5829 the `BIO_*printf` functions could overflow while calculating the length of a
5830 string and cause an OOB read when printing very long strings.
5832 Additionally the internal `doapr_outch` function can attempt to write to an
5833 OOB memory location (at an offset from the NULL pointer) in the event of a
5834 memory allocation failure. In 1.0.2 and below this could be caused where
5835 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5836 could be in processing a very long "%s" format string. Memory leaks can
5839 The first issue may mask the second issue dependent on compiler behaviour.
5840 These problems could enable attacks where large amounts of untrusted data
5841 is passed to the `BIO_*printf` functions. If applications use these functions
5842 in this way then they could be vulnerable. OpenSSL itself uses these
5843 functions when printing out human-readable dumps of ASN.1 data. Therefore
5844 applications that print this data could be vulnerable if the data is from
5845 untrusted sources. OpenSSL command line applications could also be
5846 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5847 as command line arguments.
5849 Libssl is not considered directly vulnerable. Additionally certificates etc
5850 received via remote connections via libssl are also unlikely to be able to
5851 trigger these issues because of message size limits enforced within libssl.
5853 This issue was reported to OpenSSL Guido Vranken.
5858 * Side channel attack on modular exponentiation
5860 A side-channel attack was found which makes use of cache-bank conflicts on
5861 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5862 of RSA keys. The ability to exploit this issue is limited as it relies on
5863 an attacker who has control of code in a thread running on the same
5864 hyper-threaded core as the victim thread which is performing decryptions.
5866 This issue was reported to OpenSSL by Yuval Yarom, The University of
5867 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5868 Nadia Heninger, University of Pennsylvania with more information at
5869 <http://cachebleed.info>.
5874 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5875 if no keysize is specified with default_bits. This fixes an
5876 omission in an earlier change that changed all RSA/DSA key generation
5877 commands to use 2048 bits by default.
5881 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5883 * DH small subgroups
5885 Historically OpenSSL only ever generated DH parameters based on "safe"
5886 primes. More recently (in version 1.0.2) support was provided for
5887 generating X9.42 style parameter files such as those required for RFC 5114
5888 support. The primes used in such files may not be "safe". Where an
5889 application is using DH configured with parameters based on primes that are
5890 not "safe" then an attacker could use this fact to find a peer's private
5891 DH exponent. This attack requires that the attacker complete multiple
5892 handshakes in which the peer uses the same private DH exponent. For example
5893 this could be used to discover a TLS server's private DH exponent if it's
5894 reusing the private DH exponent or it's using a static DH ciphersuite.
5896 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5897 TLS. It is not on by default. If the option is not set then the server
5898 reuses the same private DH exponent for the life of the server process and
5899 would be vulnerable to this attack. It is believed that many popular
5900 applications do set this option and would therefore not be at risk.
5902 The fix for this issue adds an additional check where a "q" parameter is
5903 available (as is the case in X9.42 based parameters). This detects the
5904 only known attack, and is the only possible defense for static DH
5905 ciphersuites. This could have some performance impact.
5907 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5908 default and cannot be disabled. This could have some performance impact.
5910 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5915 * SSLv2 doesn't block disabled ciphers
5917 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5918 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5919 been disabled, provided that the SSLv2 protocol was not also disabled via
5922 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5923 and Sebastian Schinzel.
5928 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5930 * BN_mod_exp may produce incorrect results on x86_64
5932 There is a carry propagating bug in the x86_64 Montgomery squaring
5933 procedure. No EC algorithms are affected. Analysis suggests that attacks
5934 against RSA and DSA as a result of this defect would be very difficult to
5935 perform and are not believed likely. Attacks against DH are considered just
5936 feasible (although very difficult) because most of the work necessary to
5937 deduce information about a private key may be performed offline. The amount
5938 of resources required for such an attack would be very significant and
5939 likely only accessible to a limited number of attackers. An attacker would
5940 additionally need online access to an unpatched system using the target
5941 private key in a scenario with persistent DH parameters and a private
5942 key that is shared between multiple clients. For example this can occur by
5943 default in OpenSSL DHE based SSL/TLS ciphersuites.
5945 This issue was reported to OpenSSL by Hanno Böck.
5950 * Certificate verify crash with missing PSS parameter
5952 The signature verification routines will crash with a NULL pointer
5953 dereference if presented with an ASN.1 signature using the RSA PSS
5954 algorithm and absent mask generation function parameter. Since these
5955 routines are used to verify certificate signature algorithms this can be
5956 used to crash any certificate verification operation and exploited in a
5957 DoS attack. Any application which performs certificate verification is
5958 vulnerable including OpenSSL clients and servers which enable client
5961 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5966 * X509_ATTRIBUTE memory leak
5968 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5969 memory. This structure is used by the PKCS#7 and CMS routines so any
5970 application which reads PKCS#7 or CMS data from untrusted sources is
5971 affected. SSL/TLS is not affected.
5973 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5979 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5980 This changes the decoding behaviour for some invalid messages,
5981 though the change is mostly in the more lenient direction, and
5982 legacy behaviour is preserved as much as possible.
5986 * In DSA_generate_parameters_ex, if the provided seed is too short,
5989 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5991 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5993 * Alternate chains certificate forgery
5995 During certificate verification, OpenSSL will attempt to find an
5996 alternative certificate chain if the first attempt to build such a chain
5997 fails. An error in the implementation of this logic can mean that an
5998 attacker could cause certain checks on untrusted certificates to be
5999 bypassed, such as the CA flag, enabling them to use a valid leaf
6000 certificate to act as a CA and "issue" an invalid certificate.
6002 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6007 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6009 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6010 incompatibility in the handling of HMAC. The previous ABI has now been
6015 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6017 * Malformed ECParameters causes infinite loop
6019 When processing an ECParameters structure OpenSSL enters an infinite loop
6020 if the curve specified is over a specially malformed binary polynomial
6023 This can be used to perform denial of service against any
6024 system which processes public keys, certificate requests or
6025 certificates. This includes TLS clients and TLS servers with
6026 client authentication enabled.
6028 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6033 * Exploitable out-of-bounds read in X509_cmp_time
6035 X509_cmp_time does not properly check the length of the ASN1_TIME
6036 string and can read a few bytes out of bounds. In addition,
6037 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6040 An attacker can use this to craft malformed certificates and CRLs of
6041 various sizes and potentially cause a segmentation fault, resulting in
6042 a DoS on applications that verify certificates or CRLs. TLS clients
6043 that verify CRLs are affected. TLS clients and servers with client
6044 authentication enabled may be affected if they use custom verification
6047 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6048 independently by Hanno Böck.
6053 * PKCS7 crash with missing EnvelopedContent
6055 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6056 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6057 with missing content and trigger a NULL pointer dereference on parsing.
6059 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6060 structures from untrusted sources are affected. OpenSSL clients and
6061 servers are not affected.
6063 This issue was reported to OpenSSL by Michal Zalewski (Google).
6068 * CMS verify infinite loop with unknown hash function
6070 When verifying a signedData message the CMS code can enter an infinite loop
6071 if presented with an unknown hash function OID. This can be used to perform
6072 denial of service against any system which verifies signedData messages using
6074 This issue was reported to OpenSSL by Johannes Bauer.
6079 * Race condition handling NewSessionTicket
6081 If a NewSessionTicket is received by a multi-threaded client when attempting to
6082 reuse a previous ticket then a race condition can occur potentially leading to
6083 a double free of the ticket data.
6088 * Only support 256-bit or stronger elliptic curves with the
6089 'ecdh_auto' setting (server) or by default (client). Of supported
6090 curves, prefer P-256 (both).
6094 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6096 * ClientHello sigalgs DoS fix
6098 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6099 invalid signature algorithms extension a NULL pointer dereference will
6100 occur. This can be exploited in a DoS attack against the server.
6102 This issue was was reported to OpenSSL by David Ramos of Stanford
6106 *Stephen Henson and Matt Caswell*
6108 * Multiblock corrupted pointer fix
6110 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6111 feature only applies on 64 bit x86 architecture platforms that support AES
6112 NI instructions. A defect in the implementation of "multiblock" can cause
6113 OpenSSL's internal write buffer to become incorrectly set to NULL when
6114 using non-blocking IO. Typically, when the user application is using a
6115 socket BIO for writing, this will only result in a failed connection.
6116 However if some other BIO is used then it is likely that a segmentation
6117 fault will be triggered, thus enabling a potential DoS attack.
6119 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6124 * Segmentation fault in DTLSv1_listen fix
6126 The DTLSv1_listen function is intended to be stateless and processes the
6127 initial ClientHello from many peers. It is common for user code to loop
6128 over the call to DTLSv1_listen until a valid ClientHello is received with
6129 an associated cookie. A defect in the implementation of DTLSv1_listen means
6130 that state is preserved in the SSL object from one invocation to the next
6131 that can lead to a segmentation fault. Errors processing the initial
6132 ClientHello can trigger this scenario. An example of such an error could be
6133 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6136 This issue was reported to OpenSSL by Per Allansson.
6141 * Segmentation fault in ASN1_TYPE_cmp fix
6143 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6144 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6145 certificate signature algorithm consistency this can be used to crash any
6146 certificate verification operation and exploited in a DoS attack. Any
6147 application which performs certificate verification is vulnerable including
6148 OpenSSL clients and servers which enable client authentication.
6153 * Segmentation fault for invalid PSS parameters fix
6155 The signature verification routines will crash with a NULL pointer
6156 dereference if presented with an ASN.1 signature using the RSA PSS
6157 algorithm and invalid parameters. Since these routines are used to verify
6158 certificate signature algorithms this can be used to crash any
6159 certificate verification operation and exploited in a DoS attack. Any
6160 application which performs certificate verification is vulnerable including
6161 OpenSSL clients and servers which enable client authentication.
6163 This issue was was reported to OpenSSL by Brian Carpenter.
6168 * ASN.1 structure reuse memory corruption fix
6170 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6171 memory corruption via an invalid write. Such reuse is and has been
6172 strongly discouraged and is believed to be rare.
6174 Applications that parse structures containing CHOICE or ANY DEFINED BY
6175 components may be affected. Certificate parsing (d2i_X509 and related
6176 functions) are however not affected. OpenSSL clients and servers are
6182 * PKCS7 NULL pointer dereferences fix
6184 The PKCS#7 parsing code does not handle missing outer ContentInfo
6185 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6186 missing content and trigger a NULL pointer dereference on parsing.
6188 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6189 otherwise parse PKCS#7 structures from untrusted sources are
6190 affected. OpenSSL clients and servers are not affected.
6192 This issue was reported to OpenSSL by Michal Zalewski (Google).
6197 * DoS via reachable assert in SSLv2 servers fix
6199 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6200 servers that both support SSLv2 and enable export cipher suites by sending
6201 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6203 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6204 (OpenSSL development team).
6209 * Empty CKE with client auth and DHE fix
6211 If client auth is used then a server can seg fault in the event of a DHE
6212 ciphersuite being selected and a zero length ClientKeyExchange message
6213 being sent by the client. This could be exploited in a DoS attack.
6218 * Handshake with unseeded PRNG fix
6220 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6221 with an unseeded PRNG. The conditions are:
6222 - The client is on a platform where the PRNG has not been seeded
6223 automatically, and the user has not seeded manually
6224 - A protocol specific client method version has been used (i.e. not
6225 SSL_client_methodv23)
6226 - A ciphersuite is used that does not require additional random data from
6227 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6229 If the handshake succeeds then the client random that has been used will
6230 have been generated from a PRNG with insufficient entropy and therefore the
6231 output may be predictable.
6233 For example using the following command with an unseeded openssl will
6234 succeed on an unpatched platform:
6236 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6241 * Use After Free following d2i_ECPrivatekey error fix
6243 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6244 could cause a use after free condition. This, in turn, could cause a double
6245 free in several private key parsing functions (such as d2i_PrivateKey
6246 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6247 for applications that receive EC private keys from untrusted
6248 sources. This scenario is considered rare.
6250 This issue was discovered by the BoringSSL project and fixed in their
6256 * X509_to_X509_REQ NULL pointer deref fix
6258 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6259 the certificate key is invalid. This function is rarely used in practice.
6261 This issue was discovered by Brian Carpenter.
6266 * Removed the export ciphers from the DEFAULT ciphers
6270 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6272 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6273 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6274 So far those who have to target multiple platforms would compromise
6275 and argue that binary targeting say ARMv5 would still execute on
6276 ARMv8. "Universal" build resolves this compromise by providing
6277 near-optimal performance even on newer platforms.
6281 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6282 (other platforms pending).
6284 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6286 * Add support for the SignedCertificateTimestampList certificate and
6287 OCSP response extensions from RFC6962.
6291 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6292 for corner cases. (Certain input points at infinity could lead to
6293 bogus results, with non-infinity inputs mapped to infinity too.)
6297 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6298 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6299 common cases are optimized and there still is room for further
6300 improvements. Vector Permutation AES for Altivec is also added.
6304 * Add support for little-endian ppc64 Linux target.
6306 *Marcelo Cerri (IBM)*
6308 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6309 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6310 are optimized and there still is room for further improvements.
6311 Both 32- and 64-bit modes are supported.
6313 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6315 * Improved ARMv7 NEON support.
6319 * Support for SPARC Architecture 2011 crypto extensions, first
6320 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6321 SHA256/512, MD5, GHASH and modular exponentiation.
6323 *Andy Polyakov, David Miller*
6325 * Accelerated modular exponentiation for Intel processors, a.k.a.
6328 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6330 * Support for new and upcoming Intel processors, including AVX2,
6331 BMI and SHA ISA extensions. This includes additional "stitched"
6332 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6335 This work was sponsored by Intel Corp.
6339 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6340 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6341 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6345 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6346 this fixes a limitation in previous versions of OpenSSL.
6350 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6351 MGF1 digest and OAEP label.
6355 * Add EVP support for key wrapping algorithms, to avoid problems with
6356 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6357 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6358 algorithms and include tests cases.
6362 * Add functions to allocate and set the fields of an ECDSA_METHOD
6365 *Douglas E. Engert, Steve Henson*
6367 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6368 difference in days and seconds between two tm or ASN1_TIME structures.
6372 * Add -rev test option to s_server to just reverse order of characters
6373 received by client and send back to server. Also prints an abbreviated
6374 summary of the connection parameters.
6378 * New option -brief for s_client and s_server to print out a brief summary
6379 of connection parameters.
6383 * Add callbacks for arbitrary TLS extensions.
6385 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6387 * New option -crl_download in several openssl utilities to download CRLs
6388 from CRLDP extension in certificates.
6392 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6396 * New function X509_CRL_diff to generate a delta CRL from the difference
6397 of two full CRLs. Add support to "crl" utility.
6401 * New functions to set lookup_crls function and to retrieve
6402 X509_STORE from X509_STORE_CTX.
6406 * Print out deprecated issuer and subject unique ID fields in
6411 * Extend OCSP I/O functions so they can be used for simple general purpose
6412 HTTP as well as OCSP. New wrapper function which can be used to download
6413 CRLs using the OCSP API.
6417 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6421 * `SSL_CONF*` functions. These provide a common framework for application
6422 configuration using configuration files or command lines.
6426 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6427 message callback and prints the results. Needs compile time option
6428 "enable-ssl-trace". New options to s_client and s_server to enable
6433 * New ctrl and macro to retrieve supported points extensions.
6434 Print out extension in s_server and s_client.
6438 * New functions to retrieve certificate signature and signature
6443 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6448 * New Suite B modes for TLS code. These use and enforce the requirements
6449 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6450 only use Suite B curves. The Suite B modes can be set by using the
6451 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6455 * New chain verification flags for Suite B levels of security. Check
6456 algorithms are acceptable when flags are set in X509_verify_cert.
6460 * Make tls1_check_chain return a set of flags indicating checks passed
6461 by a certificate chain. Add additional tests to handle client
6462 certificates: checks for matching certificate type and issuer name
6467 * If an attempt is made to use a signature algorithm not in the peer
6468 preference list abort the handshake. If client has no suitable
6469 signature algorithms in response to a certificate request do not
6470 use the certificate.
6474 * If server EC tmp key is not in client preference list abort handshake.
6478 * Add support for certificate stores in CERT structure. This makes it
6479 possible to have different stores per SSL structure or one store in
6480 the parent SSL_CTX. Include distinct stores for certificate chain
6481 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6482 to build and store a certificate chain in CERT structure: returning
6483 an error if the chain cannot be built: this will allow applications
6484 to test if a chain is correctly configured.
6486 Note: if the CERT based stores are not set then the parent SSL_CTX
6487 store is used to retain compatibility with existing behaviour.
6491 * New function ssl_set_client_disabled to set a ciphersuite disabled
6492 mask based on the current session, check mask when sending client
6493 hello and checking the requested ciphersuite.
6497 * New ctrls to retrieve and set certificate types in a certificate
6498 request message. Print out received values in s_client. If certificate
6499 types is not set with custom values set sensible values based on
6500 supported signature algorithms.
6504 * Support for distinct client and server supported signature algorithms.
6508 * Add certificate callback. If set this is called whenever a certificate
6509 is required by client or server. An application can decide which
6510 certificate chain to present based on arbitrary criteria: for example
6511 supported signature algorithms. Add very simple example to s_server.
6512 This fixes many of the problems and restrictions of the existing client
6513 certificate callback: for example you can now clear an existing
6514 certificate and specify the whole chain.
6518 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6519 the certificate can be used for (if anything). Set valid_flags field
6520 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6521 to have similar checks in it.
6523 Add new "cert_flags" field to CERT structure and include a "strict mode".
6524 This enforces some TLS certificate requirements (such as only permitting
6525 certificate signature algorithms contained in the supported algorithms
6526 extension) which some implementations ignore: this option should be used
6527 with caution as it could cause interoperability issues.
6531 * Update and tidy signature algorithm extension processing. Work out
6532 shared signature algorithms based on preferences and peer algorithms
6533 and print them out in s_client and s_server. Abort handshake if no
6534 shared signature algorithms.
6538 * Add new functions to allow customised supported signature algorithms
6539 for SSL and SSL_CTX structures. Add options to s_client and s_server
6544 * New function SSL_certs_clear() to delete all references to certificates
6545 from an SSL structure. Before this once a certificate had been added
6546 it couldn't be removed.
6550 * Integrate hostname, email address and IP address checking with certificate
6551 verification. New verify options supporting checking in openssl utility.
6555 * Fixes and wildcard matching support to hostname and email checking
6556 functions. Add manual page.
6558 *Florian Weimer (Red Hat Product Security Team)*
6560 * New functions to check a hostname email or IP address against a
6561 certificate. Add options x509 utility to print results of checks against
6566 * Fix OCSP checking.
6568 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6570 * Initial experimental support for explicitly trusted non-root CAs.
6571 OpenSSL still tries to build a complete chain to a root but if an
6572 intermediate CA has a trust setting included that is used. The first
6573 setting is used: whether to trust (e.g., -addtrust option to the x509
6578 * Add -trusted_first option which attempts to find certificates in the
6579 trusted store even if an untrusted chain is also supplied.
6583 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6584 platform support for Linux and Android.
6588 * Support for linux-x32, ILP32 environment in x86_64 framework.
6592 * Experimental multi-implementation support for FIPS capable OpenSSL.
6593 When in FIPS mode the approved implementations are used as normal,
6594 when not in FIPS mode the internal unapproved versions are used instead.
6595 This means that the FIPS capable OpenSSL isn't forced to use the
6596 (often lower performance) FIPS implementations outside FIPS mode.
6600 * Transparently support X9.42 DH parameters when calling
6601 PEM_read_bio_DHparameters. This means existing applications can handle
6602 the new parameter format automatically.
6606 * Initial experimental support for X9.42 DH parameter format: mainly
6607 to support use of 'q' parameter for RFC5114 parameters.
6611 * Add DH parameters from RFC5114 including test data to dhtest.
6615 * Support for automatic EC temporary key parameter selection. If enabled
6616 the most preferred EC parameters are automatically used instead of
6617 hardcoded fixed parameters. Now a server just has to call:
6618 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6619 support ECDH and use the most appropriate parameters.
6623 * Enhance and tidy EC curve and point format TLS extension code. Use
6624 static structures instead of allocation if default values are used.
6625 New ctrls to set curves we wish to support and to retrieve shared curves.
6626 Print out shared curves in s_server. New options to s_server and s_client
6627 to set list of supported curves.
6631 * New ctrls to retrieve supported signature algorithms and
6632 supported curve values as an array of NIDs. Extend openssl utility
6633 to print out received values.
6637 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6638 between NIDs and the more common NIST names such as "P-256". Enhance
6639 ecparam utility and ECC method to recognise the NIST names for curves.
6643 * Enhance SSL/TLS certificate chain handling to support different
6644 chains for each certificate instead of one chain in the parent SSL_CTX.
6648 * Support for fixed DH ciphersuite client authentication: where both
6649 server and client use DH certificates with common parameters.
6653 * Support for fixed DH ciphersuites: those requiring DH server
6658 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6660 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6661 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6662 X509_CINF_get_signature were reverted post internal team review.
6667 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6669 * OCSP Status Request extension unbounded memory growth
6671 A malicious client can send an excessively large OCSP Status Request
6672 extension. If that client continually requests renegotiation, sending a
6673 large OCSP Status Request extension each time, then there will be unbounded
6674 memory growth on the server. This will eventually lead to a Denial Of
6675 Service attack through memory exhaustion. Servers with a default
6676 configuration are vulnerable even if they do not support OCSP. Builds using
6677 the "no-ocsp" build time option are not affected.
6679 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6684 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6687 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6693 * OOB write in MDC2_Update()
6695 An overflow can occur in MDC2_Update() either if called directly or
6696 through the EVP_DigestUpdate() function using MDC2. If an attacker
6697 is able to supply very large amounts of input data after a previous
6698 call to EVP_EncryptUpdate() with a partial block then a length check
6699 can overflow resulting in a heap corruption.
6701 The amount of data needed is comparable to SIZE_MAX which is impractical
6704 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6709 * Malformed SHA512 ticket DoS
6711 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6712 DoS attack where a malformed ticket will result in an OOB read which will
6715 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6716 a custom server callback and ticket lookup mechanism.
6718 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6723 * OOB write in BN_bn2dec()
6725 The function BN_bn2dec() does not check the return value of BN_div_word().
6726 This can cause an OOB write if an application uses this function with an
6727 overly large BIGNUM. This could be a problem if an overly large certificate
6728 or CRL is printed out from an untrusted source. TLS is not affected because
6729 record limits will reject an oversized certificate before it is parsed.
6731 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6736 * OOB read in TS_OBJ_print_bio()
6738 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6739 the total length the OID text representation would use and not the amount
6740 of data written. This will result in OOB reads when large OIDs are
6743 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6748 * Pointer arithmetic undefined behaviour
6750 Avoid some undefined pointer arithmetic
6752 A common idiom in the codebase is to check limits in the following manner:
6755 Where "p" points to some malloc'd data of SIZE bytes and
6758 "len" here could be from some externally supplied data (e.g. from a TLS
6761 The rules of C pointer arithmetic are such that "p + len" is only well
6762 defined where len <= SIZE. Therefore, the above idiom is actually
6763 undefined behaviour.
6765 For example this could cause problems if some malloc implementation
6766 provides an address for "p" such that "p + len" actually overflows for
6767 values of len that are too big and therefore p + len < limit.
6769 This issue was reported to OpenSSL by Guido Vranken
6774 * Constant time flag not preserved in DSA signing
6776 Operations in the DSA signing algorithm should run in constant time in
6777 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6778 implementation means that a non-constant time codepath is followed for
6779 certain operations. This has been demonstrated through a cache-timing
6780 attack to be sufficient for an attacker to recover the private DSA key.
6782 This issue was reported by César Pereida (Aalto University), Billy Brumley
6783 (Tampere University of Technology), and Yuval Yarom (The University of
6784 Adelaide and NICTA).
6789 * DTLS buffered message DoS
6791 In a DTLS connection where handshake messages are delivered out-of-order
6792 those messages that OpenSSL is not yet ready to process will be buffered
6793 for later use. Under certain circumstances, a flaw in the logic means that
6794 those messages do not get removed from the buffer even though the handshake
6795 has been completed. An attacker could force up to approx. 15 messages to
6796 remain in the buffer when they are no longer required. These messages will
6797 be cleared when the DTLS connection is closed. The default maximum size for
6798 a message is 100k. Therefore, the attacker could force an additional 1500k
6799 to be consumed per connection. By opening many simultaneous connections an
6800 attacker could cause a DoS attack through memory exhaustion.
6802 This issue was reported to OpenSSL by Quan Luo.
6807 * DTLS replay protection DoS
6809 A flaw in the DTLS replay attack protection mechanism means that records
6810 that arrive for future epochs update the replay protection "window" before
6811 the MAC for the record has been validated. This could be exploited by an
6812 attacker by sending a record for the next epoch (which does not have to
6813 decrypt or have a valid MAC), with a very large sequence number. This means
6814 that all subsequent legitimate packets are dropped causing a denial of
6815 service for a specific DTLS connection.
6817 This issue was reported to OpenSSL by the OCAP audit team.
6822 * Certificate message OOB reads
6824 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6825 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6826 theoretical DoS risk but this has not been observed in practice on common
6829 The messages affected are client certificate, client certificate request
6830 and server certificate. As a result the attack can only be performed
6831 against a client or a server which enables client authentication.
6833 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6838 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6840 * Prevent padding oracle in AES-NI CBC MAC check
6842 A MITM attacker can use a padding oracle attack to decrypt traffic
6843 when the connection uses an AES CBC cipher and the server support
6846 This issue was introduced as part of the fix for Lucky 13 padding
6847 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6848 constant time by making sure that always the same bytes are read and
6849 compared against either the MAC or padding bytes. But it no longer
6850 checked that there was enough data to have both the MAC and padding
6853 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6858 * Fix EVP_EncodeUpdate overflow
6860 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6861 Base64 encoding of binary data. If an attacker is able to supply very large
6862 amounts of input data then a length check can overflow resulting in a heap
6865 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6866 the `PEM_write_bio*` family of functions. These are mainly used within the
6867 OpenSSL command line applications, so any application which processes data
6868 from an untrusted source and outputs it as a PEM file should be considered
6869 vulnerable to this issue. User applications that call these APIs directly
6870 with large amounts of untrusted data may also be vulnerable.
6872 This issue was reported by Guido Vranken.
6877 * Fix EVP_EncryptUpdate overflow
6879 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6880 is able to supply very large amounts of input data after a previous call to
6881 EVP_EncryptUpdate() with a partial block then a length check can overflow
6882 resulting in a heap corruption. Following an analysis of all OpenSSL
6883 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6884 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6885 the first called function after an EVP_EncryptInit(), and therefore that
6886 specific call must be safe. The second form is where the length passed to
6887 EVP_EncryptUpdate() can be seen from the code to be some small value and
6888 therefore there is no possibility of an overflow. Since all instances are
6889 one of these two forms, it is believed that there can be no overflows in
6890 internal code due to this problem. It should be noted that
6891 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6892 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6893 of these calls have also been analysed too and it is believed there are no
6894 instances in internal usage where an overflow could occur.
6896 This issue was reported by Guido Vranken.
6901 * Prevent ASN.1 BIO excessive memory allocation
6903 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6904 a short invalid encoding can casuse allocation of large amounts of memory
6905 potentially consuming excessive resources or exhausting memory.
6907 Any application parsing untrusted data through d2i BIO functions is
6908 affected. The memory based functions such as d2i_X509() are *not* affected.
6909 Since the memory based functions are used by the TLS library, TLS
6910 applications are not affected.
6912 This issue was reported by Brian Carpenter.
6919 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6920 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6921 in arbitrary stack data being returned in the buffer.
6923 This issue was reported by Guido Vranken.
6928 * Modify behavior of ALPN to invoke callback after SNI/servername
6929 callback, such that updates to the SSL_CTX affect ALPN.
6933 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6938 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6939 methods are enabled and ssl2 is disabled the methods return NULL.
6943 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6945 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6946 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6947 provide any "EXPORT" or "LOW" strength ciphers.
6951 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6952 is by default disabled at build-time. Builds that are not configured with
6953 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6954 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6955 will need to explicitly call either of:
6957 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6959 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6961 as appropriate. Even if either of those is used, or the application
6962 explicitly uses the version-specific SSLv2_method() or its client and
6963 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6964 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6965 ciphers, and SSLv2 56-bit DES are no longer available.
6970 * Fix a double-free in DSA code
6972 A double free bug was discovered when OpenSSL parses malformed DSA private
6973 keys and could lead to a DoS attack or memory corruption for applications
6974 that receive DSA private keys from untrusted sources. This scenario is
6977 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6983 * Disable SRP fake user seed to address a server memory leak.
6985 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6987 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6988 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6989 was changed to ignore the "fake user" SRP seed, even if the seed
6992 Users should use SRP_VBASE_get1_by_user instead. Note that in
6993 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6994 also that even though configuring the SRP seed attempts to hide
6995 invalid usernames by continuing the handshake with fake
6996 credentials, this behaviour is not constant time and no strong
6997 guarantees are made that the handshake is indistinguishable from
6998 that of a valid user.
7003 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7005 In the BN_hex2bn function the number of hex digits is calculated using an
7006 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7007 large values of `i` this can result in `bn_expand` not allocating any
7008 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7009 field as NULL leading to a subsequent NULL ptr deref. For very large values
7010 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7011 In this case memory is allocated to the internal BIGNUM data field, but it
7012 is insufficiently sized leading to heap corruption. A similar issue exists
7013 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7014 is ever called by user applications with very large untrusted hex/dec data.
7015 This is anticipated to be a rare occurrence.
7017 All OpenSSL internal usage of these functions use data that is not expected
7018 to be untrusted, e.g. config file data or application command line
7019 arguments. If user developed applications generate config file data based
7020 on untrusted data then it is possible that this could also lead to security
7021 consequences. This is also anticipated to be rare.
7023 This issue was reported to OpenSSL by Guido Vranken.
7028 * Fix memory issues in `BIO_*printf` functions
7030 The internal `fmtstr` function used in processing a "%s" format string in
7031 the `BIO_*printf` functions could overflow while calculating the length of a
7032 string and cause an OOB read when printing very long strings.
7034 Additionally the internal `doapr_outch` function can attempt to write to an
7035 OOB memory location (at an offset from the NULL pointer) in the event of a
7036 memory allocation failure. In 1.0.2 and below this could be caused where
7037 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7038 could be in processing a very long "%s" format string. Memory leaks can
7041 The first issue may mask the second issue dependent on compiler behaviour.
7042 These problems could enable attacks where large amounts of untrusted data
7043 is passed to the `BIO_*printf` functions. If applications use these functions
7044 in this way then they could be vulnerable. OpenSSL itself uses these
7045 functions when printing out human-readable dumps of ASN.1 data. Therefore
7046 applications that print this data could be vulnerable if the data is from
7047 untrusted sources. OpenSSL command line applications could also be
7048 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7049 as command line arguments.
7051 Libssl is not considered directly vulnerable. Additionally certificates etc
7052 received via remote connections via libssl are also unlikely to be able to
7053 trigger these issues because of message size limits enforced within libssl.
7055 This issue was reported to OpenSSL Guido Vranken.
7060 * Side channel attack on modular exponentiation
7062 A side-channel attack was found which makes use of cache-bank conflicts on
7063 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7064 of RSA keys. The ability to exploit this issue is limited as it relies on
7065 an attacker who has control of code in a thread running on the same
7066 hyper-threaded core as the victim thread which is performing decryptions.
7068 This issue was reported to OpenSSL by Yuval Yarom, The University of
7069 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7070 Nadia Heninger, University of Pennsylvania with more information at
7071 <http://cachebleed.info>.
7076 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7077 if no keysize is specified with default_bits. This fixes an
7078 omission in an earlier change that changed all RSA/DSA key generation
7079 commands to use 2048 bits by default.
7083 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7085 * Protection for DH small subgroup attacks
7087 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7088 switched on by default and cannot be disabled. This could have some
7093 * SSLv2 doesn't block disabled ciphers
7095 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7096 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7097 been disabled, provided that the SSLv2 protocol was not also disabled via
7100 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7101 and Sebastian Schinzel.
7106 * Reject DH handshakes with parameters shorter than 1024 bits.
7110 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7112 * Certificate verify crash with missing PSS parameter
7114 The signature verification routines will crash with a NULL pointer
7115 dereference if presented with an ASN.1 signature using the RSA PSS
7116 algorithm and absent mask generation function parameter. Since these
7117 routines are used to verify certificate signature algorithms this can be
7118 used to crash any certificate verification operation and exploited in a
7119 DoS attack. Any application which performs certificate verification is
7120 vulnerable including OpenSSL clients and servers which enable client
7123 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7128 * X509_ATTRIBUTE memory leak
7130 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7131 memory. This structure is used by the PKCS#7 and CMS routines so any
7132 application which reads PKCS#7 or CMS data from untrusted sources is
7133 affected. SSL/TLS is not affected.
7135 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7141 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7142 This changes the decoding behaviour for some invalid messages,
7143 though the change is mostly in the more lenient direction, and
7144 legacy behaviour is preserved as much as possible.
7148 * In DSA_generate_parameters_ex, if the provided seed is too short,
7149 use a random seed, as already documented.
7151 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7153 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7155 * Alternate chains certificate forgery
7157 During certificate verfification, OpenSSL will attempt to find an
7158 alternative certificate chain if the first attempt to build such a chain
7159 fails. An error in the implementation of this logic can mean that an
7160 attacker could cause certain checks on untrusted certificates to be
7161 bypassed, such as the CA flag, enabling them to use a valid leaf
7162 certificate to act as a CA and "issue" an invalid certificate.
7164 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7170 * Race condition handling PSK identify hint
7172 If PSK identity hints are received by a multi-threaded client then
7173 the values are wrongly updated in the parent SSL_CTX structure. This can
7174 result in a race condition potentially leading to a double free of the
7180 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7182 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7183 incompatibility in the handling of HMAC. The previous ABI has now been
7186 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7188 * Malformed ECParameters causes infinite loop
7190 When processing an ECParameters structure OpenSSL enters an infinite loop
7191 if the curve specified is over a specially malformed binary polynomial
7194 This can be used to perform denial of service against any
7195 system which processes public keys, certificate requests or
7196 certificates. This includes TLS clients and TLS servers with
7197 client authentication enabled.
7199 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7204 * Exploitable out-of-bounds read in X509_cmp_time
7206 X509_cmp_time does not properly check the length of the ASN1_TIME
7207 string and can read a few bytes out of bounds. In addition,
7208 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7211 An attacker can use this to craft malformed certificates and CRLs of
7212 various sizes and potentially cause a segmentation fault, resulting in
7213 a DoS on applications that verify certificates or CRLs. TLS clients
7214 that verify CRLs are affected. TLS clients and servers with client
7215 authentication enabled may be affected if they use custom verification
7218 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7219 independently by Hanno Böck.
7224 * PKCS7 crash with missing EnvelopedContent
7226 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7227 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7228 with missing content and trigger a NULL pointer dereference on parsing.
7230 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7231 structures from untrusted sources are affected. OpenSSL clients and
7232 servers are not affected.
7234 This issue was reported to OpenSSL by Michal Zalewski (Google).
7239 * CMS verify infinite loop with unknown hash function
7241 When verifying a signedData message the CMS code can enter an infinite loop
7242 if presented with an unknown hash function OID. This can be used to perform
7243 denial of service against any system which verifies signedData messages using
7245 This issue was reported to OpenSSL by Johannes Bauer.
7250 * Race condition handling NewSessionTicket
7252 If a NewSessionTicket is received by a multi-threaded client when attempting to
7253 reuse a previous ticket then a race condition can occur potentially leading to
7254 a double free of the ticket data.
7259 * Reject DH handshakes with parameters shorter than 768 bits.
7261 *Kurt Roeckx and Emilia Kasper*
7263 * dhparam: generate 2048-bit parameters by default.
7265 *Kurt Roeckx and Emilia Kasper*
7267 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7269 * Segmentation fault in ASN1_TYPE_cmp fix
7271 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7272 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7273 certificate signature algorithm consistency this can be used to crash any
7274 certificate verification operation and exploited in a DoS attack. Any
7275 application which performs certificate verification is vulnerable including
7276 OpenSSL clients and servers which enable client authentication.
7281 * ASN.1 structure reuse memory corruption fix
7283 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7284 memory corruption via an invalid write. Such reuse is and has been
7285 strongly discouraged and is believed to be rare.
7287 Applications that parse structures containing CHOICE or ANY DEFINED BY
7288 components may be affected. Certificate parsing (d2i_X509 and related
7289 functions) are however not affected. OpenSSL clients and servers are
7295 * PKCS7 NULL pointer dereferences fix
7297 The PKCS#7 parsing code does not handle missing outer ContentInfo
7298 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7299 missing content and trigger a NULL pointer dereference on parsing.
7301 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7302 otherwise parse PKCS#7 structures from untrusted sources are
7303 affected. OpenSSL clients and servers are not affected.
7305 This issue was reported to OpenSSL by Michal Zalewski (Google).
7310 * DoS via reachable assert in SSLv2 servers fix
7312 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7313 servers that both support SSLv2 and enable export cipher suites by sending
7314 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7316 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7317 (OpenSSL development team).
7322 * Use After Free following d2i_ECPrivatekey error fix
7324 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7325 could cause a use after free condition. This, in turn, could cause a double
7326 free in several private key parsing functions (such as d2i_PrivateKey
7327 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7328 for applications that receive EC private keys from untrusted
7329 sources. This scenario is considered rare.
7331 This issue was discovered by the BoringSSL project and fixed in their
7337 * X509_to_X509_REQ NULL pointer deref fix
7339 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7340 the certificate key is invalid. This function is rarely used in practice.
7342 This issue was discovered by Brian Carpenter.
7347 * Removed the export ciphers from the DEFAULT ciphers
7351 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7353 * Build fixes for the Windows and OpenVMS platforms
7355 *Matt Caswell and Richard Levitte*
7357 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7359 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7360 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7361 dereference. This could lead to a Denial Of Service attack. Thanks to
7362 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7367 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7368 dtls1_buffer_record function under certain conditions. In particular this
7369 could occur if an attacker sent repeated DTLS records with the same
7370 sequence number but for the next epoch. The memory leak could be exploited
7371 by an attacker in a Denial of Service attack through memory exhaustion.
7372 Thanks to Chris Mueller for reporting this issue.
7377 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7378 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7379 method would be set to NULL which could later result in a NULL pointer
7380 dereference. Thanks to Frank Schmirler for reporting this issue.
7385 * Abort handshake if server key exchange message is omitted for ephemeral
7388 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7389 reporting this issue.
7394 * Remove non-export ephemeral RSA code on client and server. This code
7395 violated the TLS standard by allowing the use of temporary RSA keys in
7396 non-export ciphersuites and could be used by a server to effectively
7397 downgrade the RSA key length used to a value smaller than the server
7398 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7399 INRIA or reporting this issue.
7404 * Fixed issue where DH client certificates are accepted without verification.
7405 An OpenSSL server will accept a DH certificate for client authentication
7406 without the certificate verify message. This effectively allows a client to
7407 authenticate without the use of a private key. This only affects servers
7408 which trust a client certificate authority which issues certificates
7409 containing DH keys: these are extremely rare and hardly ever encountered.
7410 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7416 * Ensure that the session ID context of an SSL is updated when its
7417 SSL_CTX is updated via SSL_set_SSL_CTX.
7419 The session ID context is typically set from the parent SSL_CTX,
7420 and can vary with the CTX.
7424 * Fix various certificate fingerprint issues.
7426 By using non-DER or invalid encodings outside the signed portion of a
7427 certificate the fingerprint can be changed without breaking the signature.
7428 Although no details of the signed portion of the certificate can be changed
7429 this can cause problems with some applications: e.g. those using the
7430 certificate fingerprint for blacklists.
7432 1. Reject signatures with non zero unused bits.
7434 If the BIT STRING containing the signature has non zero unused bits reject
7435 the signature. All current signature algorithms require zero unused bits.
7437 2. Check certificate algorithm consistency.
7439 Check the AlgorithmIdentifier inside TBS matches the one in the
7440 certificate signature. NB: this will result in signature failure
7441 errors for some broken certificates.
7443 Thanks to Konrad Kraszewski from Google for reporting this issue.
7445 3. Check DSA/ECDSA signatures use DER.
7447 Re-encode DSA/ECDSA signatures and compare with the original received
7448 signature. Return an error if there is a mismatch.
7450 This will reject various cases including garbage after signature
7451 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7452 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7453 (negative or with leading zeroes).
7455 Further analysis was conducted and fixes were developed by Stephen Henson
7456 of the OpenSSL core team.
7462 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7463 results on some platforms, including x86_64. This bug occurs at random
7464 with a very low probability, and is not known to be exploitable in any
7465 way, though its exact impact is difficult to determine. Thanks to Pieter
7466 Wuille (Blockstream) who reported this issue and also suggested an initial
7467 fix. Further analysis was conducted by the OpenSSL development team and
7468 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7469 the OpenSSL core team.
7474 * Do not resume sessions on the server if the negotiated protocol
7475 version does not match the session's version. Resuming with a different
7476 version, while not strictly forbidden by the RFC, is of questionable
7477 sanity and breaks all known clients.
7479 *David Benjamin, Emilia Käsper*
7481 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7482 early CCS messages during renegotiation. (Note that because
7483 renegotiation is encrypted, this early CCS was not exploitable.)
7487 * Tighten client-side session ticket handling during renegotiation:
7488 ensure that the client only accepts a session ticket if the server sends
7489 the extension anew in the ServerHello. Previously, a TLS client would
7490 reuse the old extension state and thus accept a session ticket if one was
7491 announced in the initial ServerHello.
7493 Similarly, ensure that the client requires a session ticket if one
7494 was advertised in the ServerHello. Previously, a TLS client would
7495 ignore a missing NewSessionTicket message.
7499 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7503 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7504 sends a carefully crafted handshake message, to cause OpenSSL to fail
7505 to free up to 64k of memory causing a memory leak. This could be
7506 exploited in a Denial Of Service attack. This issue affects OpenSSL
7507 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7508 whether SRTP is used or configured. Implementations of OpenSSL that
7509 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7511 The fix was developed by the OpenSSL team.
7516 * Session Ticket Memory Leak.
7518 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7519 integrity of that ticket is first verified. In the event of a session
7520 ticket integrity check failing, OpenSSL will fail to free memory
7521 causing a memory leak. By sending a large number of invalid session
7522 tickets an attacker could exploit this issue in a Denial Of Service
7528 * Build option no-ssl3 is incomplete.
7530 When OpenSSL is configured with "no-ssl3" as a build option, servers
7531 could accept and complete an SSL 3.0 handshake, and clients could be
7532 configured to send them.
7535 *Akamai and the OpenSSL team*
7537 * Add support for TLS_FALLBACK_SCSV.
7538 Client applications doing fallback retries should call
7539 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7542 *Adam Langley, Bodo Moeller*
7544 * Add additional DigestInfo checks.
7546 Re-encode DigestInto in DER and check against the original when
7547 verifying RSA signature: this will reject any improperly encoded
7548 DigestInfo structures.
7550 Note: this is a precautionary measure and no attacks are currently known.
7554 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7556 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7557 SRP code can be overrun an internal buffer. Add sanity check that
7558 g, A, B < N to SRP code.
7560 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7561 Group for discovering this issue.
7566 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7567 TLS 1.0 instead of higher protocol versions when the ClientHello message
7568 is badly fragmented. This allows a man-in-the-middle attacker to force a
7569 downgrade to TLS 1.0 even if both the server and the client support a
7570 higher protocol version, by modifying the client's TLS records.
7572 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7573 researching this issue.
7578 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7579 to a denial of service attack. A malicious server can crash the client
7580 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7581 ciphersuite and sending carefully crafted handshake messages.
7583 Thanks to Felix Gröbert (Google) for discovering and researching this
7589 * By sending carefully crafted DTLS packets an attacker could cause openssl
7590 to leak memory. This can be exploited through a Denial of Service attack.
7591 Thanks to Adam Langley for discovering and researching this issue.
7596 * An attacker can force openssl to consume large amounts of memory whilst
7597 processing DTLS handshake messages. This can be exploited through a
7598 Denial of Service attack.
7599 Thanks to Adam Langley for discovering and researching this issue.
7604 * An attacker can force an error condition which causes openssl to crash
7605 whilst processing DTLS packets due to memory being freed twice. This
7606 can be exploited through a Denial of Service attack.
7607 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7613 * If a multithreaded client connects to a malicious server using a resumed
7614 session and the server sends an ec point format extension it could write
7615 up to 255 bytes to freed memory.
7617 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7623 * A malicious server can crash an OpenSSL client with a null pointer
7624 dereference (read) by specifying an SRP ciphersuite even though it was not
7625 properly negotiated with the client. This can be exploited through a
7626 Denial of Service attack.
7628 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7629 discovering and researching this issue.
7634 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7635 X509_name_oneline, X509_name_print_ex et al. to leak some information
7636 from the stack. Applications may be affected if they echo pretty printing
7637 output to the attacker.
7639 Thanks to Ivan Fratric (Google) for discovering this issue.
7642 *Emilia Käsper, and Steve Henson*
7644 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7645 for corner cases. (Certain input points at infinity could lead to
7646 bogus results, with non-infinity inputs mapped to infinity too.)
7650 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7652 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7653 handshake can force the use of weak keying material in OpenSSL
7654 SSL/TLS clients and servers.
7656 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7657 researching this issue. ([CVE-2014-0224])
7659 *KIKUCHI Masashi, Steve Henson*
7661 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7662 OpenSSL DTLS client the code can be made to recurse eventually crashing
7665 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7668 *Imre Rad, Steve Henson*
7670 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7671 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7672 client or server. This is potentially exploitable to run arbitrary
7673 code on a vulnerable client or server.
7675 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7677 *Jüri Aedla, Steve Henson*
7679 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7680 are subject to a denial of service attack.
7682 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7683 this issue. ([CVE-2014-3470])
7685 *Felix Gröbert, Ivan Fratric, Steve Henson*
7687 * Harmonize version and its documentation. -f flag is used to display
7690 *mancha <mancha1@zoho.com>*
7692 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7693 in i2d_ECPrivateKey.
7695 *mancha <mancha1@zoho.com>*
7697 * Fix some double frees. These are not thought to be exploitable.
7699 *mancha <mancha1@zoho.com>*
7701 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7703 * A missing bounds check in the handling of the TLS heartbeat extension
7704 can be used to reveal up to 64k of memory to a connected client or
7707 Thanks for Neel Mehta of Google Security for discovering this bug and to
7708 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7709 preparing the fix ([CVE-2014-0160])
7711 *Adam Langley, Bodo Moeller*
7713 * Fix for the attack described in the paper "Recovering OpenSSL
7714 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7715 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7716 <http://eprint.iacr.org/2014/140>
7718 Thanks to Yuval Yarom and Naomi Benger for discovering this
7719 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7721 *Yuval Yarom and Naomi Benger*
7723 * TLS pad extension: draft-agl-tls-padding-03
7725 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7726 TLS client Hello record length value would otherwise be > 255 and
7727 less that 512 pad with a dummy extension containing zeroes so it
7728 is at least 512 bytes long.
7730 *Adam Langley, Steve Henson*
7732 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7734 * Fix for TLS record tampering bug. A carefully crafted invalid
7735 handshake could crash OpenSSL with a NULL pointer exception.
7736 Thanks to Anton Johansson for reporting this issues.
7739 * Keep original DTLS digest and encryption contexts in retransmission
7740 structures so we can use the previous session parameters if they need
7741 to be resent. ([CVE-2013-6450])
7745 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7746 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7747 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7748 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7749 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7750 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7752 *Rob Stradling, Adam Langley*
7754 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7756 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7757 supporting platforms or when small records were transferred.
7759 *Andy Polyakov, Steve Henson*
7761 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7763 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7765 This addresses the flaw in CBC record processing discovered by
7766 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7767 at: <http://www.isg.rhul.ac.uk/tls/>
7769 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7770 Security Group at Royal Holloway, University of London
7771 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7772 Emilia Käsper for the initial patch.
7775 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7777 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7778 ciphersuites which can be exploited in a denial of service attack.
7779 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7780 and detecting this bug and to Wolfgang Ettlinger
7781 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7786 * Return an error when checking OCSP signatures when key is NULL.
7787 This fixes a DoS attack. ([CVE-2013-0166])
7791 * Make openssl verify return errors.
7793 *Chris Palmer <palmer@google.com> and Ben Laurie*
7795 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7796 the right response is stapled. Also change SSL_get_certificate()
7797 so it returns the certificate actually sent.
7798 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7800 *Rob Stradling <rob.stradling@comodo.com>*
7802 * Fix possible deadlock when decoding public keys.
7806 * Don't use TLS 1.0 record version number in initial client hello
7811 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7813 * Sanity check record length before skipping explicit IV in TLS
7814 1.2, 1.1 and DTLS to fix DoS attack.
7816 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7817 fuzzing as a service testing platform.
7822 * Initialise tkeylen properly when encrypting CMS messages.
7823 Thanks to Solar Designer of Openwall for reporting this issue.
7827 * In FIPS mode don't try to use composite ciphers as they are not
7832 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7834 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7835 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7836 mean any application compiled against OpenSSL 1.0.0 headers setting
7837 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7838 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7839 0x10000000L Any application which was previously compiled against
7840 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7841 will need to be recompiled as a result. Letting be results in
7842 inability to disable specifically TLS 1.1 and in client context,
7843 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7847 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7848 disable just protocol X, but all protocols above X *if* there are
7849 protocols *below* X still enabled. In more practical terms it means
7850 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7851 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7852 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7857 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7859 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7860 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7861 in CRYPTO_realloc_clean.
7863 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7864 issue and to Adam Langley <agl@chromium.org> for fixing it.
7867 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7869 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7873 * Workarounds for some broken servers that "hang" if a client hello
7874 record length exceeds 255 bytes.
7876 1. Do not use record version number > TLS 1.0 in initial client
7877 hello: some (but not all) hanging servers will now work.
7878 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7879 the number of ciphers sent in the client hello. This should be
7880 set to an even number, such as 50, for example by passing:
7881 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7882 Most broken servers should now work.
7883 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7884 TLS 1.2 client support entirely.
7888 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7892 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7894 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7895 STRING form instead of a DigestInfo.
7899 * The format used for MDC2 RSA signatures is inconsistent between EVP
7900 and the RSA_sign/RSA_verify functions. This was made more apparent when
7901 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7902 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7903 the correct format in RSA_verify so both forms transparently work.
7907 * Some servers which support TLS 1.0 can choke if we initially indicate
7908 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7909 encrypted premaster secret. As a workaround use the maximum permitted
7910 client version in client hello, this should keep such servers happy
7911 and still work with previous versions of OpenSSL.
7915 * Add support for TLS/DTLS heartbeats.
7917 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7919 * Add support for SCTP.
7921 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7923 * Improved PRNG seeding for VOS.
7925 *Paul Green <Paul.Green@stratus.com>*
7927 * Extensive assembler packs updates, most notably:
7929 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7930 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7931 - x86_64: bit-sliced AES implementation;
7932 - ARM: NEON support, contemporary platforms optimizations;
7933 - s390x: z196 support;
7934 - `*`: GHASH and GF(2^m) multiplication implementations;
7938 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7939 (removal of unnecessary code)
7941 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7943 * Add TLS key material exporter from RFC 5705.
7947 * Add DTLS-SRTP negotiation from RFC 5764.
7951 * Add Next Protocol Negotiation,
7952 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7953 disabled with a no-npn flag to config or Configure. Code donated
7956 *Adam Langley <agl@google.com> and Ben Laurie*
7958 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7959 NIST-P256, NIST-P521, with constant-time single point multiplication on
7960 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7961 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7962 Code made available under Apache License version 2.0.
7964 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7965 line to include this in your build of OpenSSL, and run "make depend" (or
7966 "make update"). This enables the following EC_METHODs:
7968 EC_GFp_nistp224_method()
7969 EC_GFp_nistp256_method()
7970 EC_GFp_nistp521_method()
7972 EC_GROUP_new_by_curve_name() will automatically use these (while
7973 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7976 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7978 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7979 all platforms. Move ssize_t definition from e_os.h to the public
7980 header file e_os2.h as it now appears in public header file cms.h
7984 * New -sigopt option to the ca, req and x509 utilities. Additional
7985 signature parameters can be passed using this option and in
7990 * Add RSA PSS signing function. This will generate and set the
7991 appropriate AlgorithmIdentifiers for PSS based on those in the
7992 corresponding EVP_MD_CTX structure. No application support yet.
7996 * Support for companion algorithm specific ASN1 signing routines.
7997 New function ASN1_item_sign_ctx() signs a pre-initialised
7998 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7999 the appropriate parameters.
8003 * Add new algorithm specific ASN1 verification initialisation function
8004 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8005 handling will be the same no matter what EVP_PKEY_METHOD is used.
8006 Add a PSS handler to support verification of PSS signatures: checked
8007 against a number of sample certificates.
8011 * Add signature printing for PSS. Add PSS OIDs.
8013 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8015 * Add algorithm specific signature printing. An individual ASN1 method
8016 can now print out signatures instead of the standard hex dump.
8018 More complex signatures (e.g. PSS) can print out more meaningful
8019 information. Include DSA version that prints out the signature
8024 * Password based recipient info support for CMS library: implementing
8029 * Split password based encryption into PBES2 and PBKDF2 functions. This
8030 neatly separates the code into cipher and PBE sections and is required
8031 for some algorithms that split PBES2 into separate pieces (such as
8032 password based CMS).
8036 * Session-handling fixes:
8037 - Fix handling of connections that are resuming with a session ID,
8038 but also support Session Tickets.
8039 - Fix a bug that suppressed issuing of a new ticket if the client
8040 presented a ticket with an expired session.
8041 - Try to set the ticket lifetime hint to something reasonable.
8042 - Make tickets shorter by excluding irrelevant information.
8043 - On the client side, don't ignore renewed tickets.
8045 *Adam Langley, Bodo Moeller (Google)*
8047 * Fix PSK session representation.
8051 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8053 This work was sponsored by Intel.
8057 * Add GCM support to TLS library. Some custom code is needed to split
8058 the IV between the fixed (from PRF) and explicit (from TLS record)
8059 portions. This adds all GCM ciphersuites supported by RFC5288 and
8060 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8061 add a special AESGCM string for GCM only.
8065 * Expand range of ctrls for AES GCM. Permit setting invocation
8066 field on decrypt and retrieval of invocation field only on encrypt.
8070 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8071 As required by RFC5289 these ciphersuites cannot be used if for
8072 versions of TLS earlier than 1.2.
8076 * For FIPS capable OpenSSL interpret a NULL default public key method
8077 as unset and return the appropriate default but do *not* set the default.
8078 This means we can return the appropriate method in applications that
8079 switch between FIPS and non-FIPS modes.
8083 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8084 ENGINE is used then we cannot handle that in the FIPS module so we
8085 keep original code iff non-FIPS operations are allowed.
8089 * Add -attime option to openssl utilities.
8091 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8093 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8097 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8098 FIPS EC methods unconditionally for now.
8102 * New build option no-ec2m to disable characteristic 2 code.
8106 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8107 all cases can be covered as some introduce binary incompatibilities.
8111 * Redirect RSA operations to FIPS module including keygen,
8112 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8116 * Add similar low-level API blocking to ciphers.
8120 * low-level digest APIs are not approved in FIPS mode: any attempt
8121 to use these will cause a fatal error. Applications that *really* want
8122 to use them can use the `private_*` version instead.
8126 * Redirect cipher operations to FIPS module for FIPS builds.
8130 * Redirect digest operations to FIPS module for FIPS builds.
8134 * Update build system to add "fips" flag which will link in fipscanister.o
8135 for static and shared library builds embedding a signature if needed.
8139 * Output TLS supported curves in preference order instead of numerical
8140 order. This is currently hardcoded for the highest order curves first.
8141 This should be configurable so applications can judge speed vs strength.
8145 * Add TLS v1.2 server support for client authentication.
8149 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8154 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8155 FIPS modules versions.
8159 * Add TLS v1.2 client side support for client authentication. Keep cache
8160 of handshake records longer as we don't know the hash algorithm to use
8161 until after the certificate request message is received.
8165 * Initial TLS v1.2 client support. Add a default signature algorithms
8166 extension including all the algorithms we support. Parse new signature
8167 format in client key exchange. Relax some ECC signing restrictions for
8168 TLS v1.2 as indicated in RFC5246.
8172 * Add server support for TLS v1.2 signature algorithms extension. Switch
8173 to new signature format when needed using client digest preference.
8174 All server ciphersuites should now work correctly in TLS v1.2. No client
8175 support yet and no support for client certificates.
8179 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8180 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8181 ciphersuites. At present only RSA key exchange ciphersuites work with
8182 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8183 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8184 and version checking.
8188 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8189 with this defined it will not be affected by any changes to ssl internal
8190 structures. Add several utility functions to allow openssl application
8191 to work with OPENSSL_NO_SSL_INTERN defined.
8195 * A long standing patch to add support for SRP from EdelWeb (Peter
8196 Sylvester and Christophe Renou) was integrated.
8197 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8198 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8201 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8205 * Permit abbreviated handshakes when renegotiating using the function
8206 SSL_renegotiate_abbreviated().
8208 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8210 * Add call to ENGINE_register_all_complete() to
8211 ENGINE_load_builtin_engines(), so some implementations get used
8212 automatically instead of needing explicit application support.
8216 * Add support for TLS key exporter as described in RFC5705.
8218 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8220 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8221 a few changes are required:
8223 Add SSL_OP_NO_TLSv1_1 flag.
8224 Add TLSv1_1 methods.
8225 Update version checking logic to handle version 1.1.
8226 Add explicit IV handling (ported from DTLS code).
8227 Add command line options to s_client/s_server.
8234 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8236 * X509_ATTRIBUTE memory leak
8238 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8239 memory. This structure is used by the PKCS#7 and CMS routines so any
8240 application which reads PKCS#7 or CMS data from untrusted sources is
8241 affected. SSL/TLS is not affected.
8243 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8249 * Race condition handling PSK identify hint
8251 If PSK identity hints are received by a multi-threaded client then
8252 the values are wrongly updated in the parent SSL_CTX structure. This can
8253 result in a race condition potentially leading to a double free of the
8259 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8261 * Malformed ECParameters causes infinite loop
8263 When processing an ECParameters structure OpenSSL enters an infinite loop
8264 if the curve specified is over a specially malformed binary polynomial
8267 This can be used to perform denial of service against any
8268 system which processes public keys, certificate requests or
8269 certificates. This includes TLS clients and TLS servers with
8270 client authentication enabled.
8272 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8277 * Exploitable out-of-bounds read in X509_cmp_time
8279 X509_cmp_time does not properly check the length of the ASN1_TIME
8280 string and can read a few bytes out of bounds. In addition,
8281 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8284 An attacker can use this to craft malformed certificates and CRLs of
8285 various sizes and potentially cause a segmentation fault, resulting in
8286 a DoS on applications that verify certificates or CRLs. TLS clients
8287 that verify CRLs are affected. TLS clients and servers with client
8288 authentication enabled may be affected if they use custom verification
8291 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8292 independently by Hanno Böck.
8297 * PKCS7 crash with missing EnvelopedContent
8299 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8300 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8301 with missing content and trigger a NULL pointer dereference on parsing.
8303 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8304 structures from untrusted sources are affected. OpenSSL clients and
8305 servers are not affected.
8307 This issue was reported to OpenSSL by Michal Zalewski (Google).
8312 * CMS verify infinite loop with unknown hash function
8314 When verifying a signedData message the CMS code can enter an infinite loop
8315 if presented with an unknown hash function OID. This can be used to perform
8316 denial of service against any system which verifies signedData messages using
8318 This issue was reported to OpenSSL by Johannes Bauer.
8323 * Race condition handling NewSessionTicket
8325 If a NewSessionTicket is received by a multi-threaded client when attempting to
8326 reuse a previous ticket then a race condition can occur potentially leading to
8327 a double free of the ticket data.
8332 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8334 * Segmentation fault in ASN1_TYPE_cmp fix
8336 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8337 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8338 certificate signature algorithm consistency this can be used to crash any
8339 certificate verification operation and exploited in a DoS attack. Any
8340 application which performs certificate verification is vulnerable including
8341 OpenSSL clients and servers which enable client authentication.
8346 * ASN.1 structure reuse memory corruption fix
8348 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8349 memory corruption via an invalid write. Such reuse is and has been
8350 strongly discouraged and is believed to be rare.
8352 Applications that parse structures containing CHOICE or ANY DEFINED BY
8353 components may be affected. Certificate parsing (d2i_X509 and related
8354 functions) are however not affected. OpenSSL clients and servers are
8360 * PKCS7 NULL pointer dereferences fix
8362 The PKCS#7 parsing code does not handle missing outer ContentInfo
8363 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8364 missing content and trigger a NULL pointer dereference on parsing.
8366 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8367 otherwise parse PKCS#7 structures from untrusted sources are
8368 affected. OpenSSL clients and servers are not affected.
8370 This issue was reported to OpenSSL by Michal Zalewski (Google).
8375 * DoS via reachable assert in SSLv2 servers fix
8377 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8378 servers that both support SSLv2 and enable export cipher suites by sending
8379 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8381 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8382 (OpenSSL development team).
8387 * Use After Free following d2i_ECPrivatekey error fix
8389 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8390 could cause a use after free condition. This, in turn, could cause a double
8391 free in several private key parsing functions (such as d2i_PrivateKey
8392 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8393 for applications that receive EC private keys from untrusted
8394 sources. This scenario is considered rare.
8396 This issue was discovered by the BoringSSL project and fixed in their
8402 * X509_to_X509_REQ NULL pointer deref fix
8404 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8405 the certificate key is invalid. This function is rarely used in practice.
8407 This issue was discovered by Brian Carpenter.
8412 * Removed the export ciphers from the DEFAULT ciphers
8416 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8418 * Build fixes for the Windows and OpenVMS platforms
8420 *Matt Caswell and Richard Levitte*
8422 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8424 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8425 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8426 dereference. This could lead to a Denial Of Service attack. Thanks to
8427 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8432 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8433 dtls1_buffer_record function under certain conditions. In particular this
8434 could occur if an attacker sent repeated DTLS records with the same
8435 sequence number but for the next epoch. The memory leak could be exploited
8436 by an attacker in a Denial of Service attack through memory exhaustion.
8437 Thanks to Chris Mueller for reporting this issue.
8442 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8443 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8444 method would be set to NULL which could later result in a NULL pointer
8445 dereference. Thanks to Frank Schmirler for reporting this issue.
8450 * Abort handshake if server key exchange message is omitted for ephemeral
8453 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8454 reporting this issue.
8459 * Remove non-export ephemeral RSA code on client and server. This code
8460 violated the TLS standard by allowing the use of temporary RSA keys in
8461 non-export ciphersuites and could be used by a server to effectively
8462 downgrade the RSA key length used to a value smaller than the server
8463 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8464 INRIA or reporting this issue.
8469 * Fixed issue where DH client certificates are accepted without verification.
8470 An OpenSSL server will accept a DH certificate for client authentication
8471 without the certificate verify message. This effectively allows a client to
8472 authenticate without the use of a private key. This only affects servers
8473 which trust a client certificate authority which issues certificates
8474 containing DH keys: these are extremely rare and hardly ever encountered.
8475 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8481 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8482 results on some platforms, including x86_64. This bug occurs at random
8483 with a very low probability, and is not known to be exploitable in any
8484 way, though its exact impact is difficult to determine. Thanks to Pieter
8485 Wuille (Blockstream) who reported this issue and also suggested an initial
8486 fix. Further analysis was conducted by the OpenSSL development team and
8487 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8488 the OpenSSL core team.
8493 * Fix various certificate fingerprint issues.
8495 By using non-DER or invalid encodings outside the signed portion of a
8496 certificate the fingerprint can be changed without breaking the signature.
8497 Although no details of the signed portion of the certificate can be changed
8498 this can cause problems with some applications: e.g. those using the
8499 certificate fingerprint for blacklists.
8501 1. Reject signatures with non zero unused bits.
8503 If the BIT STRING containing the signature has non zero unused bits reject
8504 the signature. All current signature algorithms require zero unused bits.
8506 2. Check certificate algorithm consistency.
8508 Check the AlgorithmIdentifier inside TBS matches the one in the
8509 certificate signature. NB: this will result in signature failure
8510 errors for some broken certificates.
8512 Thanks to Konrad Kraszewski from Google for reporting this issue.
8514 3. Check DSA/ECDSA signatures use DER.
8516 Reencode DSA/ECDSA signatures and compare with the original received
8517 signature. Return an error if there is a mismatch.
8519 This will reject various cases including garbage after signature
8520 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8521 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8522 (negative or with leading zeroes).
8524 Further analysis was conducted and fixes were developed by Stephen Henson
8525 of the OpenSSL core team.
8531 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8533 * Session Ticket Memory Leak.
8535 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8536 integrity of that ticket is first verified. In the event of a session
8537 ticket integrity check failing, OpenSSL will fail to free memory
8538 causing a memory leak. By sending a large number of invalid session
8539 tickets an attacker could exploit this issue in a Denial Of Service
8545 * Build option no-ssl3 is incomplete.
8547 When OpenSSL is configured with "no-ssl3" as a build option, servers
8548 could accept and complete an SSL 3.0 handshake, and clients could be
8549 configured to send them.
8552 *Akamai and the OpenSSL team*
8554 * Add support for TLS_FALLBACK_SCSV.
8555 Client applications doing fallback retries should call
8556 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8559 *Adam Langley, Bodo Moeller*
8561 * Add additional DigestInfo checks.
8563 Reencode DigestInto in DER and check against the original when
8564 verifying RSA signature: this will reject any improperly encoded
8565 DigestInfo structures.
8567 Note: this is a precautionary measure and no attacks are currently known.
8571 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8573 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8574 to a denial of service attack. A malicious server can crash the client
8575 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8576 ciphersuite and sending carefully crafted handshake messages.
8578 Thanks to Felix Gröbert (Google) for discovering and researching this
8584 * By sending carefully crafted DTLS packets an attacker could cause openssl
8585 to leak memory. This can be exploited through a Denial of Service attack.
8586 Thanks to Adam Langley for discovering and researching this issue.
8591 * An attacker can force openssl to consume large amounts of memory whilst
8592 processing DTLS handshake messages. This can be exploited through a
8593 Denial of Service attack.
8594 Thanks to Adam Langley for discovering and researching this issue.
8599 * An attacker can force an error condition which causes openssl to crash
8600 whilst processing DTLS packets due to memory being freed twice. This
8601 can be exploited through a Denial of Service attack.
8602 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8608 * If a multithreaded client connects to a malicious server using a resumed
8609 session and the server sends an ec point format extension it could write
8610 up to 255 bytes to freed memory.
8612 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8618 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8619 X509_name_oneline, X509_name_print_ex et al. to leak some information
8620 from the stack. Applications may be affected if they echo pretty printing
8621 output to the attacker.
8623 Thanks to Ivan Fratric (Google) for discovering this issue.
8626 *Emilia Käsper, and Steve Henson*
8628 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8629 for corner cases. (Certain input points at infinity could lead to
8630 bogus results, with non-infinity inputs mapped to infinity too.)
8634 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8636 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8637 handshake can force the use of weak keying material in OpenSSL
8638 SSL/TLS clients and servers.
8640 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8641 researching this issue. ([CVE-2014-0224])
8643 *KIKUCHI Masashi, Steve Henson*
8645 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8646 OpenSSL DTLS client the code can be made to recurse eventually crashing
8649 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8652 *Imre Rad, Steve Henson*
8654 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8655 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8656 client or server. This is potentially exploitable to run arbitrary
8657 code on a vulnerable client or server.
8659 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8661 *Jüri Aedla, Steve Henson*
8663 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8664 are subject to a denial of service attack.
8666 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8667 this issue. ([CVE-2014-3470])
8669 *Felix Gröbert, Ivan Fratric, Steve Henson*
8671 * Harmonize version and its documentation. -f flag is used to display
8674 *mancha <mancha1@zoho.com>*
8676 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8677 in i2d_ECPrivateKey.
8679 *mancha <mancha1@zoho.com>*
8681 * Fix some double frees. These are not thought to be exploitable.
8683 *mancha <mancha1@zoho.com>*
8685 * Fix for the attack described in the paper "Recovering OpenSSL
8686 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8687 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8688 <http://eprint.iacr.org/2014/140>
8690 Thanks to Yuval Yarom and Naomi Benger for discovering this
8691 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8693 *Yuval Yarom and Naomi Benger*
8695 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8697 * Keep original DTLS digest and encryption contexts in retransmission
8698 structures so we can use the previous session parameters if they need
8699 to be resent. ([CVE-2013-6450])
8703 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8704 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8705 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8706 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8707 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8708 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8710 *Rob Stradling, Adam Langley*
8712 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8714 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8716 This addresses the flaw in CBC record processing discovered by
8717 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8718 at: <http://www.isg.rhul.ac.uk/tls/>
8720 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8721 Security Group at Royal Holloway, University of London
8722 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8723 Emilia Käsper for the initial patch.
8726 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8728 * Return an error when checking OCSP signatures when key is NULL.
8729 This fixes a DoS attack. ([CVE-2013-0166])
8733 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8734 the right response is stapled. Also change SSL_get_certificate()
8735 so it returns the certificate actually sent.
8736 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8737 (This is a backport)
8739 *Rob Stradling <rob.stradling@comodo.com>*
8741 * Fix possible deadlock when decoding public keys.
8745 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8747 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8750 * Sanity check record length before skipping explicit IV in DTLS
8753 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8754 fuzzing as a service testing platform.
8759 * Initialise tkeylen properly when encrypting CMS messages.
8760 Thanks to Solar Designer of Openwall for reporting this issue.
8764 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8766 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8767 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8768 in CRYPTO_realloc_clean.
8770 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8771 issue and to Adam Langley <agl@chromium.org> for fixing it.
8774 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8776 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8778 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8779 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8780 content decryption and always return the same error. Note: this attack
8781 needs on average 2^20 messages so it only affects automated senders. The
8782 old behaviour can be re-enabled in the CMS code by setting the
8783 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8784 an MMA defence is not necessary.
8785 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8786 this issue. ([CVE-2012-0884])
8790 * Fix CVE-2011-4619: make sure we really are receiving a
8791 client hello before rejecting multiple SGC restarts. Thanks to
8792 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8796 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8798 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8799 Thanks to Antonio Martin, Enterprise Secure Access Research and
8800 Development, Cisco Systems, Inc. for discovering this bug and
8801 preparing a fix. ([CVE-2012-0050])
8805 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8807 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8808 of the Vaudenay padding oracle attack on CBC mode encryption
8809 which enables an efficient plaintext recovery attack against
8810 the OpenSSL implementation of DTLS. Their attack exploits timing
8811 differences arising during decryption processing. A research
8812 paper describing this attack can be found at:
8813 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8814 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8815 Security Group at Royal Holloway, University of London
8816 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8817 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8818 for preparing the fix. ([CVE-2011-4108])
8820 *Robin Seggelmann, Michael Tuexen*
8822 * Clear bytes used for block padding of SSL 3.0 records.
8825 *Adam Langley (Google)*
8827 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8828 Kadianakis <desnacked@gmail.com> for discovering this issue and
8829 Adam Langley for preparing the fix. ([CVE-2011-4619])
8831 *Adam Langley (Google)*
8833 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8835 *Andrey Kulikov <amdeich@gmail.com>*
8837 * Prevent malformed RFC3779 data triggering an assertion failure.
8838 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8839 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8841 *Rob Austein <sra@hactrn.net>*
8843 * Improved PRNG seeding for VOS.
8845 *Paul Green <Paul.Green@stratus.com>*
8847 * Fix ssl_ciph.c set-up race.
8849 *Adam Langley (Google)*
8851 * Fix spurious failures in ecdsatest.c.
8853 *Emilia Käsper (Google)*
8855 * Fix the BIO_f_buffer() implementation (which was mixing different
8856 interpretations of the `..._len` fields).
8858 *Adam Langley (Google)*
8860 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8861 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8862 threads won't reuse the same blinding coefficients.
8864 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8865 lock to call BN_BLINDING_invert_ex, and avoids one use of
8866 BN_BLINDING_update for each BN_BLINDING structure (previously,
8867 the last update always remained unused).
8869 *Emilia Käsper (Google)*
8871 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8873 *Bob Buckholz (Google)*
8875 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8877 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8878 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8880 *Kaspar Brand <ossl@velox.ch>*
8882 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8883 for multi-threaded use of ECDH. ([CVE-2011-3210])
8885 *Adam Langley (Google)*
8887 * Fix x509_name_ex_d2i memory leak on bad inputs.
8891 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8892 signature public key algorithm by using OID xref utilities instead.
8893 Before this you could only use some ECC ciphersuites with SHA1 only.
8897 * Add protection against ECDSA timing attacks as mentioned in the paper
8898 by Billy Bob Brumley and Nicola Tuveri, see:
8899 <http://eprint.iacr.org/2011/232.pdf>
8901 *Billy Bob Brumley and Nicola Tuveri*
8903 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8905 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8907 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8909 * Fix bug in string printing code: if *any* escaping is enabled we must
8910 escape the escape character (backslash) or the resulting string is
8915 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8917 * Disable code workaround for ancient and obsolete Netscape browsers
8918 and servers: an attacker can use it in a ciphersuite downgrade attack.
8919 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8923 * Fixed J-PAKE implementation error, originally discovered by
8924 Sebastien Martini, further info and confirmation from Stefan
8925 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8929 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8931 * Fix extension code to avoid race conditions which can result in a buffer
8932 overrun vulnerability: resumed sessions must not be modified as they can
8933 be shared by multiple threads. CVE-2010-3864
8937 * Fix WIN32 build system to correctly link an ENGINE directory into
8942 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8944 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8947 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8949 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8951 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8952 context. The operation can be customised via the ctrl mechanism in
8953 case ENGINEs want to include additional functionality.
8957 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8961 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8962 output hashes compatible with older versions of OpenSSL.
8964 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8966 * Fix compression algorithm handling: if resuming a session use the
8967 compression algorithm of the resumed session instead of determining
8968 it from client hello again. Don't allow server to change algorithm.
8972 * Add load_crls() function to commands tidying load_certs() too. Add option
8973 to verify utility to allow additional CRLs to be included.
8977 * Update OCSP request code to permit adding custom headers to the request:
8978 some responders need this.
8982 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8985 *Julia Lawall <julia@diku.dk>*
8987 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8988 needlessly dereferenced structures, used obsolete functions and
8989 didn't handle all updated verify codes correctly.
8993 * Disable MD2 in the default configuration.
8997 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8998 indicate the initial BIO being pushed or popped. This makes it possible
8999 to determine whether the BIO is the one explicitly called or as a result
9000 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9001 it handles reference counts correctly and doesn't zero out the I/O bio
9002 when it is not being explicitly popped. WARNING: applications which
9003 included workarounds for the old buggy behaviour will need to be modified
9004 or they could free up already freed BIOs.
9008 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9009 renaming to all platforms (within the 0.9.8 branch, this was
9010 done conditionally on Netware platforms to avoid a name clash).
9012 *Guenter <lists@gknw.net>*
9014 * Add ECDHE and PSK support to DTLS.
9016 *Michael Tuexen <tuexen@fh-muenster.de>*
9018 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9023 * Add "missing" function EVP_MD_flags() (without this the only way to
9024 retrieve a digest flags is by accessing the structure directly. Update
9025 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9026 or cipher is registered as in the "from" argument. Print out all
9027 registered digests in the dgst usage message instead of manually
9028 attempting to work them out.
9032 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9033 this allows the use of compression and extensions. Change default cipher
9034 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9035 by default unless an application cipher string requests it.
9039 * Alter match criteria in PKCS12_parse(). It used to try to use local
9040 key ids to find matching certificates and keys but some PKCS#12 files
9041 don't follow the (somewhat unwritten) rules and this strategy fails.
9042 Now just gather all certificates together and the first private key
9043 then look for the first certificate that matches the key.
9047 * Support use of registered digest and cipher names for dgst and cipher
9048 commands instead of having to add each one as a special case. So now
9055 openssl dgst -sha256 foo
9057 and this works for ENGINE based algorithms too.
9061 * Update Gost ENGINE to support parameter files.
9063 *Victor B. Wagner <vitus@cryptocom.ru>*
9065 * Support GeneralizedTime in ca utility.
9067 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9069 * Enhance the hash format used for certificate directory links. The new
9070 form uses the canonical encoding (meaning equivalent names will work
9071 even if they aren't identical) and uses SHA1 instead of MD5. This form
9072 is incompatible with the older format and as a result c_rehash should
9073 be used to rebuild symbolic links.
9077 * Make PKCS#8 the default write format for private keys, replacing the
9078 traditional format. This form is standardised, more secure and doesn't
9079 include an implicit MD5 dependency.
9083 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9084 committed to OpenSSL should pass this lot as a minimum.
9088 * Add session ticket override functionality for use by EAP-FAST.
9090 *Jouni Malinen <j@w1.fi>*
9092 * Modify HMAC functions to return a value. Since these can be implemented
9093 in an ENGINE errors can occur.
9097 * Type-checked OBJ_bsearch_ex.
9101 * Type-checked OBJ_bsearch. Also some constification necessitated
9102 by type-checking. Still to come: TXT_DB, bsearch(?),
9103 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9108 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9109 seconds to a tm structure directly, instead of going through OS
9110 specific date routines. This avoids any issues with OS routines such
9111 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9112 and X509_time_adj_ex() to cover the extended range. The existing
9113 X509_time_adj() is still usable and will no longer have any date issues.
9117 * Delta CRL support. New use deltas option which will attempt to locate
9118 and search any appropriate delta CRLs available.
9120 This work was sponsored by Google.
9124 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9125 code and add additional score elements. Validate alternate CRL paths
9126 as part of the CRL checking and indicate a new error "CRL path validation
9127 error" in this case. Applications wanting additional details can use
9128 the verify callback and check the new "parent" field. If this is not
9129 NULL CRL path validation is taking place. Existing applications won't
9130 see this because it requires extended CRL support which is off by
9133 This work was sponsored by Google.
9137 * Support for freshest CRL extension.
9139 This work was sponsored by Google.
9143 * Initial indirect CRL support. Currently only supported in the CRLs
9144 passed directly and not via lookup. Process certificate issuer
9145 CRL entry extension and lookup CRL entries by bother issuer name
9146 and serial number. Check and process CRL issuer entry in IDP extension.
9148 This work was sponsored by Google.
9152 * Add support for distinct certificate and CRL paths. The CRL issuer
9153 certificate is validated separately in this case. Only enabled if
9154 an extended CRL support flag is set: this flag will enable additional
9155 CRL functionality in future.
9157 This work was sponsored by Google.
9161 * Add support for policy mappings extension.
9163 This work was sponsored by Google.
9167 * Fixes to pathlength constraint, self issued certificate handling,
9168 policy processing to align with RFC3280 and PKITS tests.
9170 This work was sponsored by Google.
9174 * Support for name constraints certificate extension. DN, email, DNS
9175 and URI types are currently supported.
9177 This work was sponsored by Google.
9181 * To cater for systems that provide a pointer-based thread ID rather
9182 than numeric, deprecate the current numeric thread ID mechanism and
9183 replace it with a structure and associated callback type. This
9184 mechanism allows a numeric "hash" to be extracted from a thread ID in
9185 either case, and on platforms where pointers are larger than 'long',
9186 mixing is done to help ensure the numeric 'hash' is usable even if it
9187 can't be guaranteed unique. The default mechanism is to use "&errno"
9188 as a pointer-based thread ID to distinguish between threads.
9190 Applications that want to provide their own thread IDs should now use
9191 CRYPTO_THREADID_set_callback() to register a callback that will call
9192 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9194 Note that ERR_remove_state() is now deprecated, because it is tied
9195 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9196 to free the current thread's error state should be replaced by
9197 ERR_remove_thread_state(NULL).
9199 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9200 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9201 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9202 application was previously providing a numeric thread callback that
9203 was inappropriate for distinguishing threads, then uniqueness might
9204 have been obtained with &errno that happened immediately in the
9205 intermediate development versions of OpenSSL; this is no longer the
9206 case, the numeric thread callback will now override the automatic use
9209 *Geoff Thorpe, with help from Bodo Moeller*
9211 * Initial support for different CRL issuing certificates. This covers a
9212 simple case where the self issued certificates in the chain exist and
9213 the real CRL issuer is higher in the existing chain.
9215 This work was sponsored by Google.
9219 * Removed effectively defunct crypto/store from the build.
9223 * Revamp of STACK to provide stronger type-checking. Still to come:
9224 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9225 ASN1_STRING, CONF_VALUE.
9229 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9230 RAM on SSL connections. This option can save about 34k per idle SSL.
9234 * Revamp of LHASH to provide stronger type-checking. Still to come:
9235 STACK, TXT_DB, bsearch, qsort.
9239 * Initial support for Cryptographic Message Syntax (aka CMS) based
9240 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9241 support for data, signedData, compressedData, digestedData and
9242 encryptedData, envelopedData types included. Scripts to check against
9243 RFC4134 examples draft and interop and consistency checks of many
9244 content types and variants.
9248 * Add options to enc utility to support use of zlib compression BIO.
9252 * Extend mk1mf to support importing of options and assembly language
9253 files from Configure script, currently only included in VC-WIN32.
9254 The assembly language rules can now optionally generate the source
9255 files from the associated perl scripts.
9259 * Implement remaining functionality needed to support GOST ciphersuites.
9260 Interop testing has been performed using CryptoPro implementations.
9262 *Victor B. Wagner <vitus@cryptocom.ru>*
9264 * s390x assembler pack.
9268 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9273 * Implement Opaque PRF Input TLS extension as specified in
9274 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9275 official specification yet and no extension type assignment by
9276 IANA exists, this extension (for now) will have to be explicitly
9277 enabled when building OpenSSL by providing the extension number
9278 to use. For example, specify an option
9280 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9282 to the "config" or "Configure" script to enable the extension,
9283 assuming extension number 0x9527 (which is a completely arbitrary
9284 and unofficial assignment based on the MD5 hash of the Internet
9285 Draft). Note that by doing so, you potentially lose
9286 interoperability with other TLS implementations since these might
9287 be using the same extension number for other purposes.
9289 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9290 opaque PRF input value to use in the handshake. This will create
9291 an internal copy of the length-'len' string at 'src', and will
9292 return non-zero for success.
9294 To get more control and flexibility, provide a callback function
9297 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9298 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9302 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9305 Callback function 'cb' will be called in handshakes, and is
9306 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9307 Argument 'arg' is for application purposes (the value as given to
9308 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9309 be provided to the callback function). The callback function
9310 has to return non-zero to report success: usually 1 to use opaque
9311 PRF input just if possible, or 2 to enforce use of the opaque PRF
9312 input. In the latter case, the library will abort the handshake
9313 if opaque PRF input is not successfully negotiated.
9315 Arguments 'peerinput' and 'len' given to the callback function
9316 will always be NULL and 0 in the case of a client. A server will
9317 see the client's opaque PRF input through these variables if
9318 available (NULL and 0 otherwise). Note that if the server
9319 provides an opaque PRF input, the length must be the same as the
9320 length of the client's opaque PRF input.
9322 Note that the callback function will only be called when creating
9323 a new session (session resumption can resume whatever was
9324 previously negotiated), and will not be called in SSL 2.0
9325 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9326 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9327 for applications that need to enforce opaque PRF input.
9331 * Update ssl code to support digests other than SHA1+MD5 for handshake
9334 *Victor B. Wagner <vitus@cryptocom.ru>*
9336 * Add RFC4507 support to OpenSSL. This includes the corrections in
9337 RFC4507bis. The encrypted ticket format is an encrypted encoded
9338 SSL_SESSION structure, that way new session features are automatically
9341 If a client application caches session in an SSL_SESSION structure
9342 support is transparent because tickets are now stored in the encoded
9345 The SSL_CTX structure automatically generates keys for ticket
9346 protection in servers so again support should be possible
9347 with no application modification.
9349 If a client or server wishes to disable RFC4507 support then the option
9350 SSL_OP_NO_TICKET can be set.
9352 Add a TLS extension debugging callback to allow the contents of any client
9353 or server extensions to be examined.
9355 This work was sponsored by Google.
9359 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9360 OpenSSL should now compile cleanly on gcc 4.2
9362 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9364 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9365 support including streaming MAC support: this is required for GOST
9366 ciphersuite support.
9368 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9370 * Add option -stream to use PKCS#7 streaming in smime utility. New
9371 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9372 to output in BER and PEM format.
9376 * Experimental support for use of HMAC via EVP_PKEY interface. This
9377 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9378 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9379 ENGINE support for HMAC keys which are unextractable. New -mac and
9380 -macopt options to dgst utility.
9384 * New option -sigopt to dgst utility. Update dgst to use
9385 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9386 alternative signing parameters such as X9.31 or PSS in the dgst
9391 * Change ssl_cipher_apply_rule(), the internal function that does
9392 the work each time a ciphersuite string requests enabling
9393 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9394 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9395 the order of disabled ciphersuites such that those ciphersuites
9396 that most recently went from enabled to disabled not only stay
9397 in order with respect to each other, but also have higher priority
9398 than other disabled ciphersuites the next time ciphersuites are
9401 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9402 the same ciphersuites as with "HIGH" alone, but in a specific
9403 order where the PSK ciphersuites come first (since they are the
9404 most recently disabled ciphersuites when "HIGH" is parsed).
9406 Also, change ssl_create_cipher_list() (using this new
9407 functionality) such that between otherwise identical
9408 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9413 * Change ssl_create_cipher_list() so that it automatically
9414 arranges the ciphersuites in reasonable order before starting
9415 to process the rule string. Thus, the definition for "DEFAULT"
9416 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9417 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9418 This makes it much easier to arrive at a reasonable default order
9419 in applications for which anonymous ciphers are OK (meaning
9420 that you can't actually use DEFAULT).
9422 *Bodo Moeller; suggested by Victor Duchovni*
9424 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9425 processing) into multiple integers instead of setting
9426 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9427 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9428 (These masks as well as the individual bit definitions are hidden
9429 away into the non-exported interface ssl/ssl_locl.h, so this
9430 change to the definition of the SSL_CIPHER structure shouldn't
9431 affect applications.) This give us more bits for each of these
9432 categories, so there is no longer a need to coagulate AES128 and
9433 AES256 into a single algorithm bit, and to coagulate Camellia128
9434 and Camellia256 into a single algorithm bit, which has led to all
9437 Thus, among other things, the kludge introduced in 0.9.7m and
9438 0.9.8e for masking out AES256 independently of AES128 or masking
9439 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9441 With the change, we also introduce new ciphersuite aliases that
9442 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9447 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9448 Use the leftmost N bytes of the signature input if the input is
9449 larger than the prime q (with N being the size in bytes of q).
9453 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9454 it yet and it is largely untested.
9458 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9462 * Initial incomplete changes to avoid need for function casts in OpenSSL
9463 some compilers (gcc 4.2 and later) reject their use. Safestack is
9464 reimplemented. Update ASN1 to avoid use of legacy functions.
9468 * Win32/64 targets are linked with Winsock2.
9472 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9473 to external functions. This can be used to increase CRL handling
9474 efficiency especially when CRLs are very large by (for example) storing
9475 the CRL revoked certificates in a database.
9479 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9480 new CRLs added to a directory can be used. New command line option
9481 -verify_return_error to s_client and s_server. This causes real errors
9482 to be returned by the verify callback instead of carrying on no matter
9483 what. This reflects the way a "real world" verify callback would behave.
9487 * GOST engine, supporting several GOST algorithms and public key formats.
9488 Kindly donated by Cryptocom.
9492 * Partial support for Issuing Distribution Point CRL extension. CRLs
9493 partitioned by DP are handled but no indirect CRL or reason partitioning
9494 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9495 selected via a scoring technique which handles IDP and AKID in CRLs.
9499 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9500 will ultimately be used for all verify operations: this will remove the
9501 X509_STORE dependency on certificate verification and allow alternative
9502 lookup methods. X509_STORE based implementations of these two callbacks.
9506 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9507 Modify get_crl() to find a valid (unexpired) CRL if possible.
9511 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9512 this would be called X509_CRL_cmp() but that name is already used by
9513 a function that just compares CRL issuer names. Cache several CRL
9514 extensions in X509_CRL structure and cache CRLDP in X509.
9518 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9519 this maps equivalent X509_NAME structures into a consistent structure.
9520 Name comparison can then be performed rapidly using memcmp().
9524 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9529 * Allow digests to supply their own micalg string for S/MIME type using
9530 the ctrl EVP_MD_CTRL_MICALG.
9534 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9535 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9536 ctrl. It can then customise the structure before and/or after signing
9541 * New function OBJ_add_sigid() to allow application defined signature OIDs
9542 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9543 to free up any added signature OIDs.
9547 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9548 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9549 digest and cipher tables. New options added to openssl utility:
9550 list-message-digest-algorithms and list-cipher-algorithms.
9554 * Change the array representation of binary polynomials: the list
9555 of degrees of non-zero coefficients is now terminated with -1.
9556 Previously it was terminated with 0, which was also part of the
9557 value; thus, the array representation was not applicable to
9558 polynomials where t^0 has coefficient zero. This change makes
9559 the array representation useful in a more general context.
9563 * Various modifications and fixes to SSL/TLS cipher string
9564 handling. For ECC, the code now distinguishes between fixed ECDH
9565 with RSA certificates on the one hand and with ECDSA certificates
9566 on the other hand, since these are separate ciphersuites. The
9567 unused code for Fortezza ciphersuites has been removed.
9569 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9570 (not "ECDHE"). For consistency with the code for DH
9571 certificates, use of ECDH certificates is now considered ECDH
9572 authentication, not RSA or ECDSA authentication (the latter is
9573 merely the CA's signing algorithm and not actively used in the
9576 The temporary ciphersuite alias "ECCdraft" is no longer
9577 available, and ECC ciphersuites are no longer excluded from "ALL"
9578 and "DEFAULT". The following aliases now exist for RFC 4492
9579 ciphersuites, most of these by analogy with the DH case:
9581 kECDHr - ECDH cert, signed with RSA
9582 kECDHe - ECDH cert, signed with ECDSA
9583 kECDH - ECDH cert (signed with either RSA or ECDSA)
9584 kEECDH - ephemeral ECDH
9585 ECDH - ECDH cert or ephemeral ECDH
9591 AECDH - anonymous ECDH
9592 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9596 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9597 Use correct micalg parameters depending on digest(s) in signed message.
9601 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9602 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9606 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9607 an engine to register a method. Add ENGINE lookups for methods and
9608 functional reference processing.
9612 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9613 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9618 * New -resign option to smime utility. This adds one or more signers
9619 to an existing PKCS#7 signedData structure. Also -md option to use an
9620 alternative message digest algorithm for signing.
9624 * Tidy up PKCS#7 routines and add new functions to make it easier to
9625 create PKCS7 structures containing multiple signers. Update smime
9626 application to support multiple signers.
9630 * New -macalg option to pkcs12 utility to allow setting of an alternative
9635 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9636 Reorganize PBE internals to lookup from a static table using NIDs,
9637 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9638 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9639 PRF which will be automatically used with PBES2.
9643 * Replace the algorithm specific calls to generate keys in "req" with the
9648 * Update PKCS#7 enveloped data routines to use new API. This is now
9649 supported by any public key method supporting the encrypt operation. A
9650 ctrl is added to allow the public key algorithm to examine or modify
9651 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9656 * Add a ctrl to asn1 method to allow a public key algorithm to express
9657 a default digest type to use. In most cases this will be SHA1 but some
9658 algorithms (such as GOST) need to specify an alternative digest. The
9659 return value indicates how strong the preference is 1 means optional and
9660 2 is mandatory (that is it is the only supported type). Modify
9661 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9662 use the default md. Update openssl utilities to use the default digest
9663 type for signing if it is not explicitly indicated.
9667 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9668 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9669 signing method from the key type. This effectively removes the link
9670 between digests and public key types.
9674 * Add an OID cross reference table and utility functions. Its purpose is to
9675 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9676 rsaEncryption. This will allow some of the algorithm specific hackery
9677 needed to use the correct OID to be removed.
9681 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9682 structures for PKCS7_sign(). They are now set up by the relevant public
9687 * Add provisional EC pkey method with support for ECDSA and ECDH.
9691 * Add support for key derivation (agreement) in the API, DH method and
9696 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9697 public and private key formats. As a side effect these add additional
9698 command line functionality not previously available: DSA signatures can be
9699 generated and verified using pkeyutl and DH key support and generation in
9706 *Oliver Tappe <zooey@hirschkaefer.de>*
9708 * New make target "install_html_docs" installs HTML renditions of the
9711 *Oliver Tappe <zooey@hirschkaefer.de>*
9713 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9714 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9715 support key and parameter generation and add initial key generation
9716 functionality for RSA.
9720 * Add functions for main EVP_PKEY_method operations. The undocumented
9721 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9722 `EVP_PKEY_{encrypt,decrypt}_old`.
9726 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9727 key API, doesn't do much yet.
9731 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9732 public key algorithms. New option to openssl utility:
9733 "list-public-key-algorithms" to print out info.
9737 * Implement the Supported Elliptic Curves Extension for
9738 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9742 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9743 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9747 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9748 utilities such as rsa, dsa, dsaparam etc except they process any key
9753 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9754 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9755 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9760 * Initial support for pluggable public key ASN1.
9761 De-spaghettify the public key ASN1 handling. Move public and private
9762 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9763 algorithm specific handling to a single module within the relevant
9764 algorithm directory. Add functions to allow (near) opaque processing
9765 of public and private key structures.
9769 * Implement the Supported Point Formats Extension for
9770 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9774 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9775 for the psk identity [hint] and the psk callback functions to the
9776 SSL_SESSION, SSL and SSL_CTX structure.
9779 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9783 SSL_CTX_use_psk_identity_hint
9784 SSL_get_psk_identity_hint
9785 SSL_get_psk_identity
9786 SSL_use_psk_identity_hint
9788 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9790 * Add RFC 3161 compliant time stamp request creation, response generation
9791 and response verification functionality.
9793 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9795 * Add initial support for TLS extensions, specifically for the server_name
9796 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9797 have new members for a hostname. The SSL data structure has an
9798 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9799 stored in that context to allow for session resumption, even after the
9800 SSL has been switched to a new SSL_CTX in reaction to a client's
9801 server_name extension.
9803 New functions (subject to change):
9805 SSL_get_servername()
9806 SSL_get_servername_type()
9809 New CTRL codes and macros (subject to change):
9811 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9812 - SSL_CTX_set_tlsext_servername_callback()
9813 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9814 - SSL_CTX_set_tlsext_servername_arg()
9815 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9817 openssl s_client has a new '-servername ...' option.
9819 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9820 '-key2 ...', '-servername_fatal' (subject to change). This allows
9821 testing the HostName extension for a specific single hostname ('-cert'
9822 and '-key' remain fallbacks for handshakes without HostName
9823 negotiation). If the unrecognized_name alert has to be sent, this by
9824 default is a warning; it becomes fatal with the '-servername_fatal'
9827 *Peter Sylvester, Remy Allais, Christophe Renou*
9829 * Whirlpool hash implementation is added.
9833 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9834 bn(64,32). Because of instruction set limitations it doesn't have
9835 any negative impact on performance. This was done mostly in order
9836 to make it possible to share assembler modules, such as bn_mul_mont
9837 implementations, between 32- and 64-bit builds without hassle.
9841 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9842 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9847 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9848 dedicated Montgomery multiplication procedure, is introduced.
9849 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9850 "64-bit" performance on certain 32-bit targets.
9854 * New option SSL_OP_NO_COMP to disable use of compression selectively
9855 in SSL structures. New SSL ctrl to set maximum send fragment size.
9856 Save memory by setting the I/O buffer sizes dynamically instead of
9857 using the maximum available value.
9861 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9862 in addition to the text details.
9866 * Very, very preliminary EXPERIMENTAL support for printing of general
9867 ASN1 structures. This currently produces rather ugly output and doesn't
9868 handle several customised structures at all.
9872 * Integrated support for PVK file format and some related formats such
9873 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9874 these in the 'rsa' and 'dsa' utilities.
9878 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9882 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9883 place for the (very old) "NETSCAPE" format certificates which are now
9884 handled using new ASN1 code equivalents.
9888 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9889 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9890 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9894 * Modify CRL distribution points extension code to print out previously
9895 unsupported fields. Enhance extension setting code to allow setting of
9900 * Add print and set support for Issuing Distribution Point CRL extension.
9904 * Change 'Configure' script to enable Camellia by default.
9911 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9913 * When rejecting SSL/TLS records due to an incorrect version number, never
9914 update s->server with a new major version number. As of
9915 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9916 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9917 the previous behavior could result in a read attempt at NULL when
9918 receiving specific incorrect SSL/TLS records once record payload
9919 protection is active. ([CVE-2010-0740])
9921 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9923 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9924 could be crashed if the relevant tables were not present (e.g. chrooted).
9926 *Tomas Hoger <thoger@redhat.com>*
9928 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9930 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9932 *Martin Olsson, Neel Mehta*
9934 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9935 accommodate for stack sorting, always a write lock!).
9939 * On some versions of WIN32 Heap32Next is very slow. This can cause
9940 excessive delays in the RAND_poll(): over a minute. As a workaround
9941 include a time check in the inner Heap32Next loop too.
9945 * The code that handled flushing of data in SSL/TLS originally used the
9946 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9947 the problem outlined in PR#1949. The fix suggested there however can
9948 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9949 of Apache). So instead simplify the code to flush unconditionally.
9950 This should be fine since flushing with no data to flush is a no op.
9954 * Handle TLS versions 2.0 and later properly and correctly use the
9955 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9956 off ancient servers have a habit of sticking around for a while...
9960 * Modify compression code so it frees up structures without using the
9961 ex_data callbacks. This works around a problem where some applications
9962 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9963 restarting) then use compression (e.g. SSL with compression) later.
9964 This results in significant per-connection memory leaks and
9965 has caused some security issues including CVE-2008-1678 and
9970 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9971 change when encrypting or decrypting.
9975 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9976 connect and renegotiate with servers which do not support RI.
9977 Until RI is more widely deployed this option is enabled by default.
9981 * Add "missing" ssl ctrls to clear options and mode.
9985 * If client attempts to renegotiate and doesn't support RI respond with
9986 a no_renegotiation alert as required by RFC5746. Some renegotiating
9987 TLS clients will continue a connection gracefully when they receive
9988 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9989 waiting for a server hello which it will never receive. Now we treat a
9990 received no_renegotiation alert as a fatal error. This is because
9991 applications requesting a renegotiation might well expect it to succeed
9992 and would have no code in place to handle the server denying it so the
9993 only safe thing to do is to terminate the connection.
9997 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9998 peer supports secure renegotiation and 0 otherwise. Print out peer
9999 renegotiation support in s_client/s_server.
10003 * Replace the highly broken and deprecated SPKAC certification method with
10004 the updated NID creation version. This should correctly handle UTF8.
10008 * Implement RFC5746. Re-enable renegotiation but require the extension
10009 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10010 turns out to be a bad idea. It has been replaced by
10011 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10012 SSL_CTX_set_options(). This is really not recommended unless you
10013 know what you are doing.
10015 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10017 * Fixes to stateless session resumption handling. Use initial_ctx when
10018 issuing and attempting to decrypt tickets in case it has changed during
10019 servername handling. Use a non-zero length session ID when attempting
10020 stateless session resumption: this makes it possible to determine if
10021 a resumption has occurred immediately after receiving server hello
10022 (several places in OpenSSL subtly assume this) instead of later in
10027 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10028 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10029 fixes for a few places where the return code is not checked
10032 *Julia Lawall <julia@diku.dk>*
10034 * Add --strict-warnings option to Configure script to include devteam
10035 warnings in other configurations.
10039 * Add support for --libdir option and LIBDIR variable in makefiles. This
10040 makes it possible to install openssl libraries in locations which
10041 have names other than "lib", for example "/usr/lib64" which some
10044 *Steve Henson, based on patch from Jeremy Utley*
10046 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10047 X690 8.9.12 and can produce some misleading textual output of OIDs.
10049 *Steve Henson, reported by Dan Kaminsky*
10051 * Delete MD2 from algorithm tables. This follows the recommendation in
10052 several standards that it is not used in new applications due to
10053 several cryptographic weaknesses. For binary compatibility reasons
10054 the MD2 API is still compiled in by default.
10058 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10063 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10064 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10067 *Guenter <lists@gknw.net>*
10069 * Fix the server certificate chain building code to use X509_verify_cert(),
10070 it used to have an ad-hoc builder which was unable to cope with anything
10071 other than a simple chain.
10073 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10075 * Don't check self signed certificate signatures in X509_verify_cert()
10076 by default (a flag can override this): it just wastes time without
10077 adding any security. As a useful side effect self signed root CAs
10078 with non-FIPS digests are now usable in FIPS mode.
10082 * In dtls1_process_out_of_seq_message() the check if the current message
10083 is already buffered was missing. For every new message was memory
10084 allocated, allowing an attacker to perform an denial of service attack
10085 with sending out of seq handshake messages until there is no memory
10086 left. Additionally every future message was buffered, even if the
10087 sequence number made no sense and would be part of another handshake.
10088 So only messages with sequence numbers less than 10 in advance will be
10089 buffered. ([CVE-2009-1378])
10091 *Robin Seggelmann, discovered by Daniel Mentz*
10093 * Records are buffered if they arrive with a future epoch to be
10094 processed after finishing the corresponding handshake. There is
10095 currently no limitation to this buffer allowing an attacker to perform
10096 a DOS attack with sending records with future epochs until there is no
10097 memory left. This patch adds the pqueue_size() function to determine
10098 the size of a buffer and limits the record buffer to 100 entries.
10101 *Robin Seggelmann, discovered by Daniel Mentz*
10103 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10104 parent structure is freed. ([CVE-2009-1379])
10108 * Handle non-blocking I/O properly in SSL_shutdown() call.
10110 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10112 * Add `2.5.4.*` OIDs
10114 *Ilya O. <vrghost@gmail.com>*
10116 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10118 * Disable renegotiation completely - this fixes a severe security
10119 problem ([CVE-2009-3555]) at the cost of breaking all
10120 renegotiation. Renegotiation can be re-enabled by setting
10121 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10122 run-time. This is really not recommended unless you know what
10127 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10129 * Don't set val to NULL when freeing up structures, it is freed up by
10130 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10131 zeroing past the valid field. ([CVE-2009-0789])
10133 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10135 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10136 checked correctly. This would allow some invalid signed attributes to
10137 appear to verify correctly. ([CVE-2009-0591])
10139 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10141 * Reject UniversalString and BMPString types with invalid lengths. This
10142 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10143 a legal length. ([CVE-2009-0590])
10147 * Set S/MIME signing as the default purpose rather than setting it
10148 unconditionally. This allows applications to override it at the store
10153 * Permit restricted recursion of ASN1 strings. This is needed in practice
10154 to handle some structures.
10158 * Improve efficiency of mem_gets: don't search whole buffer each time
10161 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10163 * New -hex option for openssl rand.
10167 * Print out UTF8String and NumericString when parsing ASN1.
10171 * Support NumericString type for name components.
10175 * Allow CC in the environment to override the automatically chosen
10176 compiler. Note that nothing is done to ensure flags work with the
10181 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10183 * Properly check EVP_VerifyFinal() and similar return values
10186 *Ben Laurie, Bodo Moeller, Google Security Team*
10188 * Enable TLS extensions by default.
10192 * Allow the CHIL engine to be loaded, whether the application is
10193 multithreaded or not. (This does not release the developer from the
10194 obligation to set up the dynamic locking callbacks.)
10196 *Sander Temme <sander@temme.net>*
10198 * Use correct exit code if there is an error in dgst command.
10200 *Steve Henson; problem pointed out by Roland Dirlewanger*
10202 * Tweak Configure so that you need to say "experimental-jpake" to enable
10203 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10207 * Add experimental JPAKE support, including demo authentication in
10208 s_client and s_server.
10212 * Set the comparison function in v3_addr_canonize().
10214 *Rob Austein <sra@hactrn.net>*
10216 * Add support for XMPP STARTTLS in s_client.
10218 *Philip Paeps <philip@freebsd.org>*
10220 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10221 to ensure that even with this option, only ciphersuites in the
10222 server's preference list will be accepted. (Note that the option
10223 applies only when resuming a session, so the earlier behavior was
10224 just about the algorithm choice for symmetric cryptography.)
10228 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10230 * Fix NULL pointer dereference if a DTLS server received
10231 ChangeCipherSpec as first record ([CVE-2009-1386]).
10235 * Fix a state transition in s3_srvr.c and d1_srvr.c
10236 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10238 *Nagendra Modadugu*
10240 * The fix in 0.9.8c that supposedly got rid of unsafe
10241 double-checked locking was incomplete for RSA blinding,
10242 addressing just one layer of what turns out to have been
10243 doubly unsafe triple-checked locking.
10245 So now fix this for real by retiring the MONT_HELPER macro
10246 in crypto/rsa/rsa_eay.c.
10248 *Bodo Moeller; problem pointed out by Marius Schilder*
10250 * Various precautionary measures:
10252 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10254 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10255 (NB: This would require knowledge of the secret session ticket key
10256 to exploit, in which case you'd be SOL either way.)
10258 - Change bn_nist.c so that it will properly handle input BIGNUMs
10259 outside the expected range.
10261 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10264 *Neel Mehta, Bodo Moeller*
10266 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10267 the load fails. Useful for distros.
10269 *Ben Laurie and the FreeBSD team*
10271 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10275 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10279 * Expand ENGINE to support engine supplied SSL client certificate functions.
10281 This work was sponsored by Logica.
10285 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10286 keystores. Support for SSL/TLS client authentication too.
10287 Not compiled unless enable-capieng specified to Configure.
10289 This work was sponsored by Logica.
10293 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10294 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10295 attribute creation routines such as certificate requests and PKCS#12
10300 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10302 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10303 handshake which could lead to a client crash as found using the
10304 Codenomicon TLS test suite ([CVE-2008-1672])
10306 *Steve Henson, Mark Cox*
10308 * Fix double free in TLS server name extensions which could lead to
10309 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10313 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10315 Clear the error queue to ensure that error entries left from
10316 older function calls do not interfere with the correct operation.
10318 *Lutz Jaenicke, Erik de Castro Lopo*
10320 * Remove root CA certificates of commercial CAs:
10322 The OpenSSL project does not recommend any specific CA and does not
10323 have any policy with respect to including or excluding any CA.
10324 Therefore, it does not make any sense to ship an arbitrary selection
10325 of root CA certificates with the OpenSSL software.
10329 * RSA OAEP patches to fix two separate invalid memory reads.
10330 The first one involves inputs when 'lzero' is greater than
10331 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10332 before the beginning of from). The second one involves inputs where
10333 the 'db' section contains nothing but zeroes (there is a one-byte
10334 invalid read after the end of 'db').
10336 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10338 * Partial backport from 0.9.9-dev:
10340 Introduce bn_mul_mont (dedicated Montgomery multiplication
10341 procedure) as a candidate for BIGNUM assembler implementation.
10342 While 0.9.9-dev uses assembler for various architectures, only
10343 x86_64 is available by default here in the 0.9.8 branch, and
10344 32-bit x86 is available through a compile-time setting.
10346 To try the 32-bit x86 assembler implementation, use Configure
10347 option "enable-montasm" (which exists only for this backport).
10349 As "enable-montasm" for 32-bit x86 disclaims code stability
10350 anyway, in this constellation we activate additional code
10351 backported from 0.9.9-dev for further performance improvements,
10352 namely BN_from_montgomery_word. (To enable this otherwise,
10353 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10355 *Andy Polyakov (backport partially by Bodo Moeller)*
10357 * Add TLS session ticket callback. This allows an application to set
10358 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10359 values. This is useful for key rollover for example where several key
10360 sets may exist with different names.
10364 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10365 This was broken until now in 0.9.8 releases, such that the only way
10366 a registered ENGINE could be used (assuming it initialises
10367 successfully on the host) was to explicitly set it as the default
10368 for the relevant algorithms. This is in contradiction with 0.9.7
10369 behaviour and the documentation. With this fix, when an ENGINE is
10370 registered into a given algorithm's table of implementations, the
10371 'uptodate' flag is reset so that auto-discovery will be used next
10372 time a new context for that algorithm attempts to select an
10375 *Ian Lister (tweaked by Geoff Thorpe)*
10377 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10378 implementation in the following ways:
10380 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10383 Lack of BER streaming support means one pass streaming processing is
10384 only supported if data is detached: setting the streaming flag is
10385 ignored for embedded content.
10387 CMS support is disabled by default and must be explicitly enabled
10388 with the enable-cms configuration option.
10392 * Update the GMP engine glue to do direct copies between BIGNUM and
10393 mpz_t when openssl and GMP use the same limb size. Otherwise the
10394 existing "conversion via a text string export" trick is still used.
10396 *Paul Sheer <paulsheer@gmail.com>*
10398 * Zlib compression BIO. This is a filter BIO which compressed and
10399 uncompresses any data passed through it.
10403 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10404 RFC3394 compatible AES key wrapping.
10408 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10409 sets string data without copying. X509_ALGOR_set0() and
10410 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10411 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10412 from an X509_ATTRIBUTE structure optionally checking it occurs only
10413 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10418 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10419 to get the expected BN_FLG_CONSTTIME behavior.
10421 *Bodo Moeller (Google)*
10425 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10426 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10427 - added some more tests to do_tests.pl
10428 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10429 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10430 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10431 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10432 - various changes to netware.pl to enable gcc-cross builds on Win32
10434 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10435 - various changes to fix missing prototype warnings
10436 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10437 - added AES, WHIRLPOOL and CPUID assembler code to build files
10438 - added missing AES assembler make rules to mk1mf.pl
10439 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10441 *Guenter Knauf <eflash@gmx.net>*
10443 * Implement certificate status request TLS extension defined in RFC3546.
10444 A client can set the appropriate parameters and receive the encoded
10445 OCSP response via a callback. A server can query the supplied parameters
10446 and set the encoded OCSP response in the callback. Add simplified examples
10447 to s_client and s_server.
10451 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10453 * Fix various bugs:
10454 + Binary incompatibility of ssl_ctx_st structure
10455 + DTLS interoperation with non-compliant servers
10456 + Don't call get_session_cb() without proposed session
10457 + Fix ia64 assembler code
10459 *Andy Polyakov, Steve Henson*
10461 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10463 * DTLS Handshake overhaul. There were longstanding issues with
10464 OpenSSL DTLS implementation, which were making it impossible for
10465 RFC 4347 compliant client to communicate with OpenSSL server.
10466 Unfortunately just fixing these incompatibilities would "cut off"
10467 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10468 server keeps tolerating non RFC compliant syntax. The opposite is
10469 not true, 0.9.8f client can not communicate with earlier server.
10470 This update even addresses CVE-2007-4995.
10474 * Changes to avoid need for function casts in OpenSSL: some compilers
10475 (gcc 4.2 and later) reject their use.
10476 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10479 * Add RFC4507 support to OpenSSL. This includes the corrections in
10480 RFC4507bis. The encrypted ticket format is an encrypted encoded
10481 SSL_SESSION structure, that way new session features are automatically
10484 If a client application caches session in an SSL_SESSION structure
10485 support is transparent because tickets are now stored in the encoded
10488 The SSL_CTX structure automatically generates keys for ticket
10489 protection in servers so again support should be possible
10490 with no application modification.
10492 If a client or server wishes to disable RFC4507 support then the option
10493 SSL_OP_NO_TICKET can be set.
10495 Add a TLS extension debugging callback to allow the contents of any client
10496 or server extensions to be examined.
10498 This work was sponsored by Google.
10502 * Add initial support for TLS extensions, specifically for the server_name
10503 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10504 have new members for a hostname. The SSL data structure has an
10505 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10506 stored in that context to allow for session resumption, even after the
10507 SSL has been switched to a new SSL_CTX in reaction to a client's
10508 server_name extension.
10510 New functions (subject to change):
10512 SSL_get_servername()
10513 SSL_get_servername_type()
10516 New CTRL codes and macros (subject to change):
10518 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10519 - SSL_CTX_set_tlsext_servername_callback()
10520 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10521 - SSL_CTX_set_tlsext_servername_arg()
10522 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10524 openssl s_client has a new '-servername ...' option.
10526 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10527 '-key2 ...', '-servername_fatal' (subject to change). This allows
10528 testing the HostName extension for a specific single hostname ('-cert'
10529 and '-key' remain fallbacks for handshakes without HostName
10530 negotiation). If the unrecognized_name alert has to be sent, this by
10531 default is a warning; it becomes fatal with the '-servername_fatal'
10534 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10536 * Add AES and SSE2 assembly language support to VC++ build.
10540 * Mitigate attack on final subtraction in Montgomery reduction.
10544 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10545 (which previously caused an internal error).
10549 * Squeeze another 10% out of IGE mode when in != out.
10553 * AES IGE mode speedup.
10555 *Dean Gaudet (Google)*
10557 * Add the Korean symmetric 128-bit cipher SEED (see
10558 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10559 add SEED ciphersuites from RFC 4162:
10561 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10562 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10563 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10564 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10566 To minimize changes between patchlevels in the OpenSSL 0.9.8
10567 series, SEED remains excluded from compilation unless OpenSSL
10568 is configured with 'enable-seed'.
10570 *KISA, Bodo Moeller*
10572 * Mitigate branch prediction attacks, which can be practical if a
10573 single processor is shared, allowing a spy process to extract
10574 information. For detailed background information, see
10575 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10576 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10577 and Necessary Software Countermeasures"). The core of the change
10578 are new versions BN_div_no_branch() and
10579 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10580 respectively, which are slower, but avoid the security-relevant
10581 conditional branches. These are automatically called by BN_div()
10582 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10583 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10584 remove a conditional branch.
10586 BN_FLG_CONSTTIME is the new name for the previous
10587 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10588 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10589 in the exponent causes BN_mod_exp_mont() to use the alternative
10590 implementation in BN_mod_exp_mont_consttime().) The old name
10591 remains as a deprecated alias.
10593 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10594 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10595 constant-time implementations for more than just exponentiation.
10596 Here too the old name is kept as a deprecated alias.
10598 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10599 the BN_BLINDING structure gets an independent copy of the
10600 modulus. This means that the previous `BIGNUM *m` argument to
10601 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10602 essentially becomes `const BIGNUM *m`, although we can't actually
10603 change this in the header file before 0.9.9. It allows
10604 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10605 enable BN_FLG_CONSTTIME.
10607 *Matthew D Wood (Intel Corp)*
10609 * In the SSL/TLS server implementation, be strict about session ID
10610 context matching (which matters if an application uses a single
10611 external cache for different purposes). Previously,
10612 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10613 set. This did ensure strict client verification, but meant that,
10614 with applications using a single external cache for quite
10615 different requirements, clients could circumvent ciphersuite
10616 restrictions for a given session ID context by starting a session
10617 in a different context.
10621 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10622 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10623 authentication-only ciphersuites.
10627 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10628 not complete and could lead to a possible single byte overflow
10629 ([CVE-2007-5135]) [Ben Laurie]
10631 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10633 * Since AES128 and AES256 (and similarly Camellia128 and
10634 Camellia256) share a single mask bit in the logic of
10635 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10636 kludge to work properly if AES128 is available and AES256 isn't
10637 (or if Camellia128 is available and Camellia256 isn't).
10641 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10642 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10643 When a point or a seed is encoded in a BIT STRING, we need to
10644 prevent the removal of trailing zero bits to get the proper DER
10645 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10646 of a NamedBitList, for which trailing 0 bits need to be removed.)
10650 * Have SSL/TLS server implementation tolerate "mismatched" record
10651 protocol version while receiving ClientHello even if the
10652 ClientHello is fragmented. (The server can't insist on the
10653 particular protocol version it has chosen before the ServerHello
10654 message has informed the client about his choice.)
10658 * Add RFC 3779 support.
10660 *Rob Austein for ARIN, Ben Laurie*
10662 * Load error codes if they are not already present instead of using a
10663 static variable. This allows them to be cleanly unloaded and reloaded.
10664 Improve header file function name parsing.
10668 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10669 or CAPABILITY handshake as required by RFCs.
10671 *Goetz Babin-Ebell*
10673 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10675 * Introduce limits to prevent malicious keys being able to
10676 cause a denial of service. ([CVE-2006-2940])
10678 *Steve Henson, Bodo Moeller*
10680 * Fix ASN.1 parsing of certain invalid structures that can result
10681 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10683 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10684 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10686 * Fix SSL client code which could crash if connecting to a
10687 malicious SSLv2 server. ([CVE-2006-4343])
10689 *Tavis Ormandy and Will Drewry, Google Security Team*
10691 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10692 match only those. Before that, "AES256-SHA" would be interpreted
10693 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10694 the same strength classification in 0.9.7h) as we currently only
10695 have a single AES bit in the ciphersuite description bitmap.
10696 That change, however, also applied to ciphersuite strings such as
10697 "RC4-MD5" that intentionally matched multiple ciphersuites --
10698 namely, SSL 2.0 ciphersuites in addition to the more common ones
10699 from SSL 3.0/TLS 1.0.
10701 So we change the selection algorithm again: Naming an explicit
10702 ciphersuite selects this one ciphersuite, and any other similar
10703 ciphersuite (same bitmap) from *other* protocol versions.
10704 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10705 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10707 Since SSL 2.0 does not have any ciphersuites for which the
10708 128/256 bit distinction would be relevant, this works for now.
10709 The proper fix will be to use different bits for AES128 and
10710 AES256, which would have avoided the problems from the beginning;
10711 however, bits are scarce, so we can only do this in a new release
10712 (not just a patchlevel) when we can change the SSL_CIPHER
10713 definition to split the single 'unsigned long mask' bitmap into
10714 multiple values to extend the available space.
10718 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10720 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10721 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10723 * Add AES IGE and biIGE modes.
10727 * Change the Unix randomness entropy gathering to use poll() when
10728 possible instead of select(), since the latter has some
10729 undesirable limitations.
10731 *Darryl Miles via Richard Levitte and Bodo Moeller*
10733 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10734 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10735 cannot be implicitly activated as part of, e.g., the "AES" alias.
10736 However, please upgrade to OpenSSL 0.9.9[-dev] for
10737 non-experimental use of the ECC ciphersuites to get TLS extension
10738 support, which is required for curve and point format negotiation
10739 to avoid potential handshake problems.
10743 * Disable rogue ciphersuites:
10745 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10746 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10747 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10749 The latter two were purportedly from
10750 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10753 Also deactivate the remaining ciphersuites from
10754 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10755 unofficial, and the ID has long expired.
10759 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10760 dual-core machines) and other potential thread-safety issues.
10764 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10765 versions), which is now available for royalty-free use
10766 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10767 Also, add Camellia TLS ciphersuites from RFC 4132.
10769 To minimize changes between patchlevels in the OpenSSL 0.9.8
10770 series, Camellia remains excluded from compilation unless OpenSSL
10771 is configured with 'enable-camellia'.
10775 * Disable the padding bug check when compression is in use. The padding
10776 bug check assumes the first packet is of even length, this is not
10777 necessarily true if compression is enabled and can result in false
10778 positives causing handshake failure. The actual bug test is ancient
10779 code so it is hoped that implementations will either have fixed it by
10780 now or any which still have the bug do not support compression.
10784 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10786 * When applying a cipher rule check to see if string match is an explicit
10787 cipher suite and only match that one cipher suite if it is.
10791 * Link in manifests for VC++ if needed.
10793 *Austin Ziegler <halostatue@gmail.com>*
10795 * Update support for ECC-based TLS ciphersuites according to
10796 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10797 TLS extensions, which are supported starting with the 0.9.9
10798 branch, not in the OpenSSL 0.9.8 branch).
10802 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10803 opaque EVP_CIPHER_CTX handling.
10807 * Fixes and enhancements to zlib compression code. We now only use
10808 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10809 to conform with the standards mentioned here:
10810 <http://www.zlib.net/DLL_FAQ.txt>
10811 Static zlib linking now works on Windows and the new --with-zlib-include
10812 --with-zlib-lib options to Configure can be used to supply the location
10813 of the headers and library. Gracefully handle case where zlib library
10818 * Several fixes and enhancements to the OID generation code. The old code
10819 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10820 handle numbers larger than ULONG_MAX, truncated printing and had a
10821 non standard OBJ_obj2txt() behaviour.
10825 * Add support for building of engines under engine/ as shared libraries
10826 under VC++ build system.
10830 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10831 Hopefully, we will not see any false combination of paths any more.
10835 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10837 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10838 (part of SSL_OP_ALL). This option used to disable the
10839 countermeasure against man-in-the-middle protocol-version
10840 rollback in the SSL 2.0 server implementation, which is a bad
10841 idea. ([CVE-2005-2969])
10843 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10844 for Information Security, National Institute of Advanced Industrial
10845 Science and Technology [AIST], Japan)*
10847 * Add two function to clear and return the verify parameter flags.
10851 * Keep cipherlists sorted in the source instead of sorting them at
10852 runtime, thus removing the need for a lock.
10856 * Avoid some small subgroup attacks in Diffie-Hellman.
10858 *Nick Mathewson and Ben Laurie*
10860 * Add functions for well-known primes.
10864 * Extended Windows CE support.
10866 *Satoshi Nakamura and Andy Polyakov*
10868 * Initialize SSL_METHOD structures at compile time instead of during
10869 runtime, thus removing the need for a lock.
10873 * Make PKCS7_decrypt() work even if no certificate is supplied by
10874 attempting to decrypt each encrypted key in turn. Add support to
10879 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10881 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10884 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10888 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10889 key into the same file any more.
10893 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10897 * Add -utf8 command line and config file option to 'ca'.
10899 *Stefan <stf@udoma.org*
10901 * Removed the macro des_crypt(), as it seems to conflict with some
10902 libraries. Use DES_crypt().
10906 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10907 involves renaming the source and generated shared-libs for
10908 both. The engines will accept the corrected or legacy ids
10909 ('ncipher' and '4758_cca' respectively) when binding. NB,
10910 this only applies when building 'shared'.
10912 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10914 * Add attribute functions to EVP_PKEY structure. Modify
10915 PKCS12_create() to recognize a CSP name attribute and
10916 use it. Make -CSP option work again in pkcs12 utility.
10920 * Add new functionality to the bn blinding code:
10921 - automatic re-creation of the BN_BLINDING parameters after
10922 a fixed number of uses (currently 32)
10923 - add new function for parameter creation
10924 - introduce flags to control the update behaviour of the
10925 BN_BLINDING parameters
10926 - hide BN_BLINDING structure
10927 Add a second BN_BLINDING slot to the RSA structure to improve
10928 performance when a single RSA object is shared among several
10933 * Add support for DTLS.
10935 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10937 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10938 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10942 * Remove buggy and incomplete DH cert support from
10943 ssl/ssl_rsa.c and ssl/s3_both.c
10947 * Use SHA-1 instead of MD5 as the default digest algorithm for
10948 the `apps/openssl` commands.
10952 * Compile clean with "-Wall -Wmissing-prototypes
10953 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10954 DEBUG_SAFESTACK must also be set.
10958 * Change ./Configure so that certain algorithms can be disabled by default.
10959 The new counterpiece to "no-xxx" is "enable-xxx".
10961 The patented RC5 and MDC2 algorithms will now be disabled unless
10962 "enable-rc5" and "enable-mdc2", respectively, are specified.
10964 (IDEA remains enabled despite being patented. This is because IDEA
10965 is frequently required for interoperability, and there is no license
10966 fee for non-commercial use. As before, "no-idea" can be used to
10967 avoid this algorithm.)
10971 * Add processing of proxy certificates (see RFC 3820). This work was
10972 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10973 EGEE (Enabling Grids for E-science in Europe).
10977 * RC4 performance overhaul on modern architectures/implementations, such
10978 as Intel P4, IA-64 and AMD64.
10982 * New utility extract-section.pl. This can be used specify an alternative
10983 section number in a pod file instead of having to treat each file as
10984 a separate case in Makefile. This can be done by adding two lines to the
10987 =for comment openssl_section:XXX
10989 The blank line is mandatory.
10993 * New arguments -certform, -keyform and -pass for s_client and s_server
10994 to allow alternative format key and certificate files and passphrase
10999 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11000 update associated structures and add various utility functions.
11002 Add new policy related verify parameters, include policy checking in
11003 standard verify code. Enhance 'smime' application with extra parameters
11004 to support policy checking and print out.
11008 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11009 Nehemiah processors. These extensions support AES encryption in hardware
11010 as well as RNG (though RNG support is currently disabled).
11012 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11014 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11018 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11020 *Andy Polyakov and a number of other people*
11022 * Improved PowerPC platform support. Most notably BIGNUM assembler
11023 implementation contributed by IBM.
11025 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11027 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11028 exponent rather than 'unsigned long'. There is a corresponding change to
11029 the new 'rsa_keygen' element of the RSA_METHOD structure.
11031 *Jelte Jansen, Geoff Thorpe*
11033 * Functionality for creating the initial serial number file is now
11034 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11036 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11037 number file to 1, which is bound to cause problems. To avoid
11038 the problems while respecting compatibility between different 0.9.7
11039 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11040 CA.pl for serial number initialization. With the new release 0.9.8,
11041 we can fix the problem directly in the 'ca' utility.)
11045 * Reduced header interdependencies by declaring more opaque objects in
11046 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11047 give fewer recursive includes, which could break lazy source code - so
11048 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11049 developers should define this symbol when building and using openssl to
11050 ensure they track the recommended behaviour, interfaces, [etc], but
11051 backwards-compatible behaviour prevails when this isn't defined.
11055 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11059 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11060 This will generate a random key of the appropriate length based on the
11061 cipher context. The EVP_CIPHER can provide its own random key generation
11062 routine to support keys of a specific form. This is used in the des and
11063 3des routines to generate a key of the correct parity. Update S/MIME
11064 code to use new functions and hence generate correct parity DES keys.
11065 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11066 valid (weak or incorrect parity).
11070 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11071 as looking them up. This is useful when the verified structure may contain
11072 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11073 present unless the new PKCS7_NO_CRL flag is asserted.
11077 * Extend ASN1 oid configuration module. It now additionally accepts the
11080 shortName = some long name, 1.2.3.4
11084 * Reimplemented the BN_CTX implementation. There is now no more static
11085 limitation on the number of variables it can handle nor the depth of the
11086 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11087 information can now expand as required, and rather than having a single
11088 static array of bignums, BN_CTX now uses a linked-list of such arrays
11089 allowing it to expand on demand whilst maintaining the usefulness of
11090 BN_CTX's "bundling".
11094 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11095 to allow all RSA operations to function using a single BN_CTX.
11099 * Preliminary support for certificate policy evaluation and checking. This
11100 is initially intended to pass the tests outlined in "Conformance Testing
11101 of Relying Party Client Certificate Path Processing Logic" v1.07.
11105 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11106 remained unused and not that useful. A variety of other little bignum
11107 tweaks and fixes have also been made continuing on from the audit (see
11112 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11113 associated ASN1, EVP and SSL functions and old ASN1 macros.
11117 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11118 and this should never fail. So the return value from the use of
11119 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11120 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11124 * BN_CTX_get() should return zero-valued bignums, providing the same
11125 initialised value as BN_new().
11127 *Geoff Thorpe, suggested by Ulf Möller*
11129 * Support for inhibitAnyPolicy certificate extension.
11133 * An audit of the BIGNUM code is underway, for which debugging code is
11134 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11135 is considered valid when processing BIGNUMs, and causes execution to
11136 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11137 further steps are taken to deliberately pollute unused data in BIGNUM
11138 structures to try and expose faulty code further on. For now, openssl will
11139 (in its default mode of operation) continue to tolerate the inconsistent
11140 forms that it has tolerated in the past, but authors and packagers should
11141 consider trying openssl and their own applications when compiled with
11142 these debugging symbols defined. It will help highlight potential bugs in
11143 their own code, and will improve the test coverage for OpenSSL itself. At
11144 some point, these tighter rules will become openssl's default to improve
11145 maintainability, though the assert()s and other overheads will remain only
11146 in debugging configurations. See bn.h for more details.
11148 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11150 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11151 that can only be obtained through BN_CTX_new() (which implicitly
11152 initialises it). The presence of this function only made it possible
11153 to overwrite an existing structure (and cause memory leaks).
11157 * Because of the callback-based approach for implementing LHASH as a
11158 template type, lh_insert() adds opaque objects to hash-tables and
11159 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11160 to clean up those corresponding objects before destroying the hash table
11161 (and losing the object pointers). So some over-zealous constifications in
11162 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11163 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11164 prototyped to have "const" restrictions on the object pointers they are
11165 given (and so aren't required to cast them away any more).
11169 * The tmdiff.h API was so ugly and minimal that our own timing utility
11170 (speed) prefers to use its own implementation. The two implementations
11171 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11172 its object type properly exposed (MS_TM) instead of casting to/from
11173 `char *`. This may still change yet if someone realises MS_TM and
11175 aren't necessarily the greatest nomenclatures - but this is what was used
11176 internally to the implementation so I've used that for now.
11180 * Ensure that deprecated functions do not get compiled when
11181 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11182 the self-tests were still using deprecated key-generation functions so
11183 these have been updated also.
11187 * Reorganise PKCS#7 code to separate the digest location functionality
11188 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11189 New function PKCS7_set_digest() to set the digest type for PKCS#7
11190 digestedData type. Add additional code to correctly generate the
11191 digestedData type and add support for this type in PKCS7 initialization
11196 * New function PKCS7_set0_type_other() this initializes a PKCS7
11197 structure of type "other".
11201 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11202 sure the loop does correctly stop and breaking ("division by zero")
11203 modulus operations are not performed. The (pre-generated) prime
11204 table crypto/bn/bn_prime.h was already correct, but it could not be
11205 re-generated on some platforms because of the "division by zero"
11206 situation in the script.
11208 *Ralf S. Engelschall*
11210 * Update support for ECC-based TLS ciphersuites according to
11211 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11212 SHA-1 now is only used for "small" curves (where the
11213 representation of a field element takes up to 24 bytes); for
11214 larger curves, the field element resulting from ECDH is directly
11215 used as premaster secret.
11217 *Douglas Stebila (Sun Microsystems Laboratories)*
11219 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11220 curve secp160r1 to the tests.
11222 *Douglas Stebila (Sun Microsystems Laboratories)*
11224 * Add the possibility to load symbols globally with DSO.
11226 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11228 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11229 control of the error stack.
11233 * Add support for STORE in ENGINE.
11237 * Add the STORE type. The intention is to provide a common interface
11238 to certificate and key stores, be they simple file-based stores, or
11239 HSM-type store, or LDAP stores, or...
11240 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11244 * Add a generic structure called OPENSSL_ITEM. This can be used to
11245 pass a list of arguments to any function as well as provide a way
11246 for a function to pass data back to the caller.
11250 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11251 works like BUF_strdup() but can be used to duplicate a portion of
11252 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11257 * Add the function sk_find_ex() which works like sk_find(), but will
11258 return an index to an element even if an exact match couldn't be
11259 found. The index is guaranteed to point at the element where the
11260 searched-for key would be inserted to preserve sorting order.
11264 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11265 takes an extra flags argument for optional functionality. Currently,
11266 the following flags are defined:
11268 OBJ_BSEARCH_VALUE_ON_NOMATCH
11269 This one gets OBJ_bsearch_ex() to return a pointer to the first
11270 element where the comparing function returns a negative or zero
11273 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11274 This one gets OBJ_bsearch_ex() to return a pointer to the first
11275 element where the comparing function returns zero. This is useful
11276 if there are more than one element where the comparing function
11281 * Make it possible to create self-signed certificates with 'openssl ca'
11282 in such a way that the self-signed certificate becomes part of the
11283 CA database and uses the same mechanisms for serial number generation
11284 as all other certificate signing. The new flag '-selfsign' enables
11285 this functionality. Adapt CA.sh and CA.pl.in.
11289 * Add functionality to check the public key of a certificate request
11290 against a given private. This is useful to check that a certificate
11291 request can be signed by that key (self-signing).
11295 * Make it possible to have multiple active certificates with the same
11296 subject in the CA index file. This is done only if the keyword
11297 'unique_subject' is set to 'no' in the main CA section (default
11298 if 'CA_default') of the configuration file. The value is saved
11299 with the database itself in a separate index attribute file,
11300 named like the index file with '.attr' appended to the name.
11304 * Generate multi-valued AVAs using '+' notation in config files for
11309 * Support for nameConstraints certificate extension.
11313 * Support for policyConstraints certificate extension.
11317 * Support for policyMappings certificate extension.
11321 * Make sure the default DSA_METHOD implementation only uses its
11322 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11323 and change its own handlers to be NULL so as to remove unnecessary
11324 indirection. This lets alternative implementations fallback to the
11325 default implementation more easily.
11329 * Support for directoryName in GeneralName related extensions
11334 * Make it possible to link applications using Makefile.shared.
11335 Make that possible even when linking against static libraries!
11339 * Support for single pass processing for S/MIME signing. This now
11340 means that S/MIME signing can be done from a pipe, in addition
11341 cleartext signing (multipart/signed type) is effectively streaming
11342 and the signed data does not need to be all held in memory.
11344 This is done with a new flag PKCS7_STREAM. When this flag is set
11345 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11346 is done after the data is output (and digests calculated) in
11347 SMIME_write_PKCS7().
11351 * Add full support for -rpath/-R, both in shared libraries and
11352 applications, at least on the platforms where it's known how
11357 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11358 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11359 will now compute a table of multiples of the generator that
11360 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11361 faster (notably in the case of a single point multiplication,
11362 scalar * generator).
11364 *Nils Larsch, Bodo Moeller*
11366 * IPv6 support for certificate extensions. The various extensions
11367 which use the IP:a.b.c.d can now take IPv6 addresses using the
11368 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11373 * Added an ENGINE that implements RSA by performing private key
11374 exponentiations with the GMP library. The conversions to and from
11375 GMP's mpz_t format aren't optimised nor are any montgomery forms
11376 cached, and on x86 it appears OpenSSL's own performance has caught up.
11377 However there are likely to be other architectures where GMP could
11378 provide a boost. This ENGINE is not built in by default, but it can be
11379 specified at Configure time and should be accompanied by the necessary
11380 linker additions, eg;
11381 ./config -DOPENSSL_USE_GMP -lgmp
11385 * "openssl engine" will not display ENGINE/DSO load failure errors when
11386 testing availability of engines with "-t" - the old behaviour is
11387 produced by increasing the feature's verbosity with "-tt".
11391 * ECDSA routines: under certain error conditions uninitialized BN objects
11392 could be freed. Solution: make sure initialization is performed early
11393 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11398 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11399 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11400 software implementations. For DSA and DH, parameter generation can
11401 also be overridden by providing the appropriate method callbacks.
11405 * Change the "progress" mechanism used in key-generation and
11406 primality testing to functions that take a new BN_GENCB pointer in
11407 place of callback/argument pairs. The new API functions have `_ex`
11408 postfixes and the older functions are reimplemented as wrappers for
11409 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11410 declarations of the old functions to help (graceful) attempts to
11411 migrate to the new functions. Also, the new key-generation API
11412 functions operate on a caller-supplied key-structure and return
11413 success/failure rather than returning a key or NULL - this is to
11414 help make "keygen" another member function of RSA_METHOD etc.
11416 Example for using the new callback interface:
11418 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11419 void *my_arg = ...;
11422 BN_GENCB_set(&my_cb, my_callback, my_arg);
11424 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11425 /* For the meaning of a, b in calls to my_callback(), see the
11426 * documentation of the function that calls the callback.
11427 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11428 * my_callback should return 1 if it wants BN_is_prime_ex()
11429 * to continue, or 0 to stop.
11434 * Change the ZLIB compression method to be stateful, and make it
11435 available to TLS with the number defined in
11436 draft-ietf-tls-compression-04.txt.
11440 * Add the ASN.1 structures and functions for CertificatePair, which
11441 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11443 CertificatePair ::= SEQUENCE {
11444 forward [0] Certificate OPTIONAL,
11445 reverse [1] Certificate OPTIONAL,
11446 -- at least one of the pair shall be present -- }
11448 Also implement the PEM functions to read and write certificate
11449 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11451 This needed to be defined, mostly for the sake of the LDAP
11452 attribute crossCertificatePair, but may prove useful elsewhere as
11457 * Make it possible to inhibit symlinking of shared libraries in
11458 Makefile.shared, for Cygwin's sake.
11462 * Extend the BIGNUM API by creating a function
11463 void BN_set_negative(BIGNUM *a, int neg);
11464 and a macro that behave like
11465 int BN_is_negative(const BIGNUM *a);
11467 to avoid the need to access 'a->neg' directly in applications.
11471 * Implement fast modular reduction for pseudo-Mersenne primes
11472 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11473 EC_GROUP_new_curve_GFp() will now automatically use this
11476 *Nils Larsch <nla@trustcenter.de>*
11478 * Add new lock type (CRYPTO_LOCK_BN).
11482 * Change the ENGINE framework to automatically load engines
11483 dynamically from specific directories unless they could be
11484 found to already be built in or loaded. Move all the
11485 current engines except for the cryptodev one to a new
11486 directory engines/.
11487 The engines in engines/ are built as shared libraries if
11488 the "shared" options was given to ./Configure or ./config.
11489 Otherwise, they are inserted in libcrypto.a.
11490 /usr/local/ssl/engines is the default directory for dynamic
11491 engines, but that can be overridden at configure time through
11492 the usual use of --prefix and/or --openssldir, and at run
11493 time with the environment variable OPENSSL_ENGINES.
11495 *Geoff Thorpe and Richard Levitte*
11497 * Add Makefile.shared, a helper makefile to build shared
11498 libraries. Adapt Makefile.org.
11502 * Add version info to Win32 DLLs.
11504 *Peter 'Luna' Runestig" <peter@runestig.com>*
11506 * Add new 'medium level' PKCS#12 API. Certificates and keys
11507 can be added using this API to created arbitrary PKCS#12
11508 files while avoiding the low-level API.
11510 New options to PKCS12_create(), key or cert can be NULL and
11511 will then be omitted from the output file. The encryption
11512 algorithm NIDs can be set to -1 for no encryption, the mac
11513 iteration count can be set to 0 to omit the mac.
11515 Enhance pkcs12 utility by making the -nokeys and -nocerts
11516 options work when creating a PKCS#12 file. New option -nomac
11517 to omit the mac, NONE can be set for an encryption algorithm.
11518 New code is modified to use the enhanced PKCS12_create()
11519 instead of the low-level API.
11523 * Extend ASN1 encoder to support indefinite length constructed
11524 encoding. This can output sequences tags and octet strings in
11525 this form. Modify pk7_asn1.c to support indefinite length
11526 encoding. This is experimental and needs additional code to
11527 be useful, such as an ASN1 bio and some enhanced streaming
11530 Extend template encode functionality so that tagging is passed
11531 down to the template encoder.
11535 * Let 'openssl req' fail if an argument to '-newkey' is not
11536 recognized instead of using RSA as a default.
11540 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11541 As these are not official, they are not included in "ALL";
11542 the "ECCdraft" ciphersuite group alias can be used to select them.
11544 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11546 * Add ECDH engine support.
11548 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11550 * Add ECDH in new directory crypto/ecdh/.
11552 *Douglas Stebila (Sun Microsystems Laboratories)*
11554 * Let BN_rand_range() abort with an error after 100 iterations
11555 without success (which indicates a broken PRNG).
11559 * Change BN_mod_sqrt() so that it verifies that the input value
11560 is really the square of the return value. (Previously,
11561 BN_mod_sqrt would show GIGO behaviour.)
11565 * Add named elliptic curves over binary fields from X9.62, SECG,
11566 and WAP/WTLS; add OIDs that were still missing.
11568 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11570 * Extend the EC library for elliptic curves over binary fields
11571 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11574 EC_GF2m_simple_method
11578 EC_GROUP_new_curve_GF2m
11579 EC_GROUP_set_curve_GF2m
11580 EC_GROUP_get_curve_GF2m
11581 EC_POINT_set_affine_coordinates_GF2m
11582 EC_POINT_get_affine_coordinates_GF2m
11583 EC_POINT_set_compressed_coordinates_GF2m
11585 Point compression for binary fields is disabled by default for
11586 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11589 As binary polynomials are represented as BIGNUMs, various members
11590 of the EC_GROUP and EC_POINT data structures can be shared
11591 between the implementations for prime fields and binary fields;
11592 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11593 are essentially identical to their `..._GFp` counterparts.
11594 (For simplicity, the `..._GFp` prefix has been dropped from
11595 various internal method names.)
11597 An internal 'field_div' method (similar to 'field_mul' and
11598 'field_sqr') has been added; this is used only for binary fields.
11600 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11602 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11603 through methods ('mul', 'precompute_mult').
11605 The generic implementations (now internally called 'ec_wNAF_mul'
11606 and 'ec_wNAF_precomputed_mult') remain the default if these
11607 methods are undefined.
11609 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11611 * New function EC_GROUP_get_degree, which is defined through
11612 EC_METHOD. For curves over prime fields, this returns the bit
11613 length of the modulus.
11615 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11617 * New functions EC_GROUP_dup, EC_POINT_dup.
11618 (These simply call ..._new and ..._copy).
11620 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11622 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11623 Polynomials are represented as BIGNUMs (where the sign bit is not
11624 used) in the following functions [macros]:
11627 BN_GF2m_sub [= BN_GF2m_add]
11628 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11629 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11630 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11632 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11633 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11634 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11635 BN_GF2m_cmp [= BN_ucmp]
11637 (Note that only the 'mod' functions are actually for fields GF(2^m).
11638 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11640 For some functions, an the irreducible polynomial defining a
11641 field can be given as an 'unsigned int[]' with strictly
11642 decreasing elements giving the indices of those bits that are set;
11643 i.e., p[] represents the polynomial
11644 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11646 p[0] > p[1] > ... > p[k] = 0.
11647 This applies to the following functions:
11650 BN_GF2m_mod_mul_arr
11651 BN_GF2m_mod_sqr_arr
11652 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11653 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11654 BN_GF2m_mod_exp_arr
11655 BN_GF2m_mod_sqrt_arr
11656 BN_GF2m_mod_solve_quad_arr
11660 Conversion can be performed by the following functions:
11665 bntest.c has additional tests for binary polynomial arithmetic.
11667 Two implementations for BN_GF2m_mod_div() are available.
11668 The default algorithm simply uses BN_GF2m_mod_inv() and
11669 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11670 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11671 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11673 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11675 * Add new error code 'ERR_R_DISABLED' that can be used when some
11676 functionality is disabled at compile-time.
11678 *Douglas Stebila <douglas.stebila@sun.com>*
11680 * Change default behaviour of 'openssl asn1parse' so that more
11681 information is visible when viewing, e.g., a certificate:
11683 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11684 mode the content of non-printable OCTET STRINGs is output in a
11685 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11686 avoid the appearance of a printable string.
11688 *Nils Larsch <nla@trustcenter.de>*
11690 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11692 EC_GROUP_set_asn1_flag()
11693 EC_GROUP_get_asn1_flag()
11694 EC_GROUP_set_point_conversion_form()
11695 EC_GROUP_get_point_conversion_form()
11696 These control ASN1 encoding details:
11697 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11698 has been set to OPENSSL_EC_NAMED_CURVE.
11699 - Points are encoded in uncompressed form by default; options for
11700 asn1_for are as for point2oct, namely
11701 POINT_CONVERSION_COMPRESSED
11702 POINT_CONVERSION_UNCOMPRESSED
11703 POINT_CONVERSION_HYBRID
11705 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11707 EC_GROUP_set_seed()
11708 EC_GROUP_get0_seed()
11709 EC_GROUP_get_seed_len()
11710 This is used only for ASN1 purposes (so far).
11712 *Nils Larsch <nla@trustcenter.de>*
11714 * Add 'field_type' member to EC_METHOD, which holds the NID
11715 of the appropriate field type OID. The new function
11716 EC_METHOD_get_field_type() returns this value.
11718 *Nils Larsch <nla@trustcenter.de>*
11721 EC_POINT_point2bn()
11722 EC_POINT_bn2point()
11723 EC_POINT_point2hex()
11724 EC_POINT_hex2point()
11725 providing useful interfaces to EC_POINT_point2oct() and
11726 EC_POINT_oct2point().
11728 *Nils Larsch <nla@trustcenter.de>*
11730 * Change internals of the EC library so that the functions
11731 EC_GROUP_set_generator()
11732 EC_GROUP_get_generator()
11733 EC_GROUP_get_order()
11734 EC_GROUP_get_cofactor()
11735 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11736 to methods, which would lead to unnecessary code duplication when
11737 adding different types of curves.
11739 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11741 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11742 arithmetic, and such that modified wNAFs are generated
11743 (which avoid length expansion in many cases).
11747 * Add a function EC_GROUP_check_discriminant() (defined via
11748 EC_METHOD) that verifies that the curve discriminant is non-zero.
11750 Add a function EC_GROUP_check() that makes some sanity tests
11751 on a EC_GROUP, its generator and order. This includes
11752 EC_GROUP_check_discriminant().
11754 *Nils Larsch <nla@trustcenter.de>*
11756 * Add ECDSA in new directory crypto/ecdsa/.
11758 Add applications 'openssl ecparam' and 'openssl ecdsa'
11759 (these are based on 'openssl dsaparam' and 'openssl dsa').
11761 ECDSA support is also included in various other files across the
11762 library. Most notably,
11763 - 'openssl req' now has a '-newkey ecdsa:file' option;
11764 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11765 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11766 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11767 them suitable for ECDSA where domain parameters must be
11768 extracted before the specific public key;
11769 - ECDSA engine support has been added.
11771 *Nils Larsch <nla@trustcenter.de>*
11773 * Include some named elliptic curves, and add OIDs from X9.62,
11774 SECG, and WAP/WTLS. Each curve can be obtained from the new
11776 EC_GROUP_new_by_curve_name(),
11777 and the list of available named curves can be obtained with
11778 EC_get_builtin_curves().
11779 Also add a 'curve_name' member to EC_GROUP objects, which can be
11781 EC_GROUP_set_curve_name()
11782 EC_GROUP_get_curve_name()
11784 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11786 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11787 was actually never needed) and in BN_mul(). The removal in BN_mul()
11788 required a small change in bn_mul_part_recursive() and the addition
11789 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11790 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11791 bn_sub_words() and bn_add_words() except they take arrays with
11796 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11798 * Cleanse PEM buffers before freeing them since they may contain
11801 *Benjamin Bennett <ben@psc.edu>*
11803 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11804 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11805 authentication-only ciphersuites.
11809 * Since AES128 and AES256 share a single mask bit in the logic of
11810 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11811 kludge to work properly if AES128 is available and AES256 isn't.
11815 * Expand security boundary to match 1.1.1 module.
11819 * Remove redundant features: hash file source, editing of test vectors
11820 modify fipsld to use external fips_premain.c signature.
11824 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11825 run algorithm test programs.
11829 * Make algorithm test programs more tolerant of whitespace.
11833 * Have SSL/TLS server implementation tolerate "mismatched" record
11834 protocol version while receiving ClientHello even if the
11835 ClientHello is fragmented. (The server can't insist on the
11836 particular protocol version it has chosen before the ServerHello
11837 message has informed the client about his choice.)
11841 * Load error codes if they are not already present instead of using a
11842 static variable. This allows them to be cleanly unloaded and reloaded.
11846 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11848 * Introduce limits to prevent malicious keys being able to
11849 cause a denial of service. ([CVE-2006-2940])
11851 *Steve Henson, Bodo Moeller*
11853 * Fix ASN.1 parsing of certain invalid structures that can result
11854 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11856 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11857 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11859 * Fix SSL client code which could crash if connecting to a
11860 malicious SSLv2 server. ([CVE-2006-4343])
11862 *Tavis Ormandy and Will Drewry, Google Security Team*
11864 * Change ciphersuite string processing so that an explicit
11865 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11866 will no longer include "AES128-SHA"), and any other similar
11867 ciphersuite (same bitmap) from *other* protocol versions (so that
11868 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11869 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11870 changes from 0.9.8b and 0.9.8d.
11874 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11876 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11877 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11879 * Change the Unix randomness entropy gathering to use poll() when
11880 possible instead of select(), since the latter has some
11881 undesirable limitations.
11883 *Darryl Miles via Richard Levitte and Bodo Moeller*
11885 * Disable rogue ciphersuites:
11887 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11888 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11889 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11891 The latter two were purportedly from
11892 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11895 Also deactivate the remaining ciphersuites from
11896 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11897 unofficial, and the ID has long expired.
11901 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11902 dual-core machines) and other potential thread-safety issues.
11906 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11908 * Adapt fipsld and the build system to link against the validated FIPS
11909 module in FIPS mode.
11913 * Fixes for VC++ 2005 build under Windows.
11917 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11918 from a Windows bash shell such as MSYS. It is autodetected from the
11919 "config" script when run from a VC++ environment. Modify standard VC++
11920 build to use fipscanister.o from the GNU make build.
11924 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11926 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11927 The value now differs depending on if you build for FIPS or not.
11928 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11929 safely run with a non-FIPSed libcrypto, as it may crash because of
11930 the difference induced by this change.
11934 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11936 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11937 (part of SSL_OP_ALL). This option used to disable the
11938 countermeasure against man-in-the-middle protocol-version
11939 rollback in the SSL 2.0 server implementation, which is a bad
11940 idea. ([CVE-2005-2969])
11942 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11943 for Information Security, National Institute of Advanced Industrial
11944 Science and Technology [AIST, Japan)]*
11946 * Minimal support for X9.31 signatures and PSS padding modes. This is
11947 mainly for FIPS compliance and not fully integrated at this stage.
11951 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11952 the exponentiation using a fixed-length exponent. (Otherwise,
11953 the information leaked through timing could expose the secret key
11954 after many signatures; cf. Bleichenbacher's attack on DSA with
11959 * Make a new fixed-window mod_exp implementation the default for
11960 RSA, DSA, and DH private-key operations so that the sequence of
11961 squares and multiplies and the memory access pattern are
11962 independent of the particular secret key. This will mitigate
11963 cache-timing and potential related attacks.
11965 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11966 and this is automatically used by BN_mod_exp_mont() if the new flag
11967 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11968 will use this BN flag for private exponents unless the flag
11969 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11970 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11972 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11974 * Change the client implementation for SSLv23_method() and
11975 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11976 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11977 (Previously, the SSL 2.0 backwards compatible Client Hello
11978 message format would be used even with SSL_OP_NO_SSLv2.)
11982 * Add support for smime-type MIME parameter in S/MIME messages which some
11987 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11988 a threadsafe manner. Modify rsa code to use new function and add calls
11989 to dsa and dh code (which had race conditions before).
11993 * Include the fixed error library code in the C error file definitions
11994 instead of fixing them up at runtime. This keeps the error code
11995 structures constant.
11999 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12001 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12004 * Fixes for newer kerberos headers. NB: the casts are needed because
12005 the 'length' field is signed on one version and unsigned on another
12006 with no (?) obvious way to tell the difference, without these VC++
12007 complains. Also the "definition" of FAR (blank) is no longer included
12008 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12009 some needed definitions.
12013 * Undo Cygwin change.
12017 * Added support for proxy certificates according to RFC 3820.
12018 Because they may be a security thread to unaware applications,
12019 they must be explicitly allowed in run-time. See
12020 docs/HOWTO/proxy_certificates.txt for further information.
12024 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12026 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12027 server and client random values. Previously
12028 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12029 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12031 This change has negligible security impact because:
12033 1. Server and client random values still have 24 bytes of pseudo random
12036 2. Server and client random values are sent in the clear in the initial
12039 3. The master secret is derived using the premaster secret (48 bytes in
12040 size for static RSA ciphersuites) as well as client server and random
12043 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12046 *Stephen Henson, reported by UK NISCC*
12048 * Use Windows randomness collection on Cygwin.
12052 * Fix hang in EGD/PRNGD query when communication socket is closed
12053 prematurely by EGD/PRNGD.
12055 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12057 * Prompt for pass phrases when appropriate for PKCS12 input format.
12061 * Back-port of selected performance improvements from development
12062 branch, as well as improved support for PowerPC platforms.
12066 * Add lots of checks for memory allocation failure, error codes to indicate
12067 failure and freeing up memory if a failure occurs.
12069 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12071 * Add new -passin argument to dgst.
12075 * Perform some character comparisons of different types in X509_NAME_cmp:
12076 this is needed for some certificates that re-encode DNs into UTF8Strings
12077 (in violation of RFC3280) and can't or won't issue name rollover
12082 * Make an explicit check during certificate validation to see that
12083 the CA setting in each certificate on the chain is correct. As a
12084 side effect always do the following basic checks on extensions,
12085 not just when there's an associated purpose to the check:
12087 - if there is an unhandled critical extension (unless the user
12088 has chosen to ignore this fault)
12089 - if the path length has been exceeded (if one is set at all)
12090 - that certain extensions fit the associated purpose (if one has
12095 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12097 * Avoid a race condition when CRLs are checked in a multi threaded
12098 environment. This would happen due to the reordering of the revoked
12099 entries during signature checking and serial number lookup. Now the
12100 encoding is cached and the serial number sort performed under a lock.
12101 Add new STACK function sk_is_sorted().
12105 * Add Delta CRL to the extension code.
12109 * Various fixes to s3_pkt.c so alerts are sent properly.
12111 *David Holmes <d.holmes@f5.com>*
12113 * Reduce the chances of duplicate issuer name and serial numbers (in
12114 violation of RFC3280) using the OpenSSL certificate creation utilities.
12115 This is done by creating a random 64 bit value for the initial serial
12116 number when a serial number file is created or when a self signed
12117 certificate is created using 'openssl req -x509'. The initial serial
12118 number file is created using 'openssl x509 -next_serial' in CA.pl
12119 rather than being initialized to 1.
12123 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12125 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12126 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12128 *Joe Orton, Steve Henson*
12130 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12133 *Joe Orton, Steve Henson*
12135 * Make it possible to have multiple active certificates with the same
12136 subject in the CA index file. This is done only if the keyword
12137 'unique_subject' is set to 'no' in the main CA section (default
12138 if 'CA_default') of the configuration file. The value is saved
12139 with the database itself in a separate index attribute file,
12140 named like the index file with '.attr' appended to the name.
12144 * X509 verify fixes. Disable broken certificate workarounds when
12145 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12146 keyUsage extension present. Don't accept CRLs with unhandled critical
12147 extensions: since verify currently doesn't process CRL extensions this
12148 rejects a CRL with *any* critical extensions. Add new verify error codes
12153 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12154 A clarification of RFC2560 will require the use of OCTET STRINGs and
12155 some implementations cannot handle the current raw format. Since OpenSSL
12156 copies and compares OCSP nonces as opaque blobs without any attempt at
12157 parsing them this should not create any compatibility issues.
12161 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12162 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12163 this HMAC (and other) operations are several times slower than OpenSSL
12168 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12170 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12172 * Use the correct content when signing type "other".
12176 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12178 * Fix various bugs revealed by running the NISCC test suite:
12180 Stop out of bounds reads in the ASN1 code when presented with
12181 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12183 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12185 If verify callback ignores invalid public key errors don't try to check
12186 certificate signature with the NULL public key.
12190 * New -ignore_err option in ocsp application to stop the server
12191 exiting on the first error in a request.
12195 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12196 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12201 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12202 extra data after the compression methods not only for TLS 1.0
12203 but also for SSL 3.0 (as required by the specification).
12205 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12207 * Change X509_certificate_type() to mark the key as exported/exportable
12208 when it's 512 *bits* long, not 512 bytes.
12212 * Change AES_cbc_encrypt() so it outputs exact multiple of
12213 blocks during encryption.
12217 * Various fixes to base64 BIO and non blocking I/O. On write
12218 flushes were not handled properly if the BIO retried. On read
12219 data was not being buffered properly and had various logic bugs.
12220 This also affects blocking I/O when the data being decoded is a
12225 * Various S/MIME bugfixes and compatibility changes:
12226 output correct application/pkcs7 MIME type if
12227 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12228 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12229 of files as .eml work). Correctly handle very long lines in MIME
12234 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12236 * Countermeasure against the Klima-Pokorny-Rosa extension of
12237 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12238 a protocol version number mismatch like a decryption error
12239 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12243 * Turn on RSA blinding by default in the default implementation
12244 to avoid a timing attack. Applications that don't want it can call
12245 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12246 They would be ill-advised to do so in most cases.
12248 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12250 * Change RSA blinding code so that it works when the PRNG is not
12251 seeded (in this case, the secret RSA exponent is abused as
12252 an unpredictable seed -- if it is not unpredictable, there
12253 is no point in blinding anyway). Make RSA blinding thread-safe
12254 by remembering the creator's thread ID in rsa->blinding and
12255 having all other threads use local one-time blinding factors
12256 (this requires more computation than sharing rsa->blinding, but
12257 avoids excessive locking; and if an RSA object is not shared
12258 between threads, blinding will still be very fast).
12262 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12263 ENGINE as defaults for all supported algorithms irrespective of
12264 the 'flags' parameter. 'flags' is now honoured, so applications
12265 should make sure they are passing it correctly.
12269 * Target "mingw" now allows native Windows code to be generated in
12270 the Cygwin environment as well as with the MinGW compiler.
12274 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12276 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12277 via timing by performing a MAC computation even if incorrect
12278 block cipher padding has been found. This is a countermeasure
12279 against active attacks where the attacker has to distinguish
12280 between bad padding and a MAC verification error. ([CVE-2003-0078])
12282 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12283 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12284 Martin Vuagnoux (EPFL, Ilion)*
12286 * Make the no-err option work as intended. The intention with no-err
12287 is not to have the whole error stack handling routines removed from
12288 libcrypto, it's only intended to remove all the function name and
12289 reason texts, thereby removing some of the footprint that may not
12290 be interesting if those errors aren't displayed anyway.
12292 NOTE: it's still possible for any application or module to have its
12293 own set of error texts inserted. The routines are there, just not
12294 used by default when no-err is given.
12298 * Add support for FreeBSD on IA64.
12300 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12302 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12303 Kerberos function mit_des_cbc_cksum(). Before this change,
12304 the value returned by DES_cbc_cksum() was like the one from
12305 mit_des_cbc_cksum(), except the bytes were swapped.
12307 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12309 * Allow an application to disable the automatic SSL chain building.
12310 Before this a rather primitive chain build was always performed in
12311 ssl3_output_cert_chain(): an application had no way to send the
12312 correct chain if the automatic operation produced an incorrect result.
12314 Now the chain builder is disabled if either:
12316 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12318 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12320 The reasoning behind this is that an application would not want the
12321 auto chain building to take place if extra chain certificates are
12322 present and it might also want a means of sending no additional
12323 certificates (for example the chain has two certificates and the
12328 * Add the possibility to build without the ENGINE framework.
12330 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12332 * Under Win32 gmtime() can return NULL: check return value in
12333 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12337 * DSA routines: under certain error conditions uninitialized BN objects
12338 could be freed. Solution: make sure initialization is performed early
12339 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12340 Nils Larsch <nla@trustcenter.de> via PR#459)
12344 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12345 checked on reconnect on the client side, therefore session resumption
12346 could still fail with a "ssl session id is different" error. This
12347 behaviour is masked when SSL_OP_ALL is used due to
12348 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12349 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12350 followup to PR #377.
12354 * IA-32 assembler support enhancements: unified ELF targets, support
12355 for SCO/Caldera platforms, fix for Cygwin shared build.
12359 * Add support for FreeBSD on sparc64. As a consequence, support for
12360 FreeBSD on non-x86 processors is separate from x86 processors on
12361 the config script, much like the NetBSD support.
12363 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12365 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12367 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12370 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12371 code (06) was taken as the first octet of the session ID and the last
12372 octet was ignored consequently. As a result SSLv2 client side session
12373 caching could not have worked due to the session ID mismatch between
12375 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12380 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12381 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12386 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12387 seems that in spite of existing for more than a year, many application
12388 author have done nothing to provide the necessary callbacks, which
12389 means that this particular engine will not work properly anywhere.
12390 This is a very unfortunate situation which forces us, in the name
12391 of usability, to give the hw_ncipher.c a static lock, which is part
12393 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12394 appear in 0.9.8 or later. We EXPECT application authors to have
12395 dealt properly with this when 0.9.8 is released (unless we actually
12396 make such changes in the libcrypto locking code that changes will
12397 have to be made anyway).
12401 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12402 octets have been read, EOF or an error occurs. Without this change
12403 some truncated ASN1 structures will not produce an error.
12407 * Disable Heimdal support, since it hasn't been fully implemented.
12408 Still give the possibility to force the use of Heimdal, but with
12409 warnings and a request that patches get sent to openssl-dev.
12413 * Add the VC-CE target, introduce the WINCE sysname, and add
12414 INSTALL.WCE and appropriate conditionals to make it build.
12416 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12418 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12419 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12420 edit numbers of the version.
12422 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12424 * Introduce safe string copy and catenation functions
12425 (BUF_strlcpy() and BUF_strlcat()).
12427 *Ben Laurie (CHATS) and Richard Levitte*
12429 * Avoid using fixed-size buffers for one-line DNs.
12431 *Ben Laurie (CHATS)*
12433 * Add BUF_MEM_grow_clean() to avoid information leakage when
12434 resizing buffers containing secrets, and use where appropriate.
12436 *Ben Laurie (CHATS)*
12438 * Avoid using fixed size buffers for configuration file location.
12440 *Ben Laurie (CHATS)*
12442 * Avoid filename truncation for various CA files.
12444 *Ben Laurie (CHATS)*
12446 * Use sizeof in preference to magic numbers.
12448 *Ben Laurie (CHATS)*
12450 * Avoid filename truncation in cert requests.
12452 *Ben Laurie (CHATS)*
12454 * Add assertions to check for (supposedly impossible) buffer
12457 *Ben Laurie (CHATS)*
12459 * Don't cache truncated DNS entries in the local cache (this could
12460 potentially lead to a spoofing attack).
12462 *Ben Laurie (CHATS)*
12464 * Fix various buffers to be large enough for hex/decimal
12465 representations in a platform independent manner.
12467 *Ben Laurie (CHATS)*
12469 * Add CRYPTO_realloc_clean() to avoid information leakage when
12470 resizing buffers containing secrets, and use where appropriate.
12472 *Ben Laurie (CHATS)*
12474 * Add BIO_indent() to avoid much slightly worrying code to do
12477 *Ben Laurie (CHATS)*
12479 * Convert sprintf()/BIO_puts() to BIO_printf().
12481 *Ben Laurie (CHATS)*
12483 * buffer_gets() could terminate with the buffer only half
12486 *Ben Laurie (CHATS)*
12488 * Add assertions to prevent user-supplied crypto functions from
12489 overflowing internal buffers by having large block sizes, etc.
12491 *Ben Laurie (CHATS)*
12493 * New OPENSSL_assert() macro (similar to assert(), but enabled
12496 *Ben Laurie (CHATS)*
12498 * Eliminate unused copy of key in RC4.
12500 *Ben Laurie (CHATS)*
12502 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12504 *Ben Laurie (CHATS)*
12506 * Fix off-by-one error in EGD path.
12508 *Ben Laurie (CHATS)*
12510 * If RANDFILE path is too long, ignore instead of truncating.
12512 *Ben Laurie (CHATS)*
12514 * Eliminate unused and incorrectly sized X.509 structure
12517 *Ben Laurie (CHATS)*
12519 * Eliminate unused and dangerous function knumber().
12521 *Ben Laurie (CHATS)*
12523 * Eliminate unused and dangerous structure, KSSL_ERR.
12525 *Ben Laurie (CHATS)*
12527 * Protect against overlong session ID context length in an encoded
12528 session object. Since these are local, this does not appear to be
12531 *Ben Laurie (CHATS)*
12533 * Change from security patch (see 0.9.6e below) that did not affect
12534 the 0.9.6 release series:
12536 Remote buffer overflow in SSL3 protocol - an attacker could
12537 supply an oversized master key in Kerberos-enabled versions.
12540 *Ben Laurie (CHATS)*
12542 * Change the SSL kerb5 codes to match RFC 2712.
12546 * Make -nameopt work fully for req and add -reqopt switch.
12548 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12550 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12552 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12554 * Make sure tests can be performed even if the corresponding algorithms
12555 have been removed entirely. This was also the last step to make
12556 OpenSSL compilable with DJGPP under all reasonable conditions.
12558 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12560 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12561 to allow version independent disabling of normally unselected ciphers,
12562 which may be activated as a side-effect of selecting a single cipher.
12564 (E.g., cipher list string "RSA" enables ciphersuites that are left
12565 out of "ALL" because they do not provide symmetric encryption.
12566 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12568 *Lutz Jaenicke, Bodo Moeller*
12570 * Add appropriate support for separate platform-dependent build
12571 directories. The recommended way to make a platform-dependent
12572 build directory is the following (tested on Linux), maybe with
12575 # Place yourself outside of the OpenSSL source tree. In
12576 # this example, the environment variable OPENSSL_SOURCE
12577 # is assumed to contain the absolute OpenSSL source directory.
12578 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12579 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12580 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12581 mkdir -p `dirname $F`
12582 ln -s $OPENSSL_SOURCE/$F $F
12585 To be absolutely sure not to disturb the source tree, a "make clean"
12586 is a good thing. If it isn't successful, don't worry about it,
12587 it probably means the source directory is very clean.
12591 * Make sure any ENGINE control commands make local copies of string
12592 pointers passed to them whenever necessary. Otherwise it is possible
12593 the caller may have overwritten (or deallocated) the original string
12594 data when a later ENGINE operation tries to use the stored values.
12596 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12598 * Improve diagnostics in file reading and command-line digests.
12600 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12602 * Add AES modes CFB and OFB to the object database. Correct an
12603 error in AES-CFB decryption.
12607 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12608 allows existing EVP_CIPHER_CTX structures to be reused after
12609 calling `EVP_*Final()`. This behaviour is used by encryption
12610 BIOs and some applications. This has the side effect that
12611 applications must explicitly clean up cipher contexts with
12612 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12616 * Check the values of dna and dnb in bn_mul_recursive before calling
12617 bn_mul_comba (a non zero value means the a or b arrays do not contain
12618 n2 elements) and fallback to bn_mul_normal if either is not zero.
12622 * Fix escaping of non-ASCII characters when using the -subj option
12623 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12627 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12628 form for "surname", serialNumber has no short form.
12629 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12630 therefore remove "mail" short name for "internet 7".
12631 The OID for unique identifiers in X509 certificates is
12632 x500UniqueIdentifier, not uniqueIdentifier.
12633 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12637 * Add an "init" command to the ENGINE config module and auto initialize
12638 ENGINEs. Without any "init" command the ENGINE will be initialized
12639 after all ctrl commands have been executed on it. If init=1 the
12640 ENGINE is initialized at that point (ctrls before that point are run
12641 on the uninitialized ENGINE and after on the initialized one). If
12642 init=0 then the ENGINE will not be initialized at all.
12646 * Fix the 'app_verify_callback' interface so that the user-defined
12647 argument is actually passed to the callback: In the
12648 SSL_CTX_set_cert_verify_callback() prototype, the callback
12649 declaration has been changed from
12652 int (*cb)(X509_STORE_CTX *,void *);
12653 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12654 i=s->ctx->app_verify_callback(&ctx)
12655 has been changed into
12656 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12658 To update applications using SSL_CTX_set_cert_verify_callback(),
12659 a dummy argument can be added to their callback functions.
12661 *D. K. Smetters <smetters@parc.xerox.com>*
12663 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12665 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12667 * Add and OPENSSL_LOAD_CONF define which will cause
12668 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12669 This allows older applications to transparently support certain
12670 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12671 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12672 load the config file and OPENSSL_add_all_algorithms_conf() which will
12673 always load it have also been added.
12677 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12678 Adjust NIDs and EVP layer.
12680 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12682 * Config modules support in openssl utility.
12684 Most commands now load modules from the config file,
12685 though in a few (such as version) this isn't done
12686 because it couldn't be used for anything.
12688 In the case of ca and req the config file used is
12689 the same as the utility itself: that is the -config
12690 command line option can be used to specify an
12695 * Move default behaviour from OPENSSL_config(). If appname is NULL
12696 use "openssl_conf" if filename is NULL use default openssl config file.
12700 * Add an argument to OPENSSL_config() to allow the use of an alternative
12701 config section name. Add a new flag to tolerate a missing config file
12702 and move code to CONF_modules_load_file().
12706 * Support for crypto accelerator cards from Accelerated Encryption
12707 Processing, www.aep.ie. (Use engine 'aep')
12708 The support was copied from 0.9.6c [engine] and adapted/corrected
12709 to work with the new engine framework.
12711 *AEP Inc. and Richard Levitte*
12713 * Support for SureWare crypto accelerator cards from Baltimore
12714 Technologies. (Use engine 'sureware')
12715 The support was copied from 0.9.6c [engine] and adapted
12716 to work with the new engine framework.
12720 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12721 make the newer ENGINE framework commands for the CHIL engine work.
12723 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12725 * Make it possible to produce shared libraries on ReliantUNIX.
12727 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12729 * Add the configuration target debug-linux-ppro.
12730 Make 'openssl rsa' use the general key loading routines
12731 implemented in `apps.c`, and make those routines able to
12732 handle the key format FORMAT_NETSCAPE and the variant
12735 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12737 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12739 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12741 * Add -keyform to rsautl, and document -engine.
12743 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12745 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12746 BIO_R_NO_SUCH_FILE error code rather than the generic
12747 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12751 * Add new functions
12752 ERR_peek_last_error
12753 ERR_peek_last_error_line
12754 ERR_peek_last_error_line_data.
12755 These are similar to
12757 ERR_peek_error_line
12758 ERR_peek_error_line_data,
12759 but report on the latest error recorded rather than the first one
12760 still in the error queue.
12762 *Ben Laurie, Bodo Moeller*
12764 * default_algorithms option in ENGINE config module. This allows things
12766 default_algorithms = ALL
12767 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12771 * Preliminary ENGINE config module.
12775 * New experimental application configuration code.
12779 * Change the AES code to follow the same name structure as all other
12780 symmetric ciphers, and behave the same way. Move everything to
12781 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12783 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12785 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12787 *Ben Laurie and Theo de Raadt*
12789 * Add option to output public keys in req command.
12791 *Massimiliano Pala madwolf@openca.org*
12793 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12794 (up to about 10% better than before for P-192 and P-224).
12798 * New functions/macros
12800 SSL_CTX_set_msg_callback(ctx, cb)
12801 SSL_CTX_set_msg_callback_arg(ctx, arg)
12802 SSL_set_msg_callback(ssl, cb)
12803 SSL_set_msg_callback_arg(ssl, arg)
12805 to request calling a callback function
12807 void cb(int write_p, int version, int content_type,
12808 const void *buf, size_t len, SSL *ssl, void *arg)
12810 whenever a protocol message has been completely received
12811 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12812 protocol version according to which the SSL library interprets
12813 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12814 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12815 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12816 specification (change_cipher_spec(20), alert(21), handshake(22)).
12817 'buf' and 'len' point to the actual message, 'ssl' to the
12818 SSL object, and 'arg' is the application-defined value set by
12819 SSL[_CTX]_set_msg_callback_arg().
12821 'openssl s_client' and 'openssl s_server' have new '-msg' options
12822 to enable a callback that displays all protocol messages.
12826 * Change the shared library support so shared libraries are built as
12827 soon as the corresponding static library is finished, and thereby get
12828 openssl and the test programs linked against the shared library.
12829 This still only happens when the keyword "shard" has been given to
12830 the configuration scripts.
12832 NOTE: shared library support is still an experimental thing, and
12833 backward binary compatibility is still not guaranteed.
12835 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12837 * Add support for Subject Information Access extension.
12839 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12841 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12842 additional bytes when new memory had to be allocated, not just
12843 when reusing an existing buffer.
12847 * New command line and configuration option 'utf8' for the req command.
12848 This allows field values to be specified as UTF8 strings.
12852 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12853 runs for the former and machine-readable output for the latter.
12857 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12858 of the e-mail address in the DN (i.e., it will go into a certificate
12859 extension only). The new configuration file option 'email_in_dn = no'
12860 has the same effect.
12862 *Massimiliano Pala madwolf@openca.org*
12864 * Change all functions with names starting with `des_` to be starting
12865 with `DES_` instead. Add wrappers that are compatible with libdes,
12866 but are named `_ossl_old_des_*`. Finally, add macros that map the
12867 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12868 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12869 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12872 Since we provide two compatibility mappings, the user needs to
12873 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12874 compatibility is desired. The default (i.e., when that macro
12875 isn't defined) is OpenSSL 0.9.6c compatibility.
12877 There are also macros that enable and disable the support of old
12878 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12879 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12880 are defined, the default will apply: to support the old des routines.
12882 In either case, one must include openssl/des.h to get the correct
12883 definitions. Do not try to just include openssl/des_old.h, that
12886 NOTE: This is a major break of an old API into a new one. Software
12887 authors are encouraged to switch to the `DES_` style functions. Some
12888 time in the future, des_old.h and the libdes compatibility functions
12889 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12890 default), and then completely removed.
12894 * Test for certificates which contain unsupported critical extensions.
12895 If such a certificate is found during a verify operation it is
12896 rejected by default: this behaviour can be overridden by either
12897 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12898 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12899 X509_supported_extension() has also been added which returns 1 if a
12900 particular extension is supported.
12904 * Modify the behaviour of EVP cipher functions in similar way to digests
12905 to retain compatibility with existing code.
12909 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12910 compatibility with existing code. In particular the 'ctx' parameter does
12911 not have to be to be initialized before the call to EVP_DigestInit() and
12912 it is tidied up after a call to EVP_DigestFinal(). New function
12913 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12914 EVP_MD_CTX_copy() changed to not require the destination to be
12915 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12916 requires the destination to be valid.
12918 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12919 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12923 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12924 so that complete 'Handshake' protocol structures are kept in memory
12925 instead of overwriting 'msg_type' and 'length' with 'body' data.
12929 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12931 *Massimo Santin via Richard Levitte*
12933 * Major restructuring to the underlying ENGINE code. This includes
12934 reduction of linker bloat, separation of pure "ENGINE" manipulation
12935 (initialisation, etc) from functionality dealing with implementations
12936 of specific crypto interfaces. This change also introduces integrated
12937 support for symmetric ciphers and digest implementations - so ENGINEs
12938 can now accelerate these by providing EVP_CIPHER and EVP_MD
12939 implementations of their own. This is detailed in
12940 [crypto/engine/README.md](crypto/engine/README.md)
12941 as it couldn't be adequately described here. However, there are a few
12942 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12943 were changed in the original introduction of ENGINE code have now
12944 reverted back - the hooking from this code to ENGINE is now a good
12945 deal more passive and at run-time, operations deal directly with
12946 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12947 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12948 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12949 they were not being used by the framework as there is no concept of a
12950 BIGNUM_METHOD and they could not be generalised to the new
12951 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12952 ENGINE_cpy() has been removed as it cannot be consistently defined in
12957 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12961 * Change mkdef.pl to sort symbols that get the same entry number,
12962 and make sure the automatically generated functions `ERR_load_*`
12963 become part of libeay.num as well.
12967 * New function SSL_renegotiate_pending(). This returns true once
12968 renegotiation has been requested (either SSL_renegotiate() call
12969 or HelloRequest/ClientHello received from the peer) and becomes
12970 false once a handshake has been completed.
12971 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12972 sends a HelloRequest, but does not ensure that a handshake takes
12973 place. SSL_renegotiate_pending() is useful for checking if the
12974 client has followed the request.)
12978 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12979 By default, clients may request session resumption even during
12980 renegotiation (if session ID contexts permit); with this option,
12981 session resumption is possible only in the first handshake.
12983 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12984 more bits available for options that should not be part of
12985 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12989 * Add some demos for certificate and certificate request creation.
12993 * Make maximum certificate chain size accepted from the peer application
12994 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12995 "Douglas E. Engert" <deengert@anl.gov>.
12999 * Add support for shared libraries for Unixware-7
13000 (Boyd Lynn Gerber <gerberb@zenez.com>).
13004 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13005 be done prior to destruction. Use this to unload error strings from
13006 ENGINEs that load their own error strings. NB: This adds two new API
13007 functions to "get" and "set" this destroy handler in an ENGINE.
13011 * Alter all existing ENGINE implementations (except "openssl" and
13012 "openbsd") to dynamically instantiate their own error strings. This
13013 makes them more flexible to be built both as statically-linked ENGINEs
13014 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13015 Also, add stub code to each that makes building them as self-contained
13016 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13020 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13021 implementations into applications that are completely implemented in
13022 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13023 commands that can be used to configure what shared-library to load and
13024 to control aspects of the way it is handled. Also, made an update to
13025 the [README-Engine.md](README-Engine.md) file
13026 that brings its information up-to-date and
13027 provides some information and instructions on the "dynamic" ENGINE
13028 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13032 * Make it possible to unload ranges of ERR strings with a new
13033 "ERR_unload_strings" function.
13037 * Add a copy() function to EVP_MD.
13041 * Make EVP_MD routines take a context pointer instead of just the
13042 md_data void pointer.
13046 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13047 that the digest can only process a single chunk of data
13048 (typically because it is provided by a piece of
13049 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13050 is only going to provide a single chunk of data, and hence the
13051 framework needn't accumulate the data for oneshot drivers.
13055 * As with "ERR", make it possible to replace the underlying "ex_data"
13056 functions. This change also alters the storage and management of global
13057 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13058 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13059 index counters. The API functions that use this state have been changed
13060 to take a "class_index" rather than pointers to the class's local STACK
13061 and counter, and there is now an API function to dynamically create new
13062 classes. This centralisation allows us to (a) plug a lot of the
13063 thread-safety problems that existed, and (b) makes it possible to clean
13064 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13065 such data would previously have always leaked in application code and
13066 workarounds were in place to make the memory debugging turn a blind eye
13067 to it. Application code that doesn't use this new function will still
13068 leak as before, but their memory debugging output will announce it now
13069 rather than letting it slide.
13071 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13072 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13073 has a return value to indicate success or failure.
13077 * Make it possible to replace the underlying "ERR" functions such that the
13078 global state (2 LHASH tables and 2 locks) is only used by the "default"
13079 implementation. This change also adds two functions to "get" and "set"
13080 the implementation prior to it being automatically set the first time
13081 any other ERR function takes place. Ie. an application can call "get",
13082 pass the return value to a module it has just loaded, and that module
13083 can call its own "set" function using that value. This means the
13084 module's "ERR" operations will use (and modify) the error state in the
13085 application and not in its own statically linked copy of OpenSSL code.
13089 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13090 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13091 the operation, and provides a more encapsulated way for external code
13092 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13093 to use these functions rather than manually incrementing the counts.
13095 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13099 * Add EVP test program.
13103 * Add symmetric cipher support to ENGINE. Expect the API to change!
13107 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13108 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13109 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13110 These allow a CRL to be built without having to access X509_CRL fields
13111 directly. Modify 'ca' application to use new functions.
13115 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13116 bug workarounds. Rollback attack detection is a security feature.
13117 The problem will only arise on OpenSSL servers when TLSv1 is not
13118 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13119 Software authors not wanting to support TLSv1 will have special reasons
13120 for their choice and can explicitly enable this option.
13122 *Bodo Moeller, Lutz Jaenicke*
13124 * Rationalise EVP so it can be extended: don't include a union of
13125 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13126 (similar to those existing for EVP_CIPHER_CTX).
13131 EVP_MD_CTX_init(&md); /* new function call */
13132 EVP_DigestInit(&md, EVP_sha1());
13133 EVP_DigestUpdate(&md, in, len);
13134 EVP_DigestFinal(&md, out, NULL);
13135 EVP_MD_CTX_cleanup(&md); /* new function call */
13139 * Make DES key schedule conform to the usual scheme, as well as
13140 correcting its structure. This means that calls to DES functions
13141 now have to pass a pointer to a des_key_schedule instead of a
13142 plain des_key_schedule (which was actually always a pointer
13145 des_key_schedule ks;
13147 des_set_key_checked(..., &ks);
13148 des_ncbc_encrypt(..., &ks, ...);
13150 (Note that a later change renames 'des_...' into 'DES_...'.)
13154 * Initial reduction of linker bloat: the use of some functions, such as
13155 PEM causes large amounts of unused functions to be linked in due to
13156 poor organisation. For example pem_all.c contains every PEM function
13157 which has a knock on effect of linking in large amounts of (unused)
13158 ASN1 code. Grouping together similar functions and splitting unrelated
13159 functions prevents this.
13163 * Cleanup of EVP macros.
13167 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13168 correct `_ecb suffix`.
13172 * Add initial OCSP responder support to ocsp application. The
13173 revocation information is handled using the text based index
13174 use by the ca application. The responder can either handle
13175 requests generated internally, supplied in files (for example
13176 via a CGI script) or using an internal minimal server.
13180 * Add configuration choices to get zlib compression for TLS.
13184 * Changes to Kerberos SSL for RFC 2712 compliance:
13185 1. Implemented real KerberosWrapper, instead of just using
13186 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13187 2. Implemented optional authenticator field of KerberosWrapper.
13189 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13190 and authenticator structs; see crypto/krb5/.
13192 Generalized Kerberos calls to support multiple Kerberos libraries.
13193 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13194 via Richard Levitte*
13196 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13197 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13198 values for each of the key sizes rather than having just
13199 parameters (and 'speed' generating keys each time).
13203 * Speed up EVP routines.
13206 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13207 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13208 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13209 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13211 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13212 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13213 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13216 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13218 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13222 * Added the OS2-EMX target.
13224 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13226 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13227 New functions to support `NCONF` routines in extension code.
13228 New function `CONF_set_nconf()`
13229 to allow functions which take an `NCONF` to also handle the old `LHASH`
13230 structure: this means that the old `CONF` compatible routines can be
13231 retained (in particular w.rt. extensions) without having to duplicate the
13232 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13236 * Enhance the general user interface with mechanisms for inner control
13237 and with possibilities to have yes/no kind of prompts.
13241 * Change all calls to low-level digest routines in the library and
13242 applications to use EVP. Add missing calls to HMAC_cleanup() and
13243 don't assume HMAC_CTX can be copied using memcpy().
13245 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13247 * Add the possibility to control engines through control names but with
13248 arbitrary arguments instead of just a string.
13249 Change the key loaders to take a UI_METHOD instead of a callback
13250 function pointer. NOTE: this breaks binary compatibility with earlier
13251 versions of OpenSSL [engine].
13252 Adapt the nCipher code for these new conditions and add a card insertion
13257 * Enhance the general user interface with mechanisms to better support
13258 dialog box interfaces, application-defined prompts, the possibility
13259 to use defaults (for example default passwords from somewhere else)
13260 and interrupts/cancellations.
13264 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13265 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13269 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13270 tidy up some unnecessarily weird code in 'sk_new()').
13272 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13274 * Change the key loading routines for ENGINEs to use the same kind
13275 callback (pem_password_cb) as all other routines that need this
13280 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13281 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13282 than this minimum value is recommended.
13286 * New random seeder for OpenVMS, using the system process statistics
13287 that are easily reachable.
13291 * Windows apparently can't transparently handle global
13292 variables defined in DLLs. Initialisations such as:
13294 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13296 won't compile. This is used by the any applications that need to
13297 declare their own ASN1 modules. This was fixed by adding the option
13298 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13299 needed for static libraries under Win32.
13303 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13304 setting of purpose and trust fields. New X509_STORE trust and
13305 purpose functions and tidy up setting in other SSL functions.
13309 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13310 structure. These are inherited by X509_STORE_CTX when it is
13311 initialised. This allows various defaults to be set in the
13312 X509_STORE structure (such as flags for CRL checking and custom
13313 purpose or trust settings) for functions which only use X509_STORE_CTX
13314 internally such as S/MIME.
13316 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13317 trust settings if they are not set in X509_STORE. This allows X509_STORE
13318 purposes and trust (in S/MIME for example) to override any set by default.
13320 Add command line options for CRL checking to smime, s_client and s_server
13325 * Initial CRL based revocation checking. If the CRL checking flag(s)
13326 are set then the CRL is looked up in the X509_STORE structure and
13327 its validity and signature checked, then if the certificate is found
13328 in the CRL the verify fails with a revoked error.
13330 Various new CRL related callbacks added to X509_STORE_CTX structure.
13332 Command line options added to 'verify' application to support this.
13334 This needs some additional work, such as being able to handle multiple
13335 CRLs with different times, extension based lookup (rather than just
13336 by subject name) and ultimately more complete V2 CRL extension
13341 * Add a general user interface API (crypto/ui/). This is designed
13342 to replace things like des_read_password and friends (backward
13343 compatibility functions using this new API are provided).
13344 The purpose is to remove prompting functions from the DES code
13345 section as well as provide for prompting through dialog boxes in
13346 a window system and the like.
13350 * Add "ex_data" support to ENGINE so implementations can add state at a
13351 per-structure level rather than having to store it globally.
13355 * Make it possible for ENGINE structures to be copied when retrieved by
13356 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13357 This causes the "original" ENGINE structure to act like a template,
13358 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13359 operational state can be localised to each ENGINE structure, despite the
13360 fact they all share the same "methods". New ENGINE structures returned in
13361 this case have no functional references and the return value is the single
13362 structural reference. This matches the single structural reference returned
13363 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13368 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13369 needs to match any other type at all we need to manually clear the
13374 * Changes to the "openssl engine" utility to include;
13375 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13376 about an ENGINE's available control commands.
13377 - executing control commands from command line arguments using the
13378 '-pre' and '-post' switches. '-post' is only used if '-t' is
13379 specified and the ENGINE is successfully initialised. The syntax for
13380 the individual commands are colon-separated, for example;
13381 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13385 * New dynamic control command support for ENGINEs. ENGINEs can now
13386 declare their own commands (numbers), names (strings), descriptions,
13387 and input types for run-time discovery by calling applications. A
13388 subset of these commands are implicitly classed as "executable"
13389 depending on their input type, and only these can be invoked through
13390 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13391 can be based on user input, config files, etc). The distinction is
13392 that "executable" commands cannot return anything other than a boolean
13393 result and can only support numeric or string input, whereas some
13394 discoverable commands may only be for direct use through
13395 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13396 pointers, or other custom uses. The "executable" commands are to
13397 support parameterisations of ENGINE behaviour that can be
13398 unambiguously defined by ENGINEs and used consistently across any
13399 OpenSSL-based application. Commands have been added to all the
13400 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13401 control over shared-library paths without source code alterations.
13405 * Changed all ENGINE implementations to dynamically allocate their
13406 ENGINEs rather than declaring them statically. Apart from this being
13407 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13408 this also allows the implementations to compile without using the
13409 internal engine_int.h header.
13413 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13414 'const' value. Any code that should be able to modify a RAND_METHOD
13415 should already have non-const pointers to it (ie. they should only
13416 modify their own ones).
13420 * Made a variety of little tweaks to the ENGINE code.
13421 - "atalla" and "ubsec" string definitions were moved from header files
13422 to C code. "nuron" string definitions were placed in variables
13423 rather than hard-coded - allowing parameterisation of these values
13424 later on via ctrl() commands.
13425 - Removed unused "#if 0"'d code.
13426 - Fixed engine list iteration code so it uses ENGINE_free() to release
13427 structural references.
13428 - Constified the RAND_METHOD element of ENGINE structures.
13429 - Constified various get/set functions as appropriate and added
13430 missing functions (including a catch-all ENGINE_cpy that duplicates
13431 all ENGINE values onto a new ENGINE except reference counts/state).
13432 - Removed NULL parameter checks in get/set functions. Setting a method
13433 or function to NULL is a way of cancelling out a previously set
13434 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13435 and doesn't justify the extra error symbols and code.
13436 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13437 flags from engine_int.h to engine.h.
13438 - Changed prototypes for ENGINE handler functions (init(), finish(),
13439 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13443 * Implement binary inversion algorithm for BN_mod_inverse in addition
13444 to the algorithm using long division. The binary algorithm can be
13445 used only if the modulus is odd. On 32-bit systems, it is faster
13446 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13447 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13448 up to 450 bits. In 64-bit environments, the binary algorithm
13449 appears to be advantageous for much longer moduli; here we use it
13450 for moduli up to 2048 bits.
13454 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13455 could not support the combine flag in choice fields.
13459 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13460 extensions from a certificate request to the certificate.
13464 * Allow multiple 'certopt' and 'nameopt' options to be separated
13465 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13466 file: this allows the display of the certificate about to be
13467 signed to be customised, to allow certain fields to be included
13468 or excluded and extension details. The old system didn't display
13469 multicharacter strings properly, omitted fields not in the policy
13470 and couldn't display additional details such as extensions.
13474 * Function EC_POINTs_mul for multiple scalar multiplication
13475 of an arbitrary number of elliptic curve points
13476 \sum scalars[i]*points[i],
13477 optionally including the generator defined for the EC_GROUP:
13478 scalar*generator + \sum scalars[i]*points[i].
13480 EC_POINT_mul is a simple wrapper function for the typical case
13481 that the point list has just one item (besides the optional
13486 * First EC_METHODs for curves over GF(p):
13488 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13489 operations and provides various method functions that can also
13490 operate with faster implementations of modular arithmetic.
13492 EC_GFp_mont_method() reuses most functions that are part of
13493 EC_GFp_simple_method, but uses Montgomery arithmetic.
13495 *Bodo Moeller; point addition and point doubling
13496 implementation directly derived from source code provided by
13497 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13499 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13500 crypto/ec/ec_lib.c):
13502 Curves are EC_GROUP objects (with an optional group generator)
13503 based on EC_METHODs that are built into the library.
13505 Points are EC_POINT objects based on EC_GROUP objects.
13507 Most of the framework would be able to handle curves over arbitrary
13508 finite fields, but as there are no obvious types for fields other
13509 than GF(p), some functions are limited to that for now.
13513 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13514 that the file contains a complete HTTP response.
13518 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13519 change the def and num file printf format specifier from "%-40sXXX"
13520 to "%-39s XXX". The latter will always guarantee a space after the
13521 field while the former will cause them to run together if the field
13522 is 40 of more characters long.
13526 * Constify the cipher and digest 'method' functions and structures
13527 and modify related functions to take constant EVP_MD and EVP_CIPHER
13532 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13533 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13537 * Modify `EVP_Digest*()` routines so they now return values. Although the
13538 internal software routines can never fail additional hardware versions
13543 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13545 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13546 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13549 ERR_R_NESTED_ASN1_ERROR
13551 ERR_R_MISSING_ASN1_EOS
13552 were 4 .. 9, conflicting with
13553 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13555 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13556 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13558 Add new error code 'ERR_R_INTERNAL_ERROR'.
13562 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13567 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13568 sets the subject name for a new request or supersedes the
13569 subject name in a given request. Formats that can be parsed are
13570 'CN=Some Name, OU=myOU, C=IT'
13572 'CN=Some Name/OU=myOU/C=IT'.
13574 Add options '-batch' and '-verbose' to 'openssl req'.
13576 *Massimiliano Pala <madwolf@hackmasters.net>*
13578 * Introduce the possibility to access global variables through
13579 functions on platform were that's the best way to handle exporting
13580 global variables in shared libraries. To enable this functionality,
13581 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13582 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13583 is normally done by Configure or something similar).
13585 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13586 in the source file (foo.c) like this:
13588 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13589 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13591 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13592 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13594 OPENSSL_DECLARE_GLOBAL(int,foo);
13595 #define foo OPENSSL_GLOBAL_REF(foo)
13596 OPENSSL_DECLARE_GLOBAL(double,bar);
13597 #define bar OPENSSL_GLOBAL_REF(bar)
13599 The #defines are very important, and therefore so is including the
13600 header file everywhere where the defined globals are used.
13602 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13603 of ASN.1 items, but that structure is a bit different.
13605 The largest change is in util/mkdef.pl which has been enhanced with
13606 better and easier to understand logic to choose which symbols should
13607 go into the Windows .def files as well as a number of fixes and code
13608 cleanup (among others, algorithm keywords are now sorted
13609 lexicographically to avoid constant rewrites).
13613 * In BN_div() keep a copy of the sign of 'num' before writing the
13614 result to 'rm' because if rm==num the value will be overwritten
13615 and produce the wrong result if 'num' is negative: this caused
13616 problems with BN_mod() and BN_nnmod().
13620 * Function OCSP_request_verify(). This checks the signature on an
13621 OCSP request and verifies the signer certificate. The signer
13622 certificate is just checked for a generic purpose and OCSP request
13627 * Add OCSP_check_validity() function to check the validity of OCSP
13628 responses. OCSP responses are prepared in real time and may only
13629 be a few seconds old. Simply checking that the current time lies
13630 between thisUpdate and nextUpdate max reject otherwise valid responses
13631 caused by either OCSP responder or client clock inaccuracy. Instead
13632 we allow thisUpdate and nextUpdate to fall within a certain period of
13633 the current time. The age of the response can also optionally be
13634 checked. Two new options -validity_period and -status_age added to
13639 * If signature or public key algorithm is unrecognized print out its
13640 OID rather that just UNKNOWN.
13644 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13645 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13646 ID to be generated from the issuer certificate alone which can then be
13647 passed to OCSP_id_issuer_cmp().
13651 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13652 ASN1 modules to export functions returning ASN1_ITEM pointers
13653 instead of the ASN1_ITEM structures themselves. This adds several
13654 new macros which allow the underlying ASN1 function/structure to
13655 be accessed transparently. As a result code should not use ASN1_ITEM
13656 references directly (such as &X509_it) but instead use the relevant
13657 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13658 use of the new ASN1 code on platforms where exporting structures
13659 is problematical (for example in shared libraries) but exporting
13660 functions returning pointers to structures is not.
13664 * Add support for overriding the generation of SSL/TLS session IDs.
13665 These callbacks can be registered either in an SSL_CTX or per SSL.
13666 The purpose of this is to allow applications to control, if they wish,
13667 the arbitrary values chosen for use as session IDs, particularly as it
13668 can be useful for session caching in multiple-server environments. A
13669 command-line switch for testing this (and any client code that wishes
13670 to use such a feature) has been added to "s_server".
13672 *Geoff Thorpe, Lutz Jaenicke*
13674 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13675 of the form `#if defined(...) || defined(...) || ...` and
13676 `#if !defined(...) && !defined(...) && ...`. This also avoids
13677 the growing number of special cases it was previously handling.
13681 * Make all configuration macros available for application by making
13682 sure they are available in opensslconf.h, by giving them names starting
13683 with `OPENSSL_` to avoid conflicts with other packages and by making
13684 sure e_os2.h will cover all platform-specific cases together with
13686 Additionally, it is now possible to define configuration/platform-
13687 specific names (called "system identities"). In the C code, these
13688 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13689 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13690 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13695 * New option -set_serial to 'req' and 'x509' this allows the serial
13696 number to use to be specified on the command line. Previously self
13697 signed certificates were hard coded with serial number 0 and the
13698 CA options of 'x509' had to use a serial number in a file which was
13703 * New options to 'ca' utility to support V2 CRL entry extensions.
13704 Currently CRL reason, invalidity date and hold instruction are
13705 supported. Add new CRL extensions to V3 code and some new objects.
13709 * New function EVP_CIPHER_CTX_set_padding() this is used to
13710 disable standard block padding (aka PKCS#5 padding) in the EVP
13711 API, which was previously mandatory. This means that the data is
13712 not padded in any way and so the total length much be a multiple
13713 of the block size, otherwise an error occurs.
13717 * Initial (incomplete) OCSP SSL support.
13721 * New function OCSP_parse_url(). This splits up a URL into its host,
13722 port and path components: primarily to parse OCSP URLs. New -url
13723 option to ocsp utility.
13727 * New nonce behavior. The return value of OCSP_check_nonce() now
13728 reflects the various checks performed. Applications can decide
13729 whether to tolerate certain situations such as an absent nonce
13730 in a response when one was present in a request: the ocsp application
13731 just prints out a warning. New function OCSP_add1_basic_nonce()
13732 this is to allow responders to include a nonce in a response even if
13733 the request is nonce-less.
13737 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13738 skipped when using openssl x509 multiple times on a single input file,
13739 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13743 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13744 set string type: to handle setting ASN1_TIME structures. Fix ca
13745 utility to correctly initialize revocation date of CRLs.
13749 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13750 the clients preferred ciphersuites and rather use its own preferences.
13751 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13752 Internet Explorer by ensuring unchanged hash method during stepup.
13753 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13757 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13758 to aes and add a new 'exist' option to print out symbols that don't
13763 * Additional options to ocsp utility to allow flags to be set and
13764 additional certificates supplied.
13768 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13769 OCSP client a number of certificate to only verify the response
13774 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13775 handle the new API. Currently only ECB, CBC modes supported. Add new
13778 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13779 Encryption Standard (AES) Ciphersuites for Transport Layer
13780 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13781 not enabled by default and were not part of the "ALL" ciphersuite
13782 alias because they were not yet official; they could be
13783 explicitly requested by specifying the "AESdraft" ciphersuite
13784 group alias. In the final release of OpenSSL 0.9.7, the group
13785 alias is called "AES" and is part of "ALL".)
13787 *Ben Laurie, Steve Henson, Bodo Moeller*
13789 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13790 request to response.
13794 * Functions for OCSP responders. OCSP_request_onereq_count(),
13795 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13796 extract information from a certificate request. OCSP_response_create()
13797 creates a response and optionally adds a basic response structure.
13798 OCSP_basic_add1_status() adds a complete single response to a basic
13799 response and returns the OCSP_SINGLERESP structure just added (to allow
13800 extensions to be included for example). OCSP_basic_add1_cert() adds a
13801 certificate to a basic response and OCSP_basic_sign() signs a basic
13802 response with various flags. New helper functions ASN1_TIME_check()
13803 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13804 (converts ASN1_TIME to GeneralizedTime).
13808 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13809 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13810 structure from a certificate. X509_pubkey_digest() digests the public_key
13811 contents: this is used in various key identifiers.
13815 * Make sk_sort() tolerate a NULL argument.
13817 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13819 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13820 passed by the function are trusted implicitly. If any of them signed the
13821 response then it is assumed to be valid and is not verified.
13825 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13826 to data. This was previously part of the PKCS7 ASN1 code. This
13827 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13828 *Steve Henson, reported by Kenneth R. Robinette
13829 <support@securenetterm.com>*
13831 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13832 routines: without these tracing memory leaks is very painful.
13833 Fix leaks in PKCS12 and PKCS7 routines.
13837 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13838 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13839 effectively meant GeneralizedTime would never be used. Now it
13840 is initialised to -1 but X509_time_adj() now has to check the value
13841 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13842 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13843 *Steve Henson, reported by Kenneth R. Robinette
13844 <support@securenetterm.com>*
13846 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13847 result in a zero length in the ASN1_INTEGER structure which was
13848 not consistent with the structure when d2i_ASN1_INTEGER() was used
13849 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13850 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13851 where it did not print out a minus for negative ASN1_INTEGER.
13855 * Add summary printout to ocsp utility. The various functions which
13856 convert status values to strings have been renamed to:
13857 OCSP_response_status_str(), OCSP_cert_status_str() and
13858 OCSP_crl_reason_str() and are no longer static. New options
13859 to verify nonce values and to disable verification. OCSP response
13860 printout format cleaned up.
13864 * Add additional OCSP certificate checks. These are those specified
13865 in RFC2560. This consists of two separate checks: the CA of the
13866 certificate being checked must either be the OCSP signer certificate
13867 or the issuer of the OCSP signer certificate. In the latter case the
13868 OCSP signer certificate must contain the OCSP signing extended key
13869 usage. This check is performed by attempting to match the OCSP
13870 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13871 in the OCSP_CERTID structures of the response.
13875 * Initial OCSP certificate verification added to OCSP_basic_verify()
13876 and related routines. This uses the standard OpenSSL certificate
13877 verify routines to perform initial checks (just CA validity) and
13878 to obtain the certificate chain. Then additional checks will be
13879 performed on the chain. Currently the root CA is checked to see
13880 if it is explicitly trusted for OCSP signing. This is used to set
13881 a root CA as a global signing root: that is any certificate that
13882 chains to that CA is an acceptable OCSP signing certificate.
13886 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13887 extensions from a separate configuration file.
13888 As when reading extensions from the main configuration file,
13889 the '-extensions ...' option may be used for specifying the
13892 *Massimiliano Pala <madwolf@comune.modena.it>*
13894 * New OCSP utility. Allows OCSP requests to be generated or
13895 read. The request can be sent to a responder and the output
13896 parsed, outputted or printed in text form. Not complete yet:
13897 still needs to check the OCSP response validity.
13901 * New subcommands for 'openssl ca':
13902 `openssl ca -status <serial>` prints the status of the cert with
13903 the given serial number (according to the index file).
13904 `openssl ca -updatedb` updates the expiry status of certificates
13907 *Massimiliano Pala <madwolf@comune.modena.it>*
13909 * New '-newreq-nodes' command option to CA.pl. This is like
13910 '-newreq', but calls 'openssl req' with the '-nodes' option
13911 so that the resulting key is not encrypted.
13913 *Damien Miller <djm@mindrot.org>*
13915 * New configuration for the GNU Hurd.
13917 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13919 * Initial code to implement OCSP basic response verify. This
13920 is currently incomplete. Currently just finds the signer's
13921 certificate and verifies the signature on the response.
13925 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13926 value of OPENSSLDIR. This is available via the new '-d' option
13927 to 'openssl version', and is also included in 'openssl version -a'.
13931 * Allowing defining memory allocation callbacks that will be given
13932 file name and line number information in additional arguments
13933 (a `const char*` and an int). The basic functionality remains, as
13934 well as the original possibility to just replace malloc(),
13935 realloc() and free() by functions that do not know about these
13936 additional arguments. To register and find out the current
13937 settings for extended allocation functions, the following
13938 functions are provided:
13940 CRYPTO_set_mem_ex_functions
13941 CRYPTO_set_locked_mem_ex_functions
13942 CRYPTO_get_mem_ex_functions
13943 CRYPTO_get_locked_mem_ex_functions
13945 These work the same way as CRYPTO_set_mem_functions and friends.
13946 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13947 extended allocation function is enabled.
13948 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13949 a conventional allocation function is enabled.
13951 *Richard Levitte, Bodo Moeller*
13953 * Finish off removing the remaining LHASH function pointer casts.
13954 There should no longer be any prototype-casting required when using
13955 the LHASH abstraction, and any casts that remain are "bugs". See
13956 the callback types and macros at the head of lhash.h for details
13957 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13961 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13962 If /dev/[u]random devices are not available or do not return enough
13963 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13965 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13966 /etc/entropy will be queried once each in this sequence, querying stops
13967 when enough entropy was collected without querying more sockets.
13971 * Change the Unix RAND_poll() variant to be able to poll several
13972 random devices, as specified by DEVRANDOM, until a sufficient amount
13973 of data has been collected. We spend at most 10 ms on each file
13974 (select timeout) and read in non-blocking mode. DEVRANDOM now
13975 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13976 (previously it was just the string "/dev/urandom"), so on typical
13977 platforms the 10 ms delay will never occur.
13978 Also separate out the Unix variant to its own file, rand_unix.c.
13979 For VMS, there's a currently-empty rand_vms.c.
13983 * Move OCSP client related routines to ocsp_cl.c. These
13984 provide utility functions which an application needing
13985 to issue a request to an OCSP responder and analyse the
13986 response will typically need: as opposed to those which an
13987 OCSP responder itself would need which will be added later.
13989 OCSP_request_sign() signs an OCSP request with an API similar
13990 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13991 response. OCSP_response_get1_basic() extracts basic response
13992 from response. OCSP_resp_find_status(): finds and extracts status
13993 information from an OCSP_CERTID structure (which will be created
13994 when the request structure is built). These are built from lower
13995 level functions which work on OCSP_SINGLERESP structures but
13996 won't normally be used unless the application wishes to examine
13997 extensions in the OCSP response for example.
13999 Replace nonce routines with a pair of functions.
14000 OCSP_request_add1_nonce() adds a nonce value and optionally
14001 generates a random value. OCSP_check_nonce() checks the
14002 validity of the nonce in an OCSP response.
14006 * Change function OCSP_request_add() to OCSP_request_add0_id().
14007 This doesn't copy the supplied OCSP_CERTID and avoids the
14008 need to free up the newly created id. Change return type
14009 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14010 This can then be used to add extensions to the request.
14011 Deleted OCSP_request_new(), since most of its functionality
14012 is now in OCSP_REQUEST_new() (and the case insensitive name
14013 clash) apart from the ability to set the request name which
14014 will be added elsewhere.
14018 * Update OCSP API. Remove obsolete extensions argument from
14019 various functions. Extensions are now handled using the new
14020 OCSP extension code. New simple OCSP HTTP function which
14021 can be used to send requests and parse the response.
14025 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14026 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14027 uses the special reorder version of SET OF to sort the attributes
14028 and reorder them to match the encoded order. This resolves a long
14029 standing problem: a verify on a PKCS7 structure just after signing
14030 it used to fail because the attribute order did not match the
14031 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14032 it uses the received order. This is necessary to tolerate some broken
14033 software that does not order SET OF. This is handled by encoding
14034 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14035 to produce the required SET OF.
14039 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14040 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14041 files to get correct declarations of the ASN.1 item variables.
14045 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14046 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14047 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14048 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14049 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14050 ASN1_ITEM and no wrapper functions.
14054 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14055 replace the old function pointer based I/O routines. Change most of
14056 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14060 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14061 lines, recognize more "algorithms" that can be deselected, and make
14062 it complain about algorithm deselection that isn't recognised.
14066 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14067 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14068 to use new functions. Add NO_ASN1_OLD which can be set to remove
14069 some old style ASN1 functions: this can be used to determine if old
14070 code will still work when these eventually go away.
14074 * New extension functions for OCSP structures, these follow the
14075 same conventions as certificates and CRLs.
14079 * New function X509V3_add1_i2d(). This automatically encodes and
14080 adds an extension. Its behaviour can be customised with various
14081 flags to append, replace or delete. Various wrappers added for
14082 certificates and CRLs.
14086 * Fix to avoid calling the underlying ASN1 print routine when
14087 an extension cannot be parsed. Correct a typo in the
14088 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14092 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14093 entries for variables.
14097 * Add functionality to `apps/openssl.c` for detecting locking
14098 problems: As the program is single-threaded, all we have
14099 to do is register a locking callback using an array for
14100 storing which locks are currently held by the program.
14104 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14105 SSL_get_ex_data_X509_STORE_idx(), which is used in
14106 ssl_verify_cert_chain() and thus can be called at any time
14107 during TLS/SSL handshakes so that thread-safety is essential.
14108 Unfortunately, the ex_data design is not at all suited
14109 for multi-threaded use, so it probably should be abolished.
14113 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14115 *Broadcom, tweaked and integrated by Geoff Thorpe*
14117 * Move common extension printing code to new function
14118 X509V3_print_extensions(). Reorganise OCSP print routines and
14119 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14123 * New function X509_signature_print() to remove duplication in some
14128 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14129 set (this was treated exactly the same as SET OF previously). This
14130 is used to reorder the STACK representing the structure to match the
14131 encoding. This will be used to get round a problem where a PKCS7
14132 structure which was signed could not be verified because the STACK
14133 order did not reflect the encoded order.
14137 * Reimplement the OCSP ASN1 module using the new code.
14141 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14142 for its ASN1 operations. The old style function pointers still exist
14143 for now but they will eventually go away.
14147 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14148 completely replaces the old ASN1 functionality with a table driven
14149 encoder and decoder which interprets an ASN1_ITEM structure describing
14150 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14151 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14152 has also been converted to the new form.
14156 * Change BN_mod_exp_recp so that negative moduli are tolerated
14157 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14158 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14159 for negative moduli.
14163 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14164 of not touching the result's sign bit.
14168 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14173 * Changed the LHASH code to use prototypes for callbacks, and created
14174 macros to declare and implement thin (optionally static) functions
14175 that provide type-safety and avoid function pointer casting for the
14176 type-specific callbacks.
14180 * Added Kerberos Cipher Suites to be used with TLS, as written in
14182 *Veers Staats <staatsvr@asc.hpc.mil>,
14183 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14185 * Reformat the FAQ so the different questions and answers can be divided
14186 in sections depending on the subject.
14190 * Have the zlib compression code load ZLIB.DLL dynamically under
14195 * New function BN_mod_sqrt for computing square roots modulo a prime
14196 (using the probabilistic Tonelli-Shanks algorithm unless
14197 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14198 be handled deterministically).
14200 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14202 * Make BN_mod_inverse faster by explicitly handling small quotients
14203 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14204 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14208 * New function BN_kronecker.
14212 * Fix BN_gcd so that it works on negative inputs; the result is
14213 positive unless both parameters are zero.
14214 Previously something reasonably close to an infinite loop was
14215 possible because numbers could be growing instead of shrinking
14216 in the implementation of Euclid's algorithm.
14220 * Fix BN_is_word() and BN_is_one() macros to take into account the
14221 sign of the number in question.
14223 Fix BN_is_word(a,w) to work correctly for w == 0.
14225 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14226 because its test if the absolute value of 'a' equals 'w'.
14227 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14228 it exists mostly for use in the implementations of BN_is_zero(),
14229 BN_is_one(), and BN_is_word().
14233 * New function BN_swap.
14237 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14238 the exponentiation functions are more likely to produce reasonable
14239 results on negative inputs.
14243 * Change BN_mod_mul so that the result is always non-negative.
14244 Previously, it could be negative if one of the factors was negative;
14245 I don't think anyone really wanted that behaviour.
14249 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14250 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14251 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14252 and add new functions:
14261 BN_mod_lshift1_quick
14263 BN_mod_lshift_quick
14265 These functions always generate non-negative results.
14267 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14268 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14270 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14271 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14272 be reduced modulo `m`.
14274 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14277 The following entry accidentally appeared in the CHANGES file
14278 distributed with OpenSSL 0.9.7. The modifications described in
14279 it do *not* apply to OpenSSL 0.9.7.
14281 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14282 was actually never needed) and in BN_mul(). The removal in BN_mul()
14283 required a small change in bn_mul_part_recursive() and the addition
14284 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14285 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14286 bn_sub_words() and bn_add_words() except they take arrays with
14292 * In 'openssl passwd', verify passwords read from the terminal
14293 unless the '-salt' option is used (which usually means that
14294 verification would just waste user's time since the resulting
14295 hash is going to be compared with some given password hash)
14296 or the new '-noverify' option is used.
14298 This is an incompatible change, but it does not affect
14299 non-interactive use of 'openssl passwd' (passwords on the command
14300 line, '-stdin' option, '-in ...' option) and thus should not
14301 cause any problems.
14305 * Remove all references to RSAref, since there's no more need for it.
14309 * Make DSO load along a path given through an environment variable
14310 (SHLIB_PATH) with shl_load().
14314 * Constify the ENGINE code as a result of BIGNUM constification.
14315 Also constify the RSA code and most things related to it. In a
14316 few places, most notable in the depth of the ASN.1 code, ugly
14317 casts back to non-const were required (to be solved at a later
14322 * Make it so the openssl application has all engines loaded by default.
14326 * Constify the BIGNUM routines a little more.
14330 * Add the following functions:
14332 ENGINE_load_cswift()
14334 ENGINE_load_atalla()
14335 ENGINE_load_nuron()
14336 ENGINE_load_builtin_engines()
14338 That way, an application can itself choose if external engines that
14339 are built-in in OpenSSL shall ever be used or not. The benefit is
14340 that applications won't have to be linked with libdl or other dso
14341 libraries unless it's really needed.
14343 Changed 'openssl engine' to load all engines on demand.
14344 Changed the engine header files to avoid the duplication of some
14345 declarations (they differed!).
14349 * 'openssl engine' can now list capabilities.
14353 * Better error reporting in 'openssl engine'.
14357 * Never call load_dh_param(NULL) in s_server.
14361 * Add engine application. It can currently list engines by name and
14362 identity, and test if they are actually available.
14366 * Improve RPM specification file by forcing symbolic linking and making
14367 sure the installed documentation is also owned by root.root.
14369 *Damien Miller <djm@mindrot.org>*
14371 * Give the OpenSSL applications more possibilities to make use of
14372 keys (public as well as private) handled by engines.
14376 * Add OCSP code that comes from CertCo.
14380 * Add VMS support for the Rijndael code.
14384 * Added untested support for Nuron crypto accelerator.
14388 * Add support for external cryptographic devices. This code was
14389 previously distributed separately as the "engine" branch.
14391 *Geoff Thorpe, Richard Levitte*
14393 * Rework the filename-translation in the DSO code. It is now possible to
14394 have far greater control over how a "name" is turned into a filename
14395 depending on the operating environment and any oddities about the
14396 different shared library filenames on each system.
14400 * Support threads on FreeBSD-elf in Configure.
14404 * Fix for SHA1 assembly problem with MASM: it produces
14405 warnings about corrupt line number information when assembling
14406 with debugging information. This is caused by the overlapping
14409 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14412 NCONF_get_number() has no error checking at all. As a replacement,
14413 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14414 promoted strongly. The old NCONF_get_number is kept around for
14415 binary backward compatibility.
14416 Make it possible for methods to load from something other than a BIO,
14417 by providing a function pointer that is given a name instead of a BIO.
14418 For example, this could be used to load configuration data from an
14423 * Fix for non blocking accept BIOs. Added new I/O special reason
14424 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14425 with non blocking I/O was not possible because no retry code was
14426 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14431 * Added the beginnings of Rijndael support.
14435 * Fix for bug in DirectoryString mask setting. Add support for
14436 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14437 to allow certificate printing to more controllable, additional
14438 'certopt' option to 'x509' to allow new printing options to be
14443 * Clean old EAY MD5 hack from e_os.h.
14447 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14449 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14450 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14452 *Joe Orton, Steve Henson*
14454 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14456 * Fix additional bug revealed by the NISCC test suite:
14458 Stop bug triggering large recursion when presented with
14459 certain ASN.1 tags ([CVE-2003-0851])
14463 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14465 * Fix various bugs revealed by running the NISCC test suite:
14467 Stop out of bounds reads in the ASN1 code when presented with
14468 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14470 If verify callback ignores invalid public key errors don't try to check
14471 certificate signature with the NULL public key.
14475 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14476 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14481 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14482 extra data after the compression methods not only for TLS 1.0
14483 but also for SSL 3.0 (as required by the specification).
14485 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14487 * Change X509_certificate_type() to mark the key as exported/exportable
14488 when it's 512 *bits* long, not 512 bytes.
14492 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14494 * Countermeasure against the Klima-Pokorny-Rosa extension of
14495 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14496 a protocol version number mismatch like a decryption error
14497 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14501 * Turn on RSA blinding by default in the default implementation
14502 to avoid a timing attack. Applications that don't want it can call
14503 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14504 They would be ill-advised to do so in most cases.
14506 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14508 * Change RSA blinding code so that it works when the PRNG is not
14509 seeded (in this case, the secret RSA exponent is abused as
14510 an unpredictable seed -- if it is not unpredictable, there
14511 is no point in blinding anyway). Make RSA blinding thread-safe
14512 by remembering the creator's thread ID in rsa->blinding and
14513 having all other threads use local one-time blinding factors
14514 (this requires more computation than sharing rsa->blinding, but
14515 avoids excessive locking; and if an RSA object is not shared
14516 between threads, blinding will still be very fast).
14520 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14522 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14523 via timing by performing a MAC computation even if incorrect
14524 block cipher padding has been found. This is a countermeasure
14525 against active attacks where the attacker has to distinguish
14526 between bad padding and a MAC verification error. ([CVE-2003-0078])
14528 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14529 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14530 Martin Vuagnoux (EPFL, Ilion)*
14532 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14534 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14535 memory from its contents. This is done with a counter that will
14536 place alternating values in each byte. This can be used to solve
14537 two issues: 1) the removal of calls to memset() by highly optimizing
14538 compilers, and 2) cleansing with other values than 0, since those can
14539 be read through on certain media, for example a swap space on disk.
14543 * Bugfix: client side session caching did not work with external caching,
14544 because the session->cipher setting was not restored when reloading
14545 from the external cache. This problem was masked, when
14546 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14547 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14551 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14552 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14554 *Zeev Lieber <zeev-l@yahoo.com>*
14556 * Undo an undocumented change introduced in 0.9.6e which caused
14557 repeated calls to OpenSSL_add_all_ciphers() and
14558 OpenSSL_add_all_digests() to be ignored, even after calling
14563 * Change the default configuration reader to deal with last line not
14564 being properly terminated.
14568 * Change X509_NAME_cmp() so it applies the special rules on handling
14569 DN values that are of type PrintableString, as well as RDNs of type
14570 emailAddress where the value has the type ia5String.
14572 *stefank@valicert.com via Richard Levitte*
14574 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14575 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14576 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14577 the bitwise-OR of the two for use by the majority of applications
14578 wanting this behaviour, and update the docs. The documented
14579 behaviour and actual behaviour were inconsistent and had been
14580 changing anyway, so this is more a bug-fix than a behavioural
14583 *Geoff Thorpe, diagnosed by Nadav Har'El*
14585 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14586 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14590 * Fix initialization code race conditions in
14591 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14592 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14593 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14594 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14595 ssl2_get_cipher_by_char(),
14596 ssl3_get_cipher_by_char().
14598 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14600 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14601 the cached sessions are flushed, as the remove_cb() might use ex_data
14602 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14603 (see [openssl.org #212]).
14605 *Geoff Thorpe, Lutz Jaenicke*
14607 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14608 length, instead of the encoding length to d2i_ASN1_OBJECT.
14612 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14614 * [In 0.9.6g-engine release:]
14615 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14617 *Lynn Gazis <lgazis@rainbow.com>*
14619 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14621 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14622 and get fix the header length calculation.
14623 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14624 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14626 * Use proper error handling instead of 'assertions' in buffer
14627 overflow checks added in 0.9.6e. This prevents DoS (the
14628 assertions could call abort()).
14630 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14632 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14634 * Add various sanity checks to asn1_get_length() to reject
14635 the ASN1 length bytes if they exceed sizeof(long), will appear
14636 negative or the content length exceeds the length of the
14639 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14641 * Fix cipher selection routines: ciphers without encryption had no flags
14642 for the cipher strength set and where therefore not handled correctly
14643 by the selection routines (PR #130).
14647 * Fix EVP_dsa_sha macro.
14652 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14653 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14654 that was added in OpenSSL 0.9.6d.
14656 As the countermeasure turned out to be incompatible with some
14657 broken SSL implementations, the new option is part of SSL_OP_ALL.
14658 SSL_OP_ALL is usually employed when compatibility with weird SSL
14659 implementations is desired (e.g. '-bugs' option to 's_client' and
14660 's_server'), so the new option is automatically set in many
14665 * Changes in security patch:
14667 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14668 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14669 Air Force Materiel Command, USAF, under agreement number
14672 * Add various sanity checks to asn1_get_length() to reject
14673 the ASN1 length bytes if they exceed sizeof(long), will appear
14674 negative or the content length exceeds the length of the
14675 supplied buffer. ([CVE-2002-0659])
14677 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14679 * Assertions for various potential buffer overflows, not known to
14680 happen in practice.
14682 *Ben Laurie (CHATS)*
14684 * Various temporary buffers to hold ASCII versions of integers were
14685 too small for 64 bit platforms. ([CVE-2002-0655])
14686 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14688 * Remote buffer overflow in SSL3 protocol - an attacker could
14689 supply an oversized session ID to a client. ([CVE-2002-0656])
14691 *Ben Laurie (CHATS)*
14693 * Remote buffer overflow in SSL2 protocol - an attacker could
14694 supply an oversized client master key. ([CVE-2002-0656])
14696 *Ben Laurie (CHATS)*
14698 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14700 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14701 encoded as NULL) with id-dsa-with-sha1.
14703 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14705 * Check various `X509_...()` return values in `apps/req.c`.
14707 *Nils Larsch <nla@trustcenter.de>*
14709 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14710 an end-of-file condition would erroneously be flagged, when the CRLF
14711 was just at the end of a processed block. The bug was discovered when
14712 processing data through a buffering memory BIO handing the data to a
14713 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14714 <ptsekov@syntrex.com> and Nedelcho Stanev.
14718 * Implement a countermeasure against a vulnerability recently found
14719 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14720 before application data chunks to avoid the use of known IVs
14721 with data potentially chosen by the attacker.
14725 * Fix length checks in ssl3_get_client_hello().
14729 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14730 to prevent ssl3_read_internal() from incorrectly assuming that
14731 ssl3_read_bytes() found application data while handshake
14732 processing was enabled when in fact s->s3->in_read_app_data was
14733 merely automatically cleared during the initial handshake.
14735 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14737 * Fix object definitions for Private and Enterprise: they were not
14738 recognized in their shortname (=lowercase) representation. Extend
14739 obj_dat.pl to issue an error when using undefined keywords instead
14740 of silently ignoring the problem (Svenning Sorensen
14741 <sss@sss.dnsalias.net>).
14745 * Fix DH_generate_parameters() so that it works for 'non-standard'
14746 generators, i.e. generators other than 2 and 5. (Previously, the
14747 code did not properly initialise the 'add' and 'rem' values to
14748 BN_generate_prime().)
14750 In the new general case, we do not insist that 'generator' is
14751 actually a primitive root: This requirement is rather pointless;
14752 a generator of the order-q subgroup is just as good, if not
14757 * Map new X509 verification errors to alerts. Discovered and submitted by
14758 Tom Wu <tom@arcot.com>.
14762 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14763 returning non-zero before the data has been completely received
14764 when using non-blocking I/O.
14766 *Bodo Moeller; problem pointed out by John Hughes*
14768 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14770 *Ben Laurie, Lutz Jaenicke*
14772 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14773 Yoram Zahavi <YoramZ@gilian.com>).
14777 * Add information about CygWin 1.3 and on, and preserve proper
14778 configuration for the versions before that.
14780 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14782 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14783 check whether we deal with a copy of a session and do not delete from
14784 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14785 <izhar@checkpoint.com>.
14789 * Do not store session data into the internal session cache, if it
14790 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14791 flag is set). Proposed by Aslam <aslam@funk.com>.
14795 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14800 * [In 0.9.6d-engine release:]
14801 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14803 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14805 * Add the configuration target linux-s390x.
14807 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14809 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14810 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14811 variable as an indication that a ClientHello message has been
14812 received. As the flag value will be lost between multiple
14813 invocations of ssl3_accept when using non-blocking I/O, the
14814 function may not be aware that a handshake has actually taken
14815 place, thus preventing a new session from being added to the
14818 To avoid this problem, we now set s->new_session to 2 instead of
14819 using a local variable.
14821 *Lutz Jaenicke, Bodo Moeller*
14823 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14824 if the SSL_R_LENGTH_MISMATCH error is detected.
14826 *Geoff Thorpe, Bodo Moeller*
14828 * New 'shared_ldflag' column in Configure platform table.
14832 * Fix EVP_CIPHER_mode macro.
14834 *"Dan S. Camper" <dan@bti.net>*
14836 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14837 type, we must throw them away by setting rr->length to 0.
14839 *D P Chang <dpc@qualys.com>*
14841 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14843 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14844 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14845 worked incorrectly for those cases where range = `10..._2` and
14846 `3*range` is two bits longer than range.)
14850 * Only add signing time to PKCS7 structures if it is not already
14855 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14856 OBJ_ld_ce should be OBJ_id_ce.
14857 Also some ip-pda OIDs in crypto/objects/objects.txt were
14858 incorrect (cf. RFC 3039).
14860 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14862 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14863 returns early because it has nothing to do.
14865 *Andy Schneider <andy.schneider@bjss.co.uk>*
14867 * [In 0.9.6c-engine release:]
14868 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14870 *Andy Schneider <andy.schneider@bjss.co.uk>*
14872 * [In 0.9.6c-engine release:]
14873 Add support for Cryptographic Appliance's keyserver technology.
14874 (Use engine 'keyclient')
14876 *Cryptographic Appliances and Geoff Thorpe*
14878 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14879 is called via tools/c89.sh because arguments have to be
14880 rearranged (all '-L' options must appear before the first object
14883 *Richard Shapiro <rshapiro@abinitio.com>*
14885 * [In 0.9.6c-engine release:]
14886 Add support for Broadcom crypto accelerator cards, backported
14889 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14891 * [In 0.9.6c-engine release:]
14892 Add support for SureWare crypto accelerator cards from
14893 Baltimore Technologies. (Use engine 'sureware')
14895 *Baltimore Technologies and Mark Cox*
14897 * [In 0.9.6c-engine release:]
14898 Add support for crypto accelerator cards from Accelerated
14899 Encryption Processing, www.aep.ie. (Use engine 'aep')
14901 *AEP Inc. and Mark Cox*
14903 * Add a configuration entry for gcc on UnixWare.
14905 *Gary Benson <gbenson@redhat.com>*
14907 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14908 messages are stored in a single piece (fixed-length part and
14909 variable-length part combined) and fix various bugs found on the way.
14913 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14914 instead. BIO_gethostbyname() does not know what timeouts are
14915 appropriate, so entries would stay in cache even when they have
14917 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14919 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14920 faced with a pathologically small ClientHello fragment that does
14921 not contain client_version: Instead of aborting with an error,
14922 simply choose the highest available protocol version (i.e.,
14923 TLS 1.0 unless it is disabled). In practice, ClientHello
14924 messages are never sent like this, but this change gives us
14925 strictly correct behaviour at least for TLS.
14929 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14930 never resets s->method to s->ctx->method when called from within
14931 one of the SSL handshake functions.
14933 *Bodo Moeller; problem pointed out by Niko Baric*
14935 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14936 (sent using the client's version number) if client_version is
14937 smaller than the protocol version in use. Also change
14938 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14939 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14940 the client will at least see that alert.
14944 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14949 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14950 client receives HelloRequest while in a handshake.
14952 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14954 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14955 should end in 'break', not 'goto end' which circumvents various
14956 cleanups done in state SSL_ST_OK. But session related stuff
14957 must be disabled for SSL_ST_OK in the case that we just sent a
14960 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14961 before just sending a HelloRequest.
14963 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14965 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14966 reveal whether illegal block cipher padding was found or a MAC
14967 verification error occurred. (Neither SSLerr() codes nor alerts
14968 are directly visible to potential attackers, but the information
14969 may leak via logfiles.)
14971 Similar changes are not required for the SSL 2.0 implementation
14972 because the number of padding bytes is sent in clear for SSL 2.0,
14973 and the extra bytes are just ignored. However ssl/s2_pkt.c
14974 failed to verify that the purported number of padding bytes is in
14979 * Add OpenUNIX-8 support including shared libraries
14980 (Boyd Lynn Gerber <gerberb@zenez.com>).
14984 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14985 'wristwatch attack' using huge encoding parameters (cf.
14986 James H. Manger's CRYPTO 2001 paper). Note that the
14987 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14988 encoding parameters and hence was not vulnerable.
14992 * BN_sqr() bug fix.
14994 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14996 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14997 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14998 followed by modular reduction.
15000 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15002 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15003 equivalent based on BN_pseudo_rand() instead of BN_rand().
15007 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15008 This function was broken, as the check for a new client hello message
15009 to handle SGC did not allow these large messages.
15010 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15014 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15018 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15019 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15023 * Rework the configuration and shared library support for Tru64 Unix.
15024 The configuration part makes use of modern compiler features and
15025 still retains old compiler behavior for those that run older versions
15026 of the OS. The shared library support part includes a variant that
15027 uses the RPATH feature, and is available through the special
15028 configuration target "alpha-cc-rpath", which will never be selected
15031 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15033 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15034 with the same message size as in ssl3_get_certificate_request().
15035 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15036 messages might inadvertently be reject as too long.
15038 *Petr Lampa <lampa@fee.vutbr.cz>*
15040 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15044 * Modified SSL library such that the verify_callback that has been set
15045 specifically for an SSL object with SSL_set_verify() is actually being
15046 used. Before the change, a verify_callback set with this function was
15047 ignored and the verify_callback() set in the SSL_CTX at the time of
15048 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15049 to allow the necessary settings.
15053 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15054 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15055 done automatically (in contradiction to the requirements of the C
15056 standard). This made problems when used from OpenSSH.
15060 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15061 dh->length and always used
15063 BN_rand_range(priv_key, dh->p).
15065 BN_rand_range() is not necessary for Diffie-Hellman, and this
15066 specific range makes Diffie-Hellman unnecessarily inefficient if
15067 dh->length (recommended exponent length) is much smaller than the
15068 length of dh->p. We could use BN_rand_range() if the order of
15069 the subgroup was stored in the DH structure, but we only have
15074 BN_rand(priv_key, l, ...)
15076 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15083 RSA_eay_public_encrypt
15084 RSA_eay_private_decrypt
15085 RSA_eay_private_encrypt (signing)
15086 RSA_eay_public_decrypt (signature verification)
15088 (default implementations for RSA_public_encrypt,
15089 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15090 always reject numbers >= n.
15094 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15095 to synchronize access to 'locking_thread'. This is necessary on
15096 systems where access to 'locking_thread' (an 'unsigned long'
15097 variable) is not atomic.
15101 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15102 *before* setting the 'crypto_lock_rand' flag. The previous code had
15103 a race condition if 0 is a valid thread ID.
15105 *Travis Vitek <vitek@roguewave.com>*
15107 * Add support for shared libraries under Irix.
15109 *Albert Chin-A-Young <china@thewrittenword.com>*
15111 * Add configuration option to build on Linux on both big-endian and
15112 little-endian MIPS.
15114 *Ralf Baechle <ralf@uni-koblenz.de>*
15116 * Add the possibility to create shared libraries on HP-UX.
15120 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15122 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15123 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15124 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15125 PRNG state recovery was possible based on the output of
15126 one PRNG request appropriately sized to gain knowledge on
15127 'md' followed by enough consecutive 1-byte PRNG requests
15128 to traverse all of 'state'.
15130 1. When updating 'md_local' (the current thread's copy of 'md')
15131 during PRNG output generation, hash all of the previous
15132 'md_local' value, not just the half used for PRNG output.
15134 2. Make the number of bytes from 'state' included into the hash
15135 independent from the number of PRNG bytes requested.
15137 The first measure alone would be sufficient to avoid
15138 Markku-Juhani's attack. (Actually it had never occurred
15139 to me that the half of 'md_local' used for chaining was the
15140 half from which PRNG output bytes were taken -- I had always
15141 assumed that the secret half would be used.) The second
15142 measure makes sure that additional data from 'state' is never
15143 mixed into 'md_local' in small portions; this heuristically
15144 further strengthens the PRNG.
15148 * Fix crypto/bn/asm/mips3.s.
15152 * When only the key is given to "enc", the IV is undefined. Print out
15153 an error message in this case.
15157 * Handle special case when X509_NAME is empty in X509 printing routines.
15161 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15162 positive and less than q.
15166 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15167 used: it isn't thread safe and the add_lock_callback should handle
15170 *Paul Rose <Paul.Rose@bridge.com>*
15172 * Verify that incoming data obeys the block size in
15173 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15179 *Ulf Möller, Bodo Möller*
15181 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15182 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15183 when fixing the server behaviour for backwards-compatible 'client
15184 hello' messages. (Note that the attack is impractical against
15185 SSL 3.0 and TLS 1.0 anyway because length and version checking
15186 means that the probability of guessing a valid ciphertext is
15187 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15190 Before 0.9.5, the countermeasure (hide the error by generating a
15191 random 'decryption result') did not work properly because
15192 ERR_clear_error() was missing, meaning that SSL_get_error() would
15193 detect the supposedly ignored error.
15195 Both problems are now fixed.
15199 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15200 (previously it was 1024).
15204 * Fix for compatibility mode trust settings: ignore trust settings
15205 unless some valid trust or reject settings are present.
15209 * Fix for blowfish EVP: its a variable length cipher.
15213 * Fix various bugs related to DSA S/MIME verification. Handle missing
15214 parameters in DSA public key structures and return an error in the
15215 DSA routines if parameters are absent.
15219 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15220 in the current directory if neither $RANDFILE nor $HOME was set.
15221 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15222 caused some confusion to Windows users who haven't defined $HOME.
15223 Thus RAND_file_name() is changed again: e_os.h can define a
15224 DEFAULT_HOME, which will be used if $HOME is not set.
15225 For Windows, we use "C:"; on other platforms, we still require
15226 environment variables.
15228 * Move 'if (!initialized) RAND_poll()' into regions protected by
15229 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15230 having multiple threads call RAND_poll() concurrently.
15234 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15235 combination of a flag and a thread ID variable.
15236 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15237 flag), *other* threads can enter ssleay_add_bytes without obeying
15238 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15239 that they do not hold after the first thread unsets add_do_not_lock).
15243 * Change bctest again: '-x' expressions are not available in all
15244 versions of 'test'.
15248 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15250 * Fix a couple of memory leaks in PKCS7_dataDecode()
15252 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15254 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15255 the default extension for executables, if any. Also, make the perl
15256 scripts that use symlink() to test if it really exists and use "cp"
15257 if it doesn't. All this made OpenSSL compilable and installable in
15262 * Fix for asn1_GetSequence() for indefinite length constructed data.
15263 If SEQUENCE is length is indefinite just set c->slen to the total
15264 amount of data available.
15266 *Steve Henson, reported by shige@FreeBSD.org*
15268 *This change does not apply to 0.9.7.*
15270 * Change bctest to avoid here-documents inside command substitution
15271 (workaround for FreeBSD /bin/sh bug).
15272 For compatibility with Ultrix, avoid shell functions (introduced
15273 in the bctest version that searches along $PATH).
15277 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15278 with des_encrypt() defined on some operating systems, like Solaris
15283 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15284 On the Importance of Eliminating Errors in Cryptographic
15285 Computations, J. Cryptology 14 (2001) 2, 101-119,
15286 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15290 * MIPS assembler BIGNUM division bug fix.
15294 * Disabled incorrect Alpha assembler code.
15298 * Fix PKCS#7 decode routines so they correctly update the length
15299 after reading an EOC for the EXPLICIT tag.
15303 *This change does not apply to 0.9.7.*
15305 * Fix bug in PKCS#12 key generation routines. This was triggered
15306 if a 3DES key was generated with a 0 initial byte. Include
15307 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15308 (but broken) behaviour.
15312 * Enhance bctest to search for a working bc along $PATH and print
15315 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15317 * Fix memory leaks in err.c: free err_data string if necessary;
15318 don't write to the wrong index in ERR_set_error_data.
15322 * Implement ssl23_peek (analogous to ssl23_read), which previously
15327 * Replace rdtsc with `_emit` statements for VC++ version 5.
15329 *Jeremy Cooper <jeremy@baymoo.org>*
15331 * Make it possible to reuse SSLv2 sessions.
15335 * In copy_email() check for >= 0 as a return value for
15336 X509_NAME_get_index_by_NID() since 0 is a valid index.
15338 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15340 * Avoid coredump with unsupported or invalid public keys by checking if
15341 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15342 PKCS7_verify() fails with non detached data.
15346 * Don't use getenv in library functions when run as setuid/setgid.
15347 New function OPENSSL_issetugid().
15351 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15352 due to incorrect handling of multi-threading:
15354 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15356 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15358 3. Count how many times MemCheck_off() has been called so that
15359 nested use can be treated correctly. This also avoids
15360 inband-signalling in the previous code (which relied on the
15361 assumption that thread ID 0 is impossible).
15365 * Add "-rand" option also to s_client and s_server.
15369 * Fix CPU detection on Irix 6.x.
15370 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15371 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15373 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15378 *This change does not apply to 0.9.7.*
15380 * Use the cached encoding of an X509_NAME structure rather than
15381 copying it. This is apparently the reason for the libsafe "errors"
15382 but the code is actually correct.
15386 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15387 Bleichenbacher's DSA attack.
15388 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15389 to be set and top=0 forces the highest bit to be set; top=-1 is new
15390 and leaves the highest bit random.
15392 *Ulf Moeller, Bodo Moeller*
15394 * In the `NCONF_...`-based implementations for `CONF_...` queries
15395 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15396 a temporary CONF structure with the data component set to NULL
15397 (which gives segmentation faults in lh_retrieve).
15398 Instead, use NULL for the CONF pointer in CONF_get_string and
15399 CONF_get_number (which may use environment variables) and directly
15400 return NULL from CONF_get_section.
15404 * Fix potential buffer overrun for EBCDIC.
15408 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15409 keyUsage if basicConstraints absent for a CA.
15413 * Make SMIME_write_PKCS7() write mail header values with a format that
15414 is more generally accepted (no spaces before the semicolon), since
15415 some programs can't parse those values properly otherwise. Also make
15416 sure BIO's that break lines after each write do not create invalid
15421 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15422 macros previously used would not encode an empty SEQUENCE OF
15423 and break the signature.
15427 *This change does not apply to 0.9.7.*
15429 * Zero the premaster secret after deriving the master secret in
15434 * Add some EVP_add_digest_alias registrations (as found in
15435 OpenSSL_add_all_digests()) to SSL_library_init()
15436 aka OpenSSL_add_ssl_algorithms(). This provides improved
15437 compatibility with peers using X.509 certificates
15438 with unconventional AlgorithmIdentifier OIDs.
15442 * Fix for Irix with NO_ASM.
15444 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15446 * ./config script fixes.
15448 *Ulf Moeller, Richard Levitte*
15450 * Fix 'openssl passwd -1'.
15454 * Change PKCS12_key_gen_asc() so it can cope with non null
15455 terminated strings whose length is passed in the passlen
15456 parameter, for example from PEM callbacks. This was done
15457 by adding an extra length parameter to asc2uni().
15459 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15461 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15462 call failed, free the DSA structure.
15466 * Fix to uni2asc() to cope with zero length Unicode strings.
15467 These are present in some PKCS#12 files.
15471 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15472 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15473 when writing a 32767 byte record.
15475 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15477 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15478 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15480 (RSA objects have a reference count access to which is protected
15481 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15482 so they are meant to be shared between threads.)
15483 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15484 "Reddie, Steven" <Steven.Reddie@ca.com>*
15486 * Fix a deadlock in CRYPTO_mem_leaks().
15490 * Use better test patterns in bntest.
15494 * rand_win.c fix for Borland C.
15498 * BN_rshift bugfix for n == 0.
15502 * Add a 'bctest' script that checks for some known 'bc' bugs
15503 so that 'make test' does not abort just because 'bc' is broken.
15507 * Store verify_result within SSL_SESSION also for client side to
15508 avoid potential security hole. (Re-used sessions on the client side
15509 always resulted in verify_result==X509_V_OK, not using the original
15510 result of the server certificate verification.)
15514 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15515 SSL3_RT_APPLICATION_DATA, return 0.
15516 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15521 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15522 releases, have been re-implemented by renaming the previous
15523 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15524 and ssl3_read_internal, respectively, and adding 'peek' parameters
15525 to them. The new ssl[23]_{read,peek} functions are calls to
15526 ssl[23]_read_internal with the 'peek' flag set appropriately.
15527 A 'peek' parameter has also been added to ssl3_read_bytes, which
15528 does the actual work for ssl3_read_internal.
15532 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15533 the method-specific "init()" handler. Also clean up ex_data after
15534 calling the method-specific "finish()" handler. Previously, this was
15535 happening the other way round.
15539 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15540 The previous value, 12, was not always sufficient for BN_mod_exp().
15544 * Make sure that shared libraries get the internal name engine with
15545 the full version number and not just 0. This should mark the
15546 shared libraries as not backward compatible. Of course, this should
15547 be changed again when we can guarantee backward binary compatibility.
15551 * Fix typo in get_cert_by_subject() in by_dir.c
15553 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15555 * Rework the system to generate shared libraries:
15557 - Make note of the expected extension for the shared libraries and
15558 if there is a need for symbolic links from for example libcrypto.so.0
15559 to libcrypto.so.0.9.7. There is extended info in Configure for
15562 - Make as few rebuilds of the shared libraries as possible.
15564 - Still avoid linking the OpenSSL programs with the shared libraries.
15566 - When installing, install the shared libraries separately from the
15571 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15573 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15574 and not in SSL_clear because the latter is also used by the
15575 accept/connect functions; previously, the settings made by
15576 SSL_set_read_ahead would be lost during the handshake.
15578 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15580 * Correct util/mkdef.pl to be selective about disabled algorithms.
15581 Previously, it would create entries for disabled algorithms no
15586 * Added several new manual pages for SSL_* function.
15590 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15592 * In ssl23_get_client_hello, generate an error message when faced
15593 with an initial SSL 3.0/TLS record that is too small to contain the
15594 first two bytes of the ClientHello message, i.e. client_version.
15595 (Note that this is a pathologic case that probably has never happened
15596 in real life.) The previous approach was to use the version number
15597 from the record header as a substitute; but our protocol choice
15598 should not depend on that one because it is not authenticated
15599 by the Finished messages.
15603 * More robust randomness gathering functions for Windows.
15605 *Jeffrey Altman <jaltman@columbia.edu>*
15607 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15608 not set then we don't setup the error code for issuer check errors
15609 to avoid possibly overwriting other errors which the callback does
15610 handle. If an application does set the flag then we assume it knows
15611 what it is doing and can handle the new informational codes
15616 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15617 a general "ANY" type, as such it should be able to decode anything
15618 including tagged types. However it didn't check the class so it would
15619 wrongly interpret tagged types in the same way as their universal
15620 counterpart and unknown types were just rejected. Changed so that the
15621 tagged and unknown types are handled in the same way as a SEQUENCE:
15622 that is the encoding is stored intact. There is also a new type
15623 "V_ASN1_OTHER" which is used when the class is not universal, in this
15624 case we have no idea what the actual type is so we just lump them all
15629 * On VMS, stdout may very well lead to a file that is written to
15630 in a record-oriented fashion. That means that every write() will
15631 write a separate record, which will be read separately by the
15632 programs trying to read from it. This can be very confusing.
15634 The solution is to put a BIO filter in the way that will buffer
15635 text until a linefeed is reached, and then write everything a
15636 line at a time, so every record written will be an actual line,
15637 not chunks of lines and not (usually doesn't happen, but I've
15638 seen it once) several lines in one record. BIO_f_linebuffer() is
15641 Currently, it's a VMS-only method, because that's where it has
15642 been tested well enough.
15646 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15647 it can return incorrect results.
15648 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15649 but it was in 0.9.6-beta[12].)
15653 * Disable the check for content being present when verifying detached
15654 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15655 include zero length content when signing messages.
15659 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15660 BIO_ctrl (for BIO pairs).
15664 * Add DSO method for VMS.
15668 * Bug fix: Montgomery multiplication could produce results with the
15673 * Add RPM specification openssl.spec and modify it to build three
15674 packages. The default package contains applications, application
15675 documentation and run-time libraries. The devel package contains
15676 include files, static libraries and function documentation. The
15677 doc package contains the contents of the doc directory. The original
15678 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15682 * Add a large number of documentation files for many SSL routines.
15684 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15686 * Add a configuration entry for Sony News 4.
15688 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15690 * Don't set the two most significant bits to one when generating a
15691 random number < q in the DSA library.
15695 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15696 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15697 the underlying transport is blocking) if a handshake took place.
15698 (The default behaviour is needed by applications such as s_client
15699 and s_server that use select() to determine when to use SSL_read;
15700 but for applications that know in advance when to expect data, it
15701 just makes things more complicated.)
15705 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15710 * Add a few more EBCDIC conditionals that make `req` and `x509`
15711 work better on such systems.
15713 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15715 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15716 Update PKCS12_parse() so it copies the friendlyName and the
15717 keyid to the certificates aux info.
15721 * Fix bug in PKCS7_verify() which caused an infinite loop
15722 if there was more than one signature.
15724 *Sven Uszpelkat <su@celocom.de>*
15726 * Major change in util/mkdef.pl to include extra information
15727 about each symbol, as well as presenting variables as well
15728 as functions. This change means that there's n more need
15729 to rebuild the .num files when some algorithms are excluded.
15733 * Allow the verify time to be set by an application,
15734 rather than always using the current time.
15738 * Phase 2 verify code reorganisation. The certificate
15739 verify code now looks up an issuer certificate by a
15740 number of criteria: subject name, authority key id
15741 and key usage. It also verifies self signed certificates
15742 by the same criteria. The main comparison function is
15743 X509_check_issued() which performs these checks.
15745 Lot of changes were necessary in order to support this
15746 without completely rewriting the lookup code.
15748 Authority and subject key identifier are now cached.
15750 The LHASH 'certs' is X509_STORE has now been replaced
15751 by a STACK_OF(X509_OBJECT). This is mainly because an
15752 LHASH can't store or retrieve multiple objects with
15753 the same hash value.
15755 As a result various functions (which were all internal
15756 use only) have changed to handle the new X509_STORE
15757 structure. This will break anything that messed round
15758 with X509_STORE internally.
15760 The functions X509_STORE_add_cert() now checks for an
15761 exact match, rather than just subject name.
15763 The X509_STORE API doesn't directly support the retrieval
15764 of multiple certificates matching a given criteria, however
15765 this can be worked round by performing a lookup first
15766 (which will fill the cache with candidate certificates)
15767 and then examining the cache for matches. This is probably
15768 the best we can do without throwing out X509_LOOKUP
15769 entirely (maybe later...).
15771 The X509_VERIFY_CTX structure has been enhanced considerably.
15773 All certificate lookup operations now go via a get_issuer()
15774 callback. Although this currently uses an X509_STORE it
15775 can be replaced by custom lookups. This is a simple way
15776 to bypass the X509_STORE hackery necessary to make this
15777 work and makes it possible to use more efficient techniques
15778 in future. A very simple version which uses a simple
15779 STACK for its trusted certificate store is also provided
15780 using X509_STORE_CTX_trusted_stack().
15782 The verify_cb() and verify() callbacks now have equivalents
15783 in the X509_STORE_CTX structure.
15785 X509_STORE_CTX also has a 'flags' field which can be used
15786 to customise the verify behaviour.
15790 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15791 excludes S/MIME capabilities.
15795 * When a certificate request is read in keep a copy of the
15796 original encoding of the signed data and use it when outputting
15797 again. Signatures then use the original encoding rather than
15798 a decoded, encoded version which may cause problems if the
15799 request is improperly encoded.
15803 * For consistency with other BIO_puts implementations, call
15804 buffer_write(b, ...) directly in buffer_puts instead of calling
15807 In BIO_puts, increment b->num_write as in BIO_write.
15809 *Peter.Sylvester@EdelWeb.fr*
15811 * Fix BN_mul_word for the case where the word is 0. (We have to use
15812 BN_zero, we may not return a BIGNUM with an array consisting of
15813 words set to zero.)
15817 * Avoid calling abort() from within the library when problems are
15818 detected, except if preprocessor symbols have been defined
15819 (such as REF_CHECK, BN_DEBUG etc.).
15823 * New openssl application 'rsautl'. This utility can be
15824 used for low-level RSA operations. DER public key
15825 BIO/fp routines also added.
15829 * New Configure entry and patches for compiling on QNX 4.
15831 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15833 * A demo state-machine implementation was sponsored by
15834 Nuron (<http://www.nuron.com/>) and is now available in
15835 demos/state_machine.
15839 * New options added to the 'dgst' utility for signature
15840 generation and verification.
15844 * Unrecognized PKCS#7 content types are now handled via a
15845 catch all ASN1_TYPE structure. This allows unsupported
15846 types to be stored as a "blob" and an application can
15847 encode and decode it manually.
15851 * Fix various signed/unsigned issues to make a_strex.c
15852 compile under VC++.
15854 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15856 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15857 length if passed a buffer. ASN1_INTEGER_to_BN failed
15858 if passed a NULL BN and its argument was negative.
15860 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15862 * Modification to PKCS#7 encoding routines to output definite
15863 length encoding. Since currently the whole structures are in
15864 memory there's not real point in using indefinite length
15865 constructed encoding. However if OpenSSL is compiled with
15866 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15870 * Added BIO_vprintf() and BIO_vsnprintf().
15874 * Added more prefixes to parse for in the strings written
15875 through a logging bio, to cover all the levels that are available
15876 through syslog. The prefixes are now:
15878 PANIC, EMERG, EMR => LOG_EMERG
15879 ALERT, ALR => LOG_ALERT
15880 CRIT, CRI => LOG_CRIT
15881 ERROR, ERR => LOG_ERR
15882 WARNING, WARN, WAR => LOG_WARNING
15883 NOTICE, NOTE, NOT => LOG_NOTICE
15884 INFO, INF => LOG_INFO
15885 DEBUG, DBG => LOG_DEBUG
15887 and as before, if none of those prefixes are present at the
15888 beginning of the string, LOG_ERR is chosen.
15890 On Win32, the `LOG_*` levels are mapped according to this:
15892 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15893 LOG_WARNING => EVENTLOG_WARNING_TYPE
15894 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15898 * Made it possible to reconfigure with just the configuration
15899 argument "reconf" or "reconfigure". The command line arguments
15900 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15901 and are retrieved from there when reconfiguring.
15907 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15909 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15913 * The obj_dat.pl script was messing up the sorting of object
15914 names. The reason was that it compared the quoted version
15915 of strings as a result "OCSP" > "OCSP Signing" because
15916 " > SPACE. Changed script to store unquoted versions of
15917 names and add quotes on output. It was also omitting some
15918 names from the lookup table if they were given a default
15919 value (that is if SN is missing it is given the same
15920 value as LN and vice versa), these are now added on the
15921 grounds that if an object has a name we should be able to
15922 look it up. Finally added warning output when duplicate
15923 short or long names are found.
15927 * Changes needed for Tandem NSK.
15929 *Scott Uroff <scott@xypro.com>*
15931 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15932 RSA_padding_check_SSLv23(), special padding was never detected
15933 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15934 version rollback attacks was not effective.
15936 In s23_clnt.c, don't use special rollback-attack detection padding
15937 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15938 client; similarly, in s23_srvr.c, don't do the rollback check if
15939 SSL 2.0 is the only protocol enabled in the server.
15943 * Make it possible to get hexdumps of unprintable data with 'openssl
15944 asn1parse'. By implication, the functions ASN1_parse_dump() and
15945 BIO_dump_indent() are added.
15949 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15950 these print out strings and name structures based on various
15951 flags including RFC2253 support and proper handling of
15952 multibyte characters. Added options to the 'x509' utility
15953 to allow the various flags to be set.
15957 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15958 Also change the functions X509_cmp_current_time() and
15959 X509_gmtime_adj() work with an ASN1_TIME structure,
15960 this will enable certificates using GeneralizedTime in validity
15961 dates to be checked.
15965 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15966 negative public key encodings) on by default,
15967 NO_NEG_PUBKEY_BUG can be set to disable it.
15971 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15972 content octets. An i2c_ASN1_OBJECT is unnecessary because
15973 the encoding can be trivially obtained from the structure.
15977 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15978 not read locks (`CRYPTO_r_[un]lock`).
15982 * A first attempt at creating official support for shared
15983 libraries through configuration. I've kept it so the
15984 default is static libraries only, and the OpenSSL programs
15985 are always statically linked for now, but there are
15986 preparations for dynamic linking in place.
15987 This has been tested on Linux and Tru64.
15991 * Randomness polling function for Win9x, as described in:
15992 Peter Gutmann, Software Generation of Practically Strong
15997 * Fix so PRNG is seeded in req if using an already existing
16002 * New options to smime application. -inform and -outform
16003 allow alternative formats for the S/MIME message including
16004 PEM and DER. The -content option allows the content to be
16005 specified separately. This should allow things like Netscape
16006 form signing output easier to verify.
16010 * Fix the ASN1 encoding of tags using the 'long form'.
16014 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16015 STRING types. These convert content octets to and from the
16016 underlying type. The actual tag and length octets are
16017 already assumed to have been read in and checked. These
16018 are needed because all other string types have virtually
16019 identical handling apart from the tag. By having versions
16020 of the ASN1 functions that just operate on content octets
16021 IMPLICIT tagging can be handled properly. It also allows
16022 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16023 and ASN1_INTEGER are identical apart from the tag.
16027 * Change the handling of OID objects as follows:
16029 - New object identifiers are inserted in objects.txt, following
16030 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16031 - objects.pl is used to process obj_mac.num and create a new
16033 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16036 This is currently kind of a hack, and the perl code in objects.pl
16037 isn't very elegant, but it works as I intended. The simplest way
16038 to check that it worked correctly is to look in obj_dat.h and
16039 check the array nid_objs and make sure the objects haven't moved
16040 around (this is important!). Additions are OK, as well as
16041 consistent name changes.
16045 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16049 * Addition of the command line parameter '-rand file' to 'openssl req'.
16050 The given file adds to whatever has already been seeded into the
16051 random pool through the RANDFILE configuration file option or
16052 environment variable, or the default random state file.
16056 * mkstack.pl now sorts each macro group into lexical order.
16057 Previously the output order depended on the order the files
16058 appeared in the directory, resulting in needless rewriting
16063 * Patches to make OpenSSL compile under Win32 again. Mostly
16064 work arounds for the VC++ problem that it treats func() as
16065 func(void). Also stripped out the parts of mkdef.pl that
16066 added extra typesafe functions: these no longer exist.
16070 * Reorganisation of the stack code. The macros are now all
16071 collected in safestack.h . Each macro is defined in terms of
16072 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16073 DEBUG_SAFESTACK is now handled in terms of function casts,
16074 this has the advantage of retaining type safety without the
16075 use of additional functions. If DEBUG_SAFESTACK is not defined
16076 then the non typesafe macros are used instead. Also modified the
16077 mkstack.pl script to handle the new form. Needs testing to see
16078 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16079 the default if no major problems. Similar behaviour for ASN1_SET_OF
16080 and PKCS12_STACK_OF.
16084 * When some versions of IIS use the 'NET' form of private key the
16085 key derivation algorithm is different. Normally MD5(password) is
16086 used as a 128 bit RC4 key. In the modified case
16087 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16088 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16089 as the old Netscape_RSA functions except they have an additional
16090 'sgckey' parameter which uses the modified algorithm. Also added
16091 an -sgckey command line option to the rsa utility. Thanks to
16092 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16093 algorithm to openssl-dev.
16097 * The evp_local.h macros were using 'c.##kname' which resulted in
16098 invalid expansion on some systems (SCO 5.0.5 for example).
16099 Corrected to 'c.kname'.
16101 *Phillip Porch <root@theporch.com>*
16103 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16104 a STACK of email addresses from a certificate or request, these look
16105 in the subject name and the subject alternative name extensions and
16106 omit any duplicate addresses.
16110 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16111 This makes DSA verification about 2 % faster.
16115 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16116 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16117 plus overhead for 1024 bit moduli).
16118 This makes exponentiations about 0.5 % faster for 1024 bit
16119 exponents (as measured by "openssl speed rsa2048").
16123 * Rename memory handling macros to avoid conflicts with other
16125 Malloc => OPENSSL_malloc
16126 Malloc_locked => OPENSSL_malloc_locked
16127 Realloc => OPENSSL_realloc
16128 Free => OPENSSL_free
16132 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16133 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16137 * CygWin32 support.
16139 *John Jarvie <jjarvie@newsguy.com>*
16141 * The type-safe stack code has been rejigged. It is now only compiled
16142 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16143 by default all type-specific stack functions are "#define"d back to
16144 standard stack functions. This results in more streamlined output
16145 but retains the type-safety checking possibilities of the original
16150 * The STACK code has been cleaned up, and certain type declarations
16151 that didn't make a lot of sense have been brought in line. This has
16152 also involved a cleanup of sorts in safestack.h to more correctly
16153 map type-safe stack functions onto their plain stack counterparts.
16154 This work has also resulted in a variety of "const"ifications of
16155 lots of the code, especially `_cmp` operations which should normally
16156 be prototyped with "const" parameters anyway.
16160 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16161 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16162 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16163 where all of 'md' is used each time the PRNG is used, but 'state'
16164 is used only indexed by a cyclic counter. As entropy may not be
16165 well distributed from the beginning, 'md' is important as a
16166 chaining variable. However, the output function chains only half
16167 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16168 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16169 in all of 'state' being rewritten, with the new values depending
16170 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16174 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16175 the handshake is continued after ssl_verify_cert_chain();
16176 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16177 can lead to 'unexplainable' connection aborts later.
16179 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16181 * Major EVP API cipher revision.
16182 Add hooks for extra EVP features. This allows various cipher
16183 parameters to be set in the EVP interface. Support added for variable
16184 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16185 setting of RC2 and RC5 parameters.
16187 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16190 Remove lots of duplicated code from the EVP library. For example *every*
16191 cipher init() function handles the 'iv' in the same way according to the
16192 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16193 for CFB and OFB modes they zero ctx->num.
16195 New functionality allows removal of S/MIME code RC2 hack.
16197 Most of the routines have the same form and so can be declared in terms
16200 By shifting this to the top level EVP_CipherInit() it can be removed from
16201 all individual ciphers. If the cipher wants to handle IVs or keys
16202 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16205 Change lots of functions like EVP_EncryptUpdate() to now return a
16206 value: although software versions of the algorithms cannot fail
16207 any installed hardware versions can.
16211 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16212 this option is set, tolerate broken clients that send the negotiated
16213 protocol version number instead of the requested protocol version
16218 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16219 i.e. non-zero for export ciphersuites, zero otherwise.
16220 Previous versions had this flag inverted, inconsistent with
16221 rsa_tmp_cb (..._TMP_RSA_CB).
16223 *Bodo Moeller; problem reported by Amit Chopra*
16225 * Add missing DSA library text string. Work around for some IIS
16226 key files with invalid SEQUENCE encoding.
16230 * Add a document (doc/standards.txt) that list all kinds of standards
16231 and so on that are implemented in OpenSSL.
16235 * Enhance c_rehash script. Old version would mishandle certificates
16236 with the same subject name hash and wouldn't handle CRLs at all.
16237 Added -fingerprint option to crl utility, to support new c_rehash
16242 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16246 * Fix for SSL server purpose checking. Server checking was
16247 rejecting certificates which had extended key usage present
16248 but no ssl client purpose.
16250 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16252 * Make PKCS#12 code work with no password. The PKCS#12 spec
16253 is a little unclear about how a blank password is handled.
16254 Since the password in encoded as a BMPString with terminating
16255 double NULL a zero length password would end up as just the
16256 double NULL. However no password at all is different and is
16257 handled differently in the PKCS#12 key generation code. NS
16258 treats a blank password as zero length. MSIE treats it as no
16259 password on export: but it will try both on import. We now do
16260 the same: PKCS12_parse() tries zero length and no password if
16261 the password is set to "" or NULL (NULL is now a valid password:
16262 it wasn't before) as does the pkcs12 application.
16266 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16267 perror when PEM_read_bio_X509_REQ fails, the error message must
16268 be obtained from the error queue.
16272 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16273 it in ERR_remove_state if appropriate, and change ERR_get_state
16274 accordingly to avoid race conditions (this is necessary because
16275 thread_hash is no longer constant once set).
16279 * Bugfix for linux-elf makefile.one.
16283 * RSA_get_default_method() will now cause a default
16284 RSA_METHOD to be chosen if one doesn't exist already.
16285 Previously this was only set during a call to RSA_new()
16286 or RSA_new_method(NULL) meaning it was possible for
16287 RSA_get_default_method() to return NULL.
16291 * Added native name translation to the existing DSO code
16292 that will convert (if the flag to do so is set) filenames
16293 that are sufficiently small and have no path information
16294 into a canonical native form. Eg. "blah" converted to
16295 "libblah.so" or "blah.dll" etc.
16299 * New function ERR_error_string_n(e, buf, len) which is like
16300 ERR_error_string(e, buf), but writes at most 'len' bytes
16301 including the 0 terminator. For ERR_error_string_n, 'buf'
16304 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16306 * CONF library reworked to become more general. A new CONF
16307 configuration file reader "class" is implemented as well as a
16308 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16309 old `CONF_*` functions are still there, but are reimplemented to
16310 work in terms of the new functions. Also, a set of functions
16311 to handle the internal storage of the configuration data is
16312 provided to make it easier to write new configuration file
16313 reader "classes" (I can definitely see something reading a
16314 configuration file in XML format, for example), called `_CONF_*`,
16315 or "the configuration storage API"...
16317 The new configuration file reading functions are:
16319 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16320 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16322 NCONF_default, NCONF_WIN32
16324 NCONF_dump_fp, NCONF_dump_bio
16326 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16327 NCONF_new creates a new CONF object. This works in the same way
16328 as other interfaces in OpenSSL, like the BIO interface.
16329 `NCONF_dump_*` dump the internal storage of the configuration file,
16330 which is useful for debugging. All other functions take the same
16331 arguments as the old `CONF_*` functions with the exception of the
16332 first that must be a `CONF *` instead of a `LHASH *`.
16334 To make it easier to use the new classes with the old `CONF_*` functions,
16335 the function CONF_set_default_method is provided.
16339 * Add '-tls1' option to 'openssl ciphers', which was already
16340 mentioned in the documentation but had not been implemented.
16341 (This option is not yet really useful because even the additional
16342 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16346 * Initial DSO code added into libcrypto for letting OpenSSL (and
16347 OpenSSL-based applications) load shared libraries and bind to
16348 them in a portable way.
16350 *Geoff Thorpe, with contributions from Richard Levitte*
16352 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16354 * Make sure _lrotl and _lrotr are only used with MSVC.
16356 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16357 (the default implementation of RAND_status).
16359 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16360 to '-clrext' (= clear extensions), as intended and documented.
16361 *Bodo Moeller; inconsistency pointed out by Michael Attili
16362 <attili@amaxo.com>*
16364 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16365 was larger than the MD block size.
16367 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16369 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16370 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16371 using the passed key: if the passed key was a private key the result
16372 of X509_print(), for example, would be to print out all the private key
16377 * des_quad_cksum() byte order bug fix.
16378 *Ulf Möller, using the problem description in krb4-0.9.7, where
16379 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16381 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16384 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16386 * For easily testing in shell scripts whether some command
16387 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16388 returns with exit code 0 iff no command of the given name is available.
16389 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16390 the output goes to stdout and nothing is printed to stderr.
16391 Additional arguments are always ignored.
16393 Since for each cipher there is a command of the same name,
16394 the 'no-cipher' compilation switches can be tested this way.
16396 ('openssl no-XXX' is not able to detect pseudo-commands such
16397 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16401 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16405 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16406 is set; it will be thrown away anyway because each handshake creates
16408 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16409 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16410 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16411 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16415 * New s_client option -ign_eof: EOF at stdin is ignored, and
16416 'Q' and 'R' lose their special meanings (quit/renegotiate).
16417 This is part of what -quiet does; unlike -quiet, -ign_eof
16418 does not suppress any output.
16422 * Add compatibility options to the purpose and trust code. The
16423 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16424 accepts a certificate or CA, this was the previous behaviour,
16425 with all the associated security issues.
16427 X509_TRUST_COMPAT is the old trust behaviour: only and
16428 automatically trust self signed roots in certificate store. A
16429 new trust setting X509_TRUST_DEFAULT is used to specify that
16430 a purpose has no associated trust setting and it should instead
16431 use the value in the default purpose.
16435 * Fix the PKCS#8 DSA private key code so it decodes keys again
16436 and fix a memory leak.
16440 * In util/mkerr.pl (which implements 'make errors'), preserve
16441 reason strings from the previous version of the .c file, as
16442 the default to have only downcase letters (and digits) in
16443 automatically generated reasons codes is not always appropriate.
16447 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16448 using strerror. Previously, ERR_reason_error_string() returned
16449 library names as reason strings for SYSerr; but SYSerr is a special
16450 case where small numbers are errno values, not library numbers.
16454 * Add '-dsaparam' option to 'openssl dhparam' application. This
16455 converts DSA parameters into DH parameters. (When creating parameters,
16456 DSA_generate_parameters is used.)
16460 * Include 'length' (recommended exponent length) in C code generated
16461 by 'openssl dhparam -C'.
16465 * The second argument to set_label in perlasm was already being used
16466 so couldn't be used as a "file scope" flag. Moved to third argument
16471 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16472 instead of RAND_bytes for encryption IVs and salts.
16476 * Include RAND_status() into RAND_METHOD instead of implementing
16477 it only for md_rand.c Otherwise replacing the PRNG by calling
16478 RAND_set_rand_method would be impossible.
16482 * Don't let DSA_generate_key() enter an infinite loop if the random
16483 number generation fails.
16487 * New 'rand' application for creating pseudo-random output.
16491 * Added configuration support for Linux/IA64
16493 *Rolf Haberrecker <rolf@suse.de>*
16495 * Assembler module support for Mingw32.
16499 * Shared library support for HPUX (in shlib/).
16501 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16503 * Shared library support for Solaris gcc.
16505 *Lutz Behnke <behnke@trustcenter.de>*
16507 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16509 * PKCS7_encrypt() was adding text MIME headers twice because they
16510 were added manually and by SMIME_crlf_copy().
16514 * In bntest.c don't call BN_rand with zero bits argument.
16516 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16518 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16519 case was implemented. This caused BN_div_recp() to fail occasionally.
16523 * Add an optional second argument to the set_label() in the perl
16524 assembly language builder. If this argument exists and is set
16525 to 1 it signals that the assembler should use a symbol whose
16526 scope is the entire file, not just the current function. This
16527 is needed with MASM which uses the format label:: for this scope.
16529 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16531 * Change the ASN1 types so they are typedefs by default. Before
16532 almost all types were #define'd to ASN1_STRING which was causing
16533 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16538 * Change names of new functions to the new get1/get0 naming
16539 convention: After 'get1', the caller owns a reference count
16540 and has to call `..._free`; 'get0' returns a pointer to some
16541 data structure without incrementing reference counters.
16542 (Some of the existing 'get' functions increment a reference
16543 counter, some don't.)
16544 Similarly, 'set1' and 'add1' functions increase reference
16545 counters or duplicate objects.
16549 * Allow for the possibility of temp RSA key generation failure:
16550 the code used to assume it always worked and crashed on failure.
16554 * Fix potential buffer overrun problem in BIO_printf().
16555 *Ulf Möller, using public domain code by Patrick Powell; problem
16556 pointed out by David Sacerdote <das33@cornell.edu>*
16558 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16559 RAND_egd() and RAND_status(). In the command line application,
16560 the EGD socket can be specified like a seed file using RANDFILE
16565 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16566 Some CAs (e.g. Verisign) distribute certificates in this form.
16570 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16571 list to exclude them. This means that no special compilation option
16572 is needed to use anonymous DH: it just needs to be included in the
16577 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16578 EVP_MD_type. The old functionality is available in a new macro called
16579 EVP_MD_md(). Change code that uses it and update docs.
16583 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16584 where the `void *` argument is replaced by a function pointer argument.
16585 Previously `void *` was abused to point to functions, which works on
16586 many platforms, but is not correct. As these functions are usually
16587 called by macros defined in OpenSSL header files, most source code
16588 should work without changes.
16592 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16593 sections with information on -D... compiler switches used for
16594 compiling the library so that applications can see them. To enable
16595 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16596 must be defined. E.g.,
16597 #define OPENSSL_ALGORITHM_DEFINES
16598 #include <openssl/opensslconf.h>
16599 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16601 *Richard Levitte, Ulf and Bodo Möller*
16603 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16608 * Change the 'other' type in certificate aux info to a STACK_OF
16609 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16610 the required ASN1 format: arbitrary types determined by an OID.
16614 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16615 argument to 'req'. This is not because the function is newer or
16616 better than others it just uses the work 'NEW' in the certificate
16617 request header lines. Some software needs this.
16621 * Reorganise password command line arguments: now passwords can be
16622 obtained from various sources. Delete the PEM_cb function and make
16623 it the default behaviour: i.e. if the callback is NULL and the
16624 usrdata argument is not NULL interpret it as a null terminated pass
16625 phrase. If usrdata and the callback are NULL then the pass phrase
16626 is prompted for as usual.
16630 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16631 the support is automatically enabled. The resulting binaries will
16632 autodetect the card and use it if present.
16634 *Ben Laurie and Compaq Inc.*
16636 * Work around for Netscape hang bug. This sends certificate request
16637 and server done in one record. Since this is perfectly legal in the
16638 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16639 the bugs/SSLv3 entry for more info.
16643 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16647 * Add -rand argument to smime and pkcs12 applications and read/write
16652 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16656 * Add command line password options to the remaining applications.
16660 * Bug fix for BN_div_recp() for numerators with an even number of
16665 * More tests in bntest.c, and changed test_bn output.
16669 * ./config recognizes MacOS X now.
16673 * Bug fix for BN_div() when the first words of num and divisor are
16674 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16678 * Add support for various broken PKCS#8 formats, and command line
16679 options to produce them.
16683 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16684 get temporary BIGNUMs from a BN_CTX.
16688 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16693 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16694 include a #define from the old name to the new. The original intent
16695 was that statically linked binaries could for example just call
16696 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16697 link with digests. This never worked because SSLeay_add_all_digests()
16698 and SSLeay_add_all_ciphers() were in the same source file so calling
16699 one would link with the other. They are now in separate source files.
16703 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16707 * Use a less unusual form of the Miller-Rabin primality test (it used
16708 a binary algorithm for exponentiation integrated into the Miller-Rabin
16709 loop, our standard modexp algorithms are faster).
16713 * Support for the EBCDIC character set completed.
16715 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16717 * Source code cleanups: use const where appropriate, eliminate casts,
16718 use `void *` instead of `char *` in lhash.
16722 * Bugfix: ssl3_send_server_key_exchange was not restartable
16723 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16724 this the server could overwrite ephemeral keys that the client
16729 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16730 using 50 iterations of the Rabin-Miller test.
16732 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16733 iterations of the Rabin-Miller test as required by the appendix
16734 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16735 As BN_is_prime_fasttest includes trial division, DSA parameter
16736 generation becomes much faster.
16738 This implies a change for the callback functions in DSA_is_prime
16739 and DSA_generate_parameters: The callback function is called once
16740 for each positive witness in the Rabin-Miller test, not just
16741 occasionally in the inner loop; and the parameters to the
16742 callback function now provide an iteration count for the outer
16743 loop rather than for the current invocation of the inner loop.
16744 DSA_generate_parameters additionally can call the callback
16745 function with an 'iteration count' of -1, meaning that a
16746 candidate has passed the trial division test (when q is generated
16747 from an application-provided seed, trial division is skipped).
16751 * New function BN_is_prime_fasttest that optionally does trial
16752 division before starting the Rabin-Miller test and has
16753 an additional BN_CTX * argument (whereas BN_is_prime always
16754 has to allocate at least one BN_CTX).
16755 'callback(1, -1, cb_arg)' is called when a number has passed the
16756 trial division stage.
16760 * Fix for bug in CRL encoding. The validity dates weren't being handled
16765 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16769 * New function BN_pseudo_rand().
16773 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16774 bignum version of BN_from_montgomery() with the working code from
16775 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16780 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16781 made it impossible to use the same SSL_SESSION data structure in
16782 SSL2 clients in multiple threads.
16786 * The return value of RAND_load_file() no longer counts bytes obtained
16787 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16788 to seed the PRNG (previously an explicit byte count was required).
16790 *Ulf Möller, Bodo Möller*
16792 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16793 used `char *` instead of `void *` and had casts all over the place.
16797 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16801 * Retain source code compatibility for BN_prime_checks macro:
16802 BN_is_prime(..., BN_prime_checks, ...) now uses
16803 BN_prime_checks_for_size to determine the appropriate number of
16804 Rabin-Miller iterations.
16808 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16809 DH_CHECK_P_NOT_SAFE_PRIME.
16810 (Check if this is true? OpenPGP calls them "strong".)
16814 * Merge the functionality of "dh" and "gendh" programs into a new program
16815 "dhparam". The old programs are retained for now but will handle DH keys
16816 (instead of parameters) in future.
16820 * Make the ciphers, s_server and s_client programs check the return values
16821 when a new cipher list is set.
16825 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16826 ciphers. Before when the 56bit ciphers were enabled the sorting was
16829 The syntax for the cipher sorting has been extended to support sorting by
16830 cipher-strength (using the strength_bits hard coded in the tables).
16831 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16833 Fix a bug in the cipher-command parser: when supplying a cipher command
16834 string with an "undefined" symbol (neither command nor alphanumeric
16835 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16836 an error is flagged.
16838 Due to the strength-sorting extension, the code of the
16839 ssl_create_cipher_list() function was completely rearranged. I hope that
16840 the readability was also increased :-)
16842 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16844 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16845 for the first serial number and places 2 in the serial number file. This
16846 avoids problems when the root CA is created with serial number zero and
16847 the first user certificate has the same issuer name and serial number
16852 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16853 the new code. Add documentation for this stuff.
16857 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16858 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16859 structures and behave in an analogous way to the X509v3 functions:
16860 they shouldn't be called directly but wrapper functions should be used
16863 So we also now have some wrapper functions that call the X509at functions
16864 when passed certificate requests. (TO DO: similar things can be done with
16865 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16866 things. Some of these need some d2i or i2d and print functionality
16867 because they handle more complex structures.)
16871 * Add missing #ifndefs that caused missing symbols when building libssl
16872 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16873 NO_RSA in `ssl/s2*.c`.
16875 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16877 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16878 has a return value which indicates the quality of the random data
16879 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16880 error queue. New function RAND_pseudo_bytes() generates output that is
16881 guaranteed to be unique but not unpredictable. RAND_add is like
16882 RAND_seed, but takes an extra argument for an entropy estimate
16883 (RAND_seed always assumes full entropy).
16887 * Do more iterations of Rabin-Miller probable prime test (specifically,
16888 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16889 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16890 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16891 false-positive rate of at most 2^-80 for random input.
16895 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16899 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16900 in the 0.9.5 release), this returns the chain
16901 from an X509_CTX structure with a dup of the stack and all
16902 the X509 reference counts upped: so the stack will exist
16903 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16906 Also make SSL_SESSION_print() print out the verify return
16911 * Add manpage for the pkcs12 command. Also change the default
16912 behaviour so MAC iteration counts are used unless the new
16913 -nomaciter option is used. This improves file security and
16914 only older versions of MSIE (4.0 for example) need it.
16918 * Honor the no-xxx Configure options when creating .DEF files.
16922 * Add PKCS#10 attributes to field table: challengePassword,
16923 unstructuredName and unstructuredAddress. These are taken from
16924 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16925 international characters are used.
16927 More changes to X509_ATTRIBUTE code: allow the setting of types
16928 based on strings. Remove the 'loc' parameter when adding
16929 attributes because these will be a SET OF encoding which is sorted
16934 * Initial changes to the 'req' utility to allow request generation
16935 automation. This will allow an application to just generate a template
16936 file containing all the field values and have req construct the
16939 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16940 used all over the place including certificate requests and PKCS#7
16941 structures. They are currently handled manually where necessary with
16942 some primitive wrappers for PKCS#7. The new functions behave in a
16943 manner analogous to the X509 extension functions: they allow
16944 attributes to be looked up by NID and added.
16946 Later something similar to the X509V3 code would be desirable to
16947 automatically handle the encoding, decoding and printing of the
16948 more complex types. The string types like challengePassword can
16949 be handled by the string table functions.
16951 Also modified the multi byte string table handling. Now there is
16952 a 'global mask' which masks out certain types. The table itself
16953 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16954 is useful when for example there is only one permissible type
16955 (as in countryName) and using the mask might result in no valid
16960 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16961 SSL_get_peer_finished to allow applications to obtain the latest
16962 Finished messages sent to the peer or expected from the peer,
16963 respectively. (SSL_get_peer_finished is usually the Finished message
16964 actually received from the peer, otherwise the protocol will be aborted.)
16966 As the Finished message are message digests of the complete handshake
16967 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16968 be used for external authentication procedures when the authentication
16969 provided by SSL/TLS is not desired or is not enough.
16973 * Enhanced support for Alpha Linux is added. Now ./config checks if
16974 the host supports BWX extension and if Compaq C is present on the
16975 $PATH. Just exploiting of the BWX extension results in 20-30%
16976 performance kick for some algorithms, e.g. DES and RC4 to mention
16977 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16982 * Add support for MS "fast SGC". This is arguably a violation of the
16983 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16984 weak crypto and after checking the certificate is SGC a second one
16985 with strong crypto. MS SGC stops the first handshake after receiving
16986 the server certificate message and sends a second client hello. Since
16987 a server will typically do all the time consuming operations before
16988 expecting any further messages from the client (server key exchange
16989 is the most expensive) there is little difference between the two.
16991 To get OpenSSL to support MS SGC we have to permit a second client
16992 hello message after we have sent server done. In addition we have to
16993 reset the MAC if we do get this second client hello.
16997 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16998 if a DER encoded private key is RSA or DSA traditional format. Changed
16999 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17000 format DER encoded private key. Newer code should use PKCS#8 format which
17001 has the key type encoded in the ASN1 structure. Added DER private key
17002 support to pkcs8 application.
17006 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17007 ciphersuites has been selected (as required by the SSL 3/TLS 1
17008 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17009 is set, we interpret this as a request to violate the specification
17010 (the worst that can happen is a handshake failure, and 'correct'
17011 behaviour would result in a handshake failure anyway).
17015 * In SSL_CTX_add_session, take into account that there might be multiple
17016 SSL_SESSION structures with the same session ID (e.g. when two threads
17017 concurrently obtain them from an external cache).
17018 The internal cache can handle only one SSL_SESSION with a given ID,
17019 so if there's a conflict, we now throw out the old one to achieve
17024 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17025 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17026 some routines that use cipher OIDs: some ciphers do not have OIDs
17027 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17032 * Simplify the trust setting structure and code. Now we just have
17033 two sequences of OIDs for trusted and rejected settings. These will
17034 typically have values the same as the extended key usage extension
17035 and any application specific purposes.
17037 The trust checking code now has a default behaviour: it will just
17038 check for an object with the same NID as the passed id. Functions can
17039 be provided to override either the default behaviour or the behaviour
17040 for a given id. SSL client, server and email already have functions
17041 in place for compatibility: they check the NID and also return "trusted"
17042 if the certificate is self signed.
17046 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17047 traditional format into an EVP_PKEY structure.
17051 * Add a password callback function PEM_cb() which either prompts for
17052 a password if usr_data is NULL or otherwise assumes it is a null
17053 terminated password. Allow passwords to be passed on command line
17054 environment or config files in a few more utilities.
17058 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17059 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17060 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17061 Update documentation.
17065 * Support for ASN1 "NULL" type. This could be handled before by using
17066 ASN1_TYPE but there wasn't any function that would try to read a NULL
17067 and produce an error if it couldn't. For compatibility we also have
17068 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17069 don't allocate anything because they don't need to.
17073 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17076 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17078 * Rebuild of the memory allocation routines used by OpenSSL code and
17079 possibly others as well. The purpose is to make an interface that
17080 provide hooks so anyone can build a separate set of allocation and
17081 deallocation routines to be used by OpenSSL, for example memory
17082 pool implementations, or something else, which was previously hard
17083 since Malloc(), Realloc() and Free() were defined as macros having
17084 the values malloc, realloc and free, respectively (except for Win32
17085 compilations). The same is provided for memory debugging code.
17086 OpenSSL already comes with functionality to find memory leaks, but
17087 this gives people a chance to debug other memory problems.
17089 With these changes, a new set of functions and macros have appeared:
17091 CRYPTO_set_mem_debug_functions() [F]
17092 CRYPTO_get_mem_debug_functions() [F]
17093 CRYPTO_dbg_set_options() [F]
17094 CRYPTO_dbg_get_options() [F]
17095 CRYPTO_malloc_debug_init() [M]
17097 The memory debug functions are NULL by default, unless the library
17098 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17099 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17100 gives the standard debugging functions that come with OpenSSL) or
17101 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17102 provided by the library user) must be used. When the standard
17103 debugging functions are used, CRYPTO_dbg_set_options can be used to
17104 request additional information:
17105 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17106 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17108 Also, things like CRYPTO_set_mem_functions will always give the
17109 expected result (the new set of functions is used for allocation
17110 and deallocation) at all times, regardless of platform and compiler
17113 To finish it up, some functions that were never use in any other
17114 way than through macros have a new API and new semantic:
17116 CRYPTO_dbg_malloc()
17117 CRYPTO_dbg_realloc()
17120 All macros of value have retained their old syntax.
17122 *Richard Levitte and Bodo Moeller*
17124 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17125 ordering of SMIMECapabilities wasn't in "strength order" and there
17126 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17131 * Some ASN1 types with illegal zero length encoding (INTEGER,
17132 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17134 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17136 * Merge in my S/MIME library for OpenSSL. This provides a simple
17137 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17138 functionality to handle multipart/signed properly) and a utility
17139 called 'smime' to call all this stuff. This is based on code I
17140 originally wrote for Celo who have kindly allowed it to be
17141 included in OpenSSL.
17145 * Add variants des_set_key_checked and des_set_key_unchecked of
17146 des_set_key (aka des_key_sched). Global variable des_check_key
17147 decides which of these is called by des_set_key; this way
17148 des_check_key behaves as it always did, but applications and
17149 the library itself, which was buggy for des_check_key == 1,
17150 have a cleaner way to pick the version they need.
17154 * New function PKCS12_newpass() which changes the password of a
17159 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17160 dynamic mix. In both cases the ids can be used as an index into the
17161 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17162 functions so they accept a list of the field values and the
17163 application doesn't need to directly manipulate the X509_TRUST
17168 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17173 * Modify the way the V3 extension code looks up extensions. This now
17174 works in a similar way to the object code: we have some "standard"
17175 extensions in a static table which is searched with OBJ_bsearch()
17176 and the application can add dynamic ones if needed. The file
17177 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17178 updated whenever a new extension is added to the core code and kept
17179 in ext_nid order. There is a simple program 'tabtest.c' which checks
17180 this. New extensions are not added too often so this file can readily
17181 be maintained manually.
17183 There are two big advantages in doing things this way. The extensions
17184 can be looked up immediately and no longer need to be "added" using
17185 X509V3_add_standard_extensions(): this function now does nothing.
17186 Side note: I get *lots* of email saying the extension code doesn't
17187 work because people forget to call this function.
17188 Also no dynamic allocation is done unless new extensions are added:
17189 so if we don't add custom extensions there is no need to call
17190 X509V3_EXT_cleanup().
17194 * Modify enc utility's salting as follows: make salting the default. Add a
17195 magic header, so unsalted files fail gracefully instead of just decrypting
17196 to garbage. This is because not salting is a big security hole, so people
17197 should be discouraged from doing it.
17201 * Fixes and enhancements to the 'x509' utility. It allowed a message
17202 digest to be passed on the command line but it only used this
17203 parameter when signing a certificate. Modified so all relevant
17204 operations are affected by the digest parameter including the
17205 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17206 DSA key was used because it didn't fix the digest.
17210 * Initial certificate chain verify code. Currently tests the untrusted
17211 certificates for consistency with the verify purpose (which is set
17212 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17214 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17215 this is because it will reject chains with invalid extensions whereas
17216 every previous version of OpenSSL and SSLeay made no checks at all.
17218 Trust code: checks the root CA for the relevant trust settings. Trust
17219 settings have an initial value consistent with the verify purpose: e.g.
17220 if the verify purpose is for SSL client use it expects the CA to be
17221 trusted for SSL client use. However the default value can be changed to
17222 permit custom trust settings: one example of this would be to only trust
17223 certificates from a specific "secure" set of CAs.
17225 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17226 which should be used for version portability: especially since the
17227 verify structure is likely to change more often now.
17229 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17230 to set them. If not set then assume SSL clients will verify SSL servers
17233 Two new options to the verify program: -untrusted allows a set of
17234 untrusted certificates to be passed in and -purpose which sets the
17235 intended purpose of the certificate. If a purpose is set then the
17236 new chain verify code is used to check extension consistency.
17240 * Support for the authority information access extension.
17244 * Modify RSA and DSA PEM read routines to transparently handle
17245 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17246 public keys in a format compatible with certificate
17247 SubjectPublicKeyInfo structures. Unfortunately there were already
17248 functions called *_PublicKey_* which used various odd formats so
17249 these are retained for compatibility: however the DSA variants were
17250 never in a public release so they have been deleted. Changed dsa/rsa
17251 utilities to handle the new format: note no releases ever handled public
17252 keys so we should be OK.
17254 The primary motivation for this change is to avoid the same fiasco
17255 that dogs private keys: there are several incompatible private key
17256 formats some of which are standard and some OpenSSL specific and
17257 require various evil hacks to allow partial transparent handling and
17258 even then it doesn't work with DER formats. Given the option anything
17259 other than PKCS#8 should be dumped: but the other formats have to
17260 stay in the name of compatibility.
17262 With public keys and the benefit of hindsight one standard format
17263 is used which works with EVP_PKEY, RSA or DSA structures: though
17264 it clearly returns an error if you try to read the wrong kind of key.
17266 Added a -pubkey option to the 'x509' utility to output the public key.
17267 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17268 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17269 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17270 that do the same as the `EVP_PKEY_assign_*()` except they up the
17271 reference count of the added key (they don't "swallow" the
17276 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17277 CRLs would fail if the file contained no certificates or no CRLs:
17278 added a new function to read in both types and return the number
17279 read: this means that if none are read it will be an error. The
17280 DER versions of the certificate and CRL reader would always fail
17281 because it isn't possible to mix certificates and CRLs in DER format
17282 without choking one or the other routine. Changed this to just read
17283 a certificate: this is the best we can do. Also modified the code
17284 in `apps/verify.c` to take notice of return codes: it was previously
17285 attempting to read in certificates from NULL pointers and ignoring
17286 any errors: this is one reason why the cert and CRL reader seemed
17287 to work. It doesn't check return codes from the default certificate
17288 routines: these may well fail if the certificates aren't installed.
17292 * Code to support otherName option in GeneralName.
17296 * First update to verify code. Change the verify utility
17297 so it warns if it is passed a self signed certificate:
17298 for consistency with the normal behaviour. X509_verify
17299 has been modified to it will now verify a self signed
17300 certificate if *exactly* the same certificate appears
17301 in the store: it was previously impossible to trust a
17302 single self signed certificate. This means that:
17303 openssl verify ss.pem
17304 now gives a warning about a self signed certificate but
17305 openssl verify -CAfile ss.pem ss.pem
17310 * For servers, store verify_result in SSL_SESSION data structure
17311 (and add it to external session representation).
17312 This is needed when client certificate verifications fails,
17313 but an application-provided verification callback (set by
17314 SSL_CTX_set_cert_verify_callback) allows accepting the session
17315 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17316 but returns 1): When the session is reused, we have to set
17317 ssl->verify_result to the appropriate error code to avoid
17320 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17322 * Fix a bug in the new PKCS#7 code: it didn't consider the
17323 case in PKCS7_dataInit() where the signed PKCS7 structure
17324 didn't contain any existing data because it was being created.
17326 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17328 * Add a salt to the key derivation routines in enc.c. This
17329 forms the first 8 bytes of the encrypted file. Also add a
17330 -S option to allow a salt to be input on the command line.
17334 * New function X509_cmp(). Oddly enough there wasn't a function
17335 to compare two certificates. We do this by working out the SHA1
17336 hash and comparing that. X509_cmp() will be needed by the trust
17341 * SSL_get1_session() is like SSL_get_session(), but increments
17342 the reference count in the SSL_SESSION returned.
17344 *Geoff Thorpe <geoff@eu.c2.net>*
17346 * Fix for 'req': it was adding a null to request attributes.
17347 Also change the X509_LOOKUP and X509_INFO code to handle
17348 certificate auxiliary information.
17352 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17357 * Add the possibility to add extra information to the memory leak
17358 detecting output, to form tracebacks, showing from where each
17359 allocation was originated: CRYPTO_push_info("constant string") adds
17360 the string plus current file name and line number to a per-thread
17361 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17362 is like calling CYRPTO_pop_info() until the stack is empty.
17363 Also updated memory leak detection code to be multi-thread-safe.
17367 * Add options -text and -noout to pkcs7 utility and delete the
17368 encryption options which never did anything. Update docs.
17372 * Add options to some of the utilities to allow the pass phrase
17373 to be included on either the command line (not recommended on
17374 OSes like Unix) or read from the environment. Update the
17375 manpages and fix a few bugs.
17379 * Add a few manpages for some of the openssl commands.
17383 * Fix the -revoke option in ca. It was freeing up memory twice,
17384 leaking and not finding already revoked certificates.
17388 * Extensive changes to support certificate auxiliary information.
17389 This involves the use of X509_CERT_AUX structure and X509_AUX
17390 functions. An X509_AUX function such as PEM_read_X509_AUX()
17391 can still read in a certificate file in the usual way but it
17392 will also read in any additional "auxiliary information". By
17393 doing things this way a fair degree of compatibility can be
17394 retained: existing certificates can have this information added
17395 using the new 'x509' options.
17397 Current auxiliary information includes an "alias" and some trust
17398 settings. The trust settings will ultimately be used in enhanced
17399 certificate chain verification routines: currently a certificate
17400 can only be trusted if it is self signed and then it is trusted
17405 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17406 The problem was that one of the replacement routines had not been working
17407 since SSLeay releases. For now the offending routine has been replaced
17408 with non-optimised assembler. Even so, this now gives around 95%
17409 performance improvement for 1024 bit RSA signs.
17413 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17414 handling. Most clients have the effective key size in bits equal to
17415 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17416 A few however don't do this and instead use the size of the decrypted key
17417 to determine the RC2 key length and the AlgorithmIdentifier to determine
17418 the effective key length. In this case the effective key length can still
17419 be 40 bits but the key length can be 168 bits for example. This is fixed
17420 by manually forcing an RC2 key into the EVP_PKEY structure because the
17421 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17422 the key length and effective key length are equal.
17426 * Add a bunch of functions that should simplify the creation of
17427 X509_NAME structures. Now you should be able to do:
17428 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17429 and have it automatically work out the correct field type and fill in
17430 the structures. The more adventurous can try:
17431 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17432 and it will (hopefully) work out the correct multibyte encoding.
17436 * Change the 'req' utility to use the new field handling and multibyte
17437 copy routines. Before the DN field creation was handled in an ad hoc
17438 way in req, ca, and x509 which was rather broken and didn't support
17439 BMPStrings or UTF8Strings. Since some software doesn't implement
17440 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17441 using the dirstring_type option. See the new comment in the default
17442 openssl.cnf for more info.
17446 * Make crypto/rand/md_rand.c more robust:
17447 - Assure unique random numbers after fork().
17448 - Make sure that concurrent threads access the global counter and
17449 md serializably so that we never lose entropy in them
17450 or use exactly the same state in multiple threads.
17451 Access to the large state is not always serializable because
17452 the additional locking could be a performance killer, and
17453 md should be large enough anyway.
17457 * New file `apps/app_rand.c` with commonly needed functionality
17458 for handling the random seed file.
17460 Use the random seed file in some applications that previously did not:
17462 dsaparam -genkey (which also ignored its '-rand' option),
17465 x509 (when signing).
17466 Except on systems with /dev/urandom, it is crucial to have a random
17467 seed file at least for key creation, DSA signing, and for DH exchanges;
17468 for RSA signatures we could do without one.
17470 gendh and gendsa (unlike genrsa) used to read only the first byte
17471 of each file listed in the '-rand' option. The function as previously
17472 found in genrsa is now in app_rand.c and is used by all programs
17473 that support '-rand'.
17477 * In RAND_write_file, use mode 0600 for creating files;
17478 don't just chmod when it may be too late.
17482 * Report an error from X509_STORE_load_locations
17483 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17487 * New function ASN1_mbstring_copy() this copies a string in either
17488 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17489 into an ASN1_STRING type. A mask of permissible types is passed
17490 and it chooses the "minimal" type to use or an error if not type
17495 * Add function equivalents to the various macros in asn1.h. The old
17496 macros are retained with an `M_` prefix. Code inside the library can
17497 use the `M_` macros. External code (including the openssl utility)
17498 should *NOT* in order to be "shared library friendly".
17502 * Add various functions that can check a certificate's extensions
17503 to see if it usable for various purposes such as SSL client,
17504 server or S/MIME and CAs of these types. This is currently
17505 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17506 verification. Also added a -purpose flag to x509 utility to
17507 print out all the purposes.
17511 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17516 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17517 for, obtain and decode and extension and obtain its critical flag.
17518 This allows all the necessary extension code to be handled in a
17519 single function call.
17523 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17524 platforms. See crypto/rc4/rc4_enc.c for further details.
17528 * New -noout option to asn1parse. This causes no output to be produced
17529 its main use is when combined with -strparse and -out to extract data
17530 from a file (which may not be in ASN.1 format).
17534 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17535 when producing the local key id.
17537 *Richard Levitte <levitte@stacken.kth.se>*
17539 * New option -dhparam in s_server. This allows a DH parameter file to be
17540 stated explicitly. If it is not stated then it tries the first server
17541 certificate file. The previous behaviour hard coded the filename
17546 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17547 a public key to be input or output. For example:
17548 openssl rsa -in key.pem -pubout -out pubkey.pem
17549 Also added necessary DSA public key functions to handle this.
17553 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17554 in the message. This was handled by allowing
17555 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17557 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17559 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17560 to the end of the strings whereas this didn't. This would cause problems
17561 if strings read with d2i_ASN1_bytes() were later modified.
17563 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17565 * Fix for base64 decode bug. When a base64 bio reads only one line of
17566 data and it contains EOF it will end up returning an error. This is
17567 caused by input 46 bytes long. The cause is due to the way base64
17568 BIOs find the start of base64 encoded data. They do this by trying a
17569 trial decode on each line until they find one that works. When they
17570 do a flag is set and it starts again knowing it can pass all the
17571 data directly through the decoder. Unfortunately it doesn't reset
17572 the context it uses. This means that if EOF is reached an attempt
17573 is made to pass two EOFs through the context and this causes the
17574 resulting error. This can also cause other problems as well. As is
17575 usual with these problems it takes *ages* to find and the fix is
17576 trivial: move one line.
17578 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17580 * Ugly workaround to get s_client and s_server working under Windows. The
17581 old code wouldn't work because it needed to select() on sockets and the
17582 tty (for keypresses and to see if data could be written). Win32 only
17583 supports select() on sockets so we select() with a 1s timeout on the
17584 sockets and then see if any characters are waiting to be read, if none
17585 are present then we retry, we also assume we can always write data to
17586 the tty. This isn't nice because the code then blocks until we've
17587 received a complete line of data and it is effectively polling the
17588 keyboard at 1s intervals: however it's quite a bit better than not
17589 working at all :-) A dedicated Windows application might handle this
17590 with an event loop for example.
17594 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17595 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17596 will be called when RSA_sign() and RSA_verify() are used. This is useful
17597 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17598 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17599 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17600 This necessitated the support of an extra signature type NID_md5_sha1
17601 for SSL signatures and modifications to the SSL library to use it instead
17602 of calling RSA_public_decrypt() and RSA_private_encrypt().
17606 * Add new -verify -CAfile and -CApath options to the crl program, these
17607 will lookup a CRL issuers certificate and verify the signature in a
17608 similar way to the verify program. Tidy up the crl program so it
17609 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17610 less strict. It will now permit CRL extensions even if it is not
17611 a V2 CRL: this will allow it to tolerate some broken CRLs.
17615 * Initialize all non-automatic variables each time one of the openssl
17616 sub-programs is started (this is necessary as they may be started
17617 multiple times from the "OpenSSL>" prompt).
17619 *Lennart Bang, Bodo Moeller*
17621 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17622 removing all other RSA functionality (this is what NO_RSA does). This
17623 is so (for example) those in the US can disable those operations covered
17624 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17629 * Non-copying interface to BIO pairs.
17630 (still largely untested)
17634 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17635 ASCII string. This was handled independently in various places before.
17639 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17640 UTF8 strings a character at a time.
17644 * Use client_version from client hello to select the protocol
17645 (s23_srvr.c) and for RSA client key exchange verification
17646 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17650 * Add various utility functions to handle SPKACs, these were previously
17651 handled by poking round in the structure internals. Added new function
17652 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17653 print, verify and generate SPKACs. Based on an original idea from
17654 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17658 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17662 * Allow the config file extension section to be overwritten on the
17663 command line. Based on an original idea from Massimiliano Pala
17664 <madwolf@comune.modena.it>. The new option is called -extensions
17665 and can be applied to ca, req and x509. Also -reqexts to override
17666 the request extensions in req and -crlexts to override the crl extensions
17671 * Add new feature to the SPKAC handling in ca. Now you can include
17672 the same field multiple times by preceding it by "XXXX." for example:
17675 this is the same syntax as used in the req config file.
17679 * Allow certificate extensions to be added to certificate requests. These
17680 are specified in a 'req_extensions' option of the req section of the
17681 config file. They can be printed out with the -text option to req but
17682 are otherwise ignored at present.
17686 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17687 data read consists of only the final block it would not decrypted because
17688 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17689 A misplaced 'break' also meant the decrypted final block might not be
17690 copied until the next read.
17694 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17695 a few extra parameters to the DH structure: these will be useful if
17696 for example we want the value of 'q' or implement X9.42 DH.
17700 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17701 provides hooks that allow the default DSA functions or functions on a
17702 "per key" basis to be replaced. This allows hardware acceleration and
17703 hardware key storage to be handled without major modification to the
17704 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17705 associated functions.
17709 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17710 as "read only": it can't be written to and the buffer it points to will
17711 not be freed. Reading from a read only BIO is much more efficient than
17712 a normal memory BIO. This was added because there are several times when
17713 an area of memory needs to be read from a BIO. The previous method was
17714 to create a memory BIO and write the data to it, this results in two
17715 copies of the data and an O(n^2) reading algorithm. There is a new
17716 function BIO_new_mem_buf() which creates a read only memory BIO from
17717 an area of memory. Also modified the PKCS#7 routines to use read only
17722 * Bugfix: ssl23_get_client_hello did not work properly when called in
17723 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17724 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17725 but a retry condition occurred while trying to read the rest.
17729 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17730 NID_pkcs7_encrypted by default: this was wrong since this should almost
17731 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17732 the encrypted data type: this is a more sensible place to put it and it
17733 allows the PKCS#12 code to be tidied up that duplicated this
17738 * Changed obj_dat.pl script so it takes its input and output files on
17739 the command line. This should avoid shell escape redirection problems
17744 * Initial support for certificate extension requests, these are included
17745 in things like Xenroll certificate requests. Included functions to allow
17746 extensions to be obtained and added.
17750 * -crlf option to s_client and s_server for sending newlines as
17751 CRLF (as required by many protocols).
17755 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17757 * Install libRSAglue.a when OpenSSL is built with RSAref.
17759 *Ralf S. Engelschall*
17761 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17763 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17765 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17770 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17771 DH parameters/keys (q is lost during that conversion, but the resulting
17772 DH parameters contain its length).
17774 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17775 much faster than DH_generate_parameters (which creates parameters
17776 where `p = 2*q + 1`), and also the smaller q makes DH computations
17777 much more efficient (160-bit exponentiation instead of 1024-bit
17778 exponentiation); so this provides a convenient way to support DHE
17779 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17780 utter importance to use
17781 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17783 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17784 when such DH parameters are used, because otherwise small subgroup
17785 attacks may become possible!
17789 * Avoid memory leak in i2d_DHparams.
17793 * Allow the -k option to be used more than once in the enc program:
17794 this allows the same encrypted message to be read by multiple recipients.
17798 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17799 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17800 it will always use the numerical form of the OID, even if it has a short
17805 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17806 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17807 otherwise bn_mod_exp was called. In the case of hardware keys for example
17808 no private key components need be present and it might store extra data
17809 in the RSA structure, which cannot be accessed from bn_mod_exp.
17810 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17811 private key operations.
17815 * Added support for SPARC Linux.
17819 * pem_password_cb function type incompatibly changed from
17820 typedef int pem_password_cb(char *buf, int size, int rwflag);
17822 ....(char *buf, int size, int rwflag, void *userdata);
17823 so that applications can pass data to their callbacks:
17824 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17825 additional void * argument, which is just handed through whenever
17826 the password callback is called.
17828 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17830 New function SSL_CTX_set_default_passwd_cb_userdata.
17832 Compatibility note: As many C implementations push function arguments
17833 onto the stack in reverse order, the new library version is likely to
17834 interoperate with programs that have been compiled with the old
17835 pem_password_cb definition (PEM_whatever takes some data that
17836 happens to be on the stack as its last argument, and the callback
17837 just ignores this garbage); but there is no guarantee whatsoever that
17840 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17841 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17842 problems not only on Windows, but also on some Unix platforms.
17843 To avoid problematic command lines, these definitions are now in an
17844 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17845 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17849 * MIPS III/IV assembler module is reimplemented.
17853 * More DES library cleanups: remove references to srand/rand and
17854 delete an unused file.
17858 * Add support for the free Netwide assembler (NASM) under Win32,
17859 since not many people have MASM (ml) and it can be hard to obtain.
17860 This is currently experimental but it seems to work OK and pass all
17861 the tests. Check out INSTALL.W32 for info.
17865 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17866 without temporary keys kept an extra copy of the server key,
17867 and connections with temporary keys did not free everything in case
17872 * New function RSA_check_key and new openssl rsa option -check
17873 for verifying the consistency of RSA keys.
17875 *Ulf Moeller, Bodo Moeller*
17877 * Various changes to make Win32 compile work:
17878 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17879 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17880 comparison" warnings.
17881 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17885 * Add a debugging option to PKCS#5 v2 key generation function: when
17886 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17887 derived keys are printed to stderr.
17891 * Copy the flags in ASN1_STRING_dup().
17893 *Roman E. Pavlov <pre@mo.msk.ru>*
17895 * The x509 application mishandled signing requests containing DSA
17896 keys when the signing key was also DSA and the parameters didn't match.
17898 It was supposed to omit the parameters when they matched the signing key:
17899 the verifying software was then supposed to automatically use the CA's
17900 parameters if they were absent from the end user certificate.
17902 Omitting parameters is no longer recommended. The test was also
17903 the wrong way round! This was probably due to unusual behaviour in
17904 EVP_cmp_parameters() which returns 1 if the parameters match.
17905 This meant that parameters were omitted when they *didn't* match and
17906 the certificate was useless. Certificates signed with 'ca' didn't have
17909 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17911 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17912 The interface is as follows:
17913 Applications can use
17914 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17915 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17916 "off" is now the default.
17917 The library internally uses
17918 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17919 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17920 to disable memory-checking temporarily.
17922 Some inconsistent states that previously were possible (and were
17923 even the default) are now avoided.
17925 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17926 with each memory chunk allocated; this is occasionally more helpful
17927 than just having a counter.
17929 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17931 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17936 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17937 which largely parallels "options", but is for changing API behaviour,
17938 whereas "options" are about protocol behaviour.
17939 Initial "mode" flags are:
17941 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17942 a single record has been written.
17943 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17944 retries use the same buffer location.
17945 (But all of the contents must be
17950 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17953 * Fix problems with no-hmac etc.
17955 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17957 * New functions RSA_get_default_method(), RSA_set_method() and
17958 RSA_get_method(). These allows replacement of RSA_METHODs without having
17959 to mess around with the internals of an RSA structure.
17963 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17964 Also really enable memory leak checks in openssl.c and in some
17967 *Chad C. Mulligan, Bodo Moeller*
17969 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17970 up the length of negative integers. This has now been simplified to just
17971 store the length when it is first determined and use it later, rather
17972 than trying to keep track of where data is copied and updating it to
17974 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17976 * Add a new function PKCS7_signatureVerify. This allows the verification
17977 of a PKCS#7 signature but with the signing certificate passed to the
17978 function itself. This contrasts with PKCS7_dataVerify which assumes the
17979 certificate is present in the PKCS#7 structure. This isn't always the
17980 case: certificates can be omitted from a PKCS#7 structure and be
17981 distributed by "out of band" means (such as a certificate database).
17985 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17986 function prototypes in pem.h, also change util/mkdef.pl to add the
17987 necessary function names.
17991 * mk1mf.pl (used by Windows builds) did not properly read the
17992 options set by Configure in the top level Makefile, and Configure
17993 was not even able to write more than one option correctly.
17994 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17998 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17999 file to be loaded from a BIO or FILE pointer. The BIO version will
18000 for example allow memory BIOs to contain config info.
18004 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18005 Whoever hopes to achieve shared-library compatibility across versions
18006 must use this, not the compile-time macro.
18007 (Exercise 0.9.4: Which is the minimum library version required by
18009 Note: All this applies only to multi-threaded programs, others don't
18014 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18015 through a BIO pair triggered the default case, i.e.
18016 SSLerr(...,SSL_R_UNKNOWN_STATE).
18020 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18021 can use the SSL library even if none of the specific BIOs is
18026 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18027 for the encoded length.
18029 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18031 * Add initial documentation of the X509V3 functions.
18035 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18036 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18037 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18038 secure PKCS#8 private key format with a high iteration count.
18042 * Fix determination of Perl interpreter: A perl or perl5
18043 *directory* in $PATH was also accepted as the interpreter.
18045 *Ralf S. Engelschall*
18047 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18048 wrong with it but it was very old and did things like calling
18049 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18050 unusual formatting.
18054 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18055 to use the new extension code.
18059 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18060 with macros. This should make it easier to change their form, add extra
18061 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18066 * Add to configuration table a new entry that can specify an alternative
18067 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18068 according to Mark Crispin <MRC@Panda.COM>.
18072 * DES CBC did not update the IV. Weird.
18076 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18077 Changing the behaviour of the former might break existing programs --
18078 where IV updating is needed, des_ncbc_encrypt can be used.
18081 * When bntest is run from "make test" it drives bc to check its
18082 calculations, as well as internally checking them. If an internal check
18083 fails, it needs to cause bc to give a non-zero result or make test carries
18084 on without noticing the failure. Fixed.
18088 * DES library cleanups.
18092 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18093 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18094 ciphers. NOTE: although the key derivation function has been verified
18095 against some published test vectors it has not been extensively tested
18096 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18101 * Instead of "mkdir -p", which is not fully portable, use new
18102 Perl script "util/mkdir-p.pl".
18106 * Rewrite the way password based encryption (PBE) is handled. It used to
18107 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18108 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18109 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18110 the 'parameter' field of the AlgorithmIdentifier is passed to the
18111 underlying key generation function so it must do its own ASN1 parsing.
18112 This has also changed the EVP_PBE_CipherInit() function which now has a
18113 'parameter' argument instead of literal salt and iteration count values
18114 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18118 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18119 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18120 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18121 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18122 value was just used as a "magic string" and not used directly its
18123 value doesn't matter.
18127 * Introduce some semblance of const correctness to BN. Shame C doesn't
18132 * "linux-sparc64" configuration (ultrapenguin).
18134 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18135 "linux-sparc" configuration.
18137 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18139 * config now generates no-xxx options for missing ciphers.
18143 * Support the EBCDIC character set (work in progress).
18144 File ebcdic.c not yet included because it has a different license.
18146 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18148 * Support BS2000/OSD-POSIX.
18150 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18152 * Make callbacks for key generation use `void *` instead of `char *`.
18156 * Make S/MIME samples compile (not yet tested).
18160 * Additional typesafe stacks.
18164 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18168 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18170 * New configuration variant "sco5-gcc".
18172 * Updated some demos.
18174 *Sean O Riordain, Wade Scholine*
18176 * Add missing BIO_free at exit of pkcs12 application.
18180 * Fix memory leak in conf.c.
18184 * Updates for Win32 to assembler version of MD5.
18188 * Set #! path to perl in `apps/der_chop` to where we found it
18189 instead of using a fixed path.
18193 * SHA library changes for irix64-mips4-cc.
18197 * Improvements for VMS support.
18201 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18203 * Bignum library bug fix. IRIX 6 passes "make test" now!
18204 This also avoids the problems with SC4.2 and unpatched SC5.
18206 *Andy Polyakov <appro@fy.chalmers.se>*
18208 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18209 These are required because of the typesafe stack would otherwise break
18210 existing code. If old code used a structure member which used to be STACK
18211 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18212 sk_num or sk_value it would produce an error because the num, data members
18213 are not present in STACK_OF. Now it just produces a warning. sk_set
18214 replaces the old method of assigning a value to sk_value
18215 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18216 that does this will no longer work (and should use sk_set instead) but
18217 this could be regarded as a "questionable" behaviour anyway.
18221 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18222 correctly handle encrypted S/MIME data.
18226 * Change type of various DES function arguments from des_cblock
18227 (which means, in function argument declarations, pointer to char)
18228 to des_cblock * (meaning pointer to array with 8 char elements),
18229 which allows the compiler to do more typechecking; it was like
18230 that back in SSLeay, but with lots of ugly casts.
18232 Introduce new type const_des_cblock.
18236 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18237 problems: find RecipientInfo structure that matches recipient certificate
18238 and initialise the ASN1 structures properly based on passed cipher.
18242 * Belatedly make the BN tests actually check the results.
18246 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18247 to and from BNs: it was completely broken. New compilation option
18248 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18249 key elements as negative integers.
18253 * Reorganize and speed up MD5.
18255 *Andy Polyakov <appro@fy.chalmers.se>*
18259 *Richard Levitte <richard@levitte.org>*
18261 * New option -out to asn1parse to allow the parsed structure to be
18262 output to a file. This is most useful when combined with the -strparse
18263 option to examine the output of things like OCTET STRINGS.
18267 * Make SSL library a little more fool-proof by not requiring any longer
18268 that `SSL_set_{accept,connect}_state` be called before
18269 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18270 in many applications because usually everything *appeared* to work as
18271 intended anyway -- now it really works as intended).
18275 * Move openssl.cnf out of lib/.
18279 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18280 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18281 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18283 *Ralf S. Engelschall*
18285 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18286 handle PKCS#7 enveloped data properly.
18288 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18290 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18291 copying pointers. The cert_st handling is changed by this in
18292 various ways (and thus what used to be known as ctx->default_cert
18293 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18294 any longer when s->cert does not give us what we need).
18295 ssl_cert_instantiate becomes obsolete by this change.
18296 As soon as we've got the new code right (possibly it already is?),
18297 we have solved a couple of bugs of the earlier code where s->cert
18298 was used as if it could not have been shared with other SSL structures.
18300 Note that using the SSL API in certain dirty ways now will result
18301 in different behaviour than observed with earlier library versions:
18302 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18303 does not influence s as it used to.
18305 In order to clean up things more thoroughly, inside SSL_SESSION
18306 we don't use CERT any longer, but a new structure SESS_CERT
18307 that holds per-session data (if available); currently, this is
18308 the peer's certificate chain and, for clients, the server's certificate
18309 and temporary key. CERT holds only those values that can have
18310 meaningful defaults in an SSL_CTX.
18314 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18315 from the internal representation. Various PKCS#7 fixes: remove some
18316 evil casts and set the enc_dig_alg field properly based on the signing
18321 * Allow PKCS#12 password to be set from the command line or the
18322 environment. Let 'ca' get its config file name from the environment
18323 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18328 * Allow certificate policies extension to use an IA5STRING for the
18329 organization field. This is contrary to the PKIX definition but
18330 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18335 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18336 without disallowing inline assembler and the like for non-pedantic builds.
18340 * Support Borland C++ builder.
18342 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18348 * SHA-1 cleanups and performance enhancements.
18350 *Andy Polyakov <appro@fy.chalmers.se>*
18352 * Sparc v8plus assembler for the bignum library.
18354 *Andy Polyakov <appro@fy.chalmers.se>*
18356 * Accept any -xxx and +xxx compiler options in Configure.
18360 * Update HPUX configuration.
18364 * Add missing `sk_<type>_unshift()` function to safestack.h
18366 *Ralf S. Engelschall*
18368 * New function SSL_CTX_use_certificate_chain_file that sets the
18369 "extra_cert"s in addition to the certificate. (This makes sense
18370 only for "PEM" format files, as chains as a whole are not
18375 * Support verify_depth from the SSL API.
18376 x509_vfy.c had what can be considered an off-by-one-error:
18377 Its depth (which was not part of the external interface)
18378 was actually counting the number of certificates in a chain;
18379 now it really counts the depth.
18383 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18384 instead of X509err, which often resulted in confusing error
18385 messages since the error codes are not globally unique
18386 (e.g. an alleged error in ssl3_accept when a certificate
18387 didn't match the private key).
18389 * New function SSL_CTX_set_session_id_context that allows to set a default
18390 value (so that you don't need SSL_set_session_id_context for each
18391 connection using the SSL_CTX).
18395 * OAEP decoding bug fix.
18399 * Support INSTALL_PREFIX for package builders, as proposed by
18404 * New Configure options "threads" and "no-threads". For systems
18405 where the proper compiler options are known (currently Solaris
18406 and Linux), "threads" is the default.
18410 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18414 * Install various scripts to $(OPENSSLDIR)/misc, not to
18415 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18416 such as /usr/local/bin.
18420 * "make linux-shared" to build shared libraries.
18422 *Niels Poppe <niels@netbox.org>*
18424 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18428 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18429 extension adding in x509 utility.
18433 * Remove NOPROTO sections and error code comments.
18437 * Partial rewrite of the DEF file generator to now parse the ANSI
18442 * New Configure options --prefix=DIR and --openssldir=DIR.
18446 * Complete rewrite of the error code script(s). It is all now handled
18447 by one script at the top level which handles error code gathering,
18448 header rewriting and C source file generation. It should be much better
18449 than the old method: it now uses a modified version of Ulf's parser to
18450 read the ANSI prototypes in all header files (thus the old K&R definitions
18451 aren't needed for error creation any more) and do a better job of
18452 translating function codes into names. The old 'ASN1 error code embedded
18453 in a comment' is no longer necessary and it doesn't use .err files which
18454 have now been deleted. Also the error code call doesn't have to appear all
18455 on one line (which resulted in some large lines...).
18459 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18463 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18464 0 (which usually indicates a closed connection), but continue reading.
18468 * Fix some race conditions.
18472 * Add support for CRL distribution points extension. Add Certificate
18473 Policies and CRL distribution points documentation.
18477 * Move the autogenerated header file parts to crypto/opensslconf.h.
18481 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18482 8 of keying material. Merlin has also confirmed interop with this fix
18483 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18485 *Merlin Hughes <merlin@baltimore.ie>*
18487 * Fix lots of warnings.
18489 *Richard Levitte <levitte@stacken.kth.se>*
18491 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18492 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18494 *Richard Levitte <levitte@stacken.kth.se>*
18496 * Fix problems with sizeof(long) == 8.
18498 *Andy Polyakov <appro@fy.chalmers.se>*
18500 * Change functions to ANSI C.
18504 * Fix typos in error codes.
18506 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18508 * Remove defunct assembler files from Configure.
18512 * SPARC v8 assembler BIGNUM implementation.
18514 *Andy Polyakov <appro@fy.chalmers.se>*
18516 * Support for Certificate Policies extension: both print and set.
18517 Various additions to support the r2i method this uses.
18521 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18522 return a const string when you are expecting an allocated buffer.
18526 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18527 types DirectoryString and DisplayText.
18531 * Add code to allow r2i extensions to access the configuration database,
18532 add an LHASH database driver and add several ctx helper functions.
18536 * Fix an evil bug in bn_expand2() which caused various BN functions to
18537 fail when they extended the size of a BIGNUM.
18541 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18542 support typesafe stack.
18546 * Fix typo in SSL_[gs]et_options().
18548 *Nils Frostberg <nils@medcom.se>*
18550 * Delete various functions and files that belonged to the (now obsolete)
18551 old X509V3 handling code.
18555 * New Configure option "rsaref".
18559 * Don't auto-generate pem.h.
18563 * Introduce type-safe ASN.1 SETs.
18567 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18569 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18571 * Introduce type-safe STACKs. This will almost certainly break lots of code
18572 that links with OpenSSL (well at least cause lots of warnings), but fear
18573 not: the conversion is trivial, and it eliminates loads of evil casts. A
18574 few STACKed things have been converted already. Feel free to convert more.
18575 In the fullness of time, I'll do away with the STACK type altogether.
18579 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18580 specified in `<certfile>` by updating the entry in the index.txt file.
18581 This way one no longer has to edit the index.txt file manually for
18582 revoking a certificate. The -revoke option does the gory details now.
18584 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18586 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18587 `-text` option at all and this way the `-noout -text` combination was
18588 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18590 *Ralf S. Engelschall*
18592 * Make sure a corresponding plain text error message exists for the
18593 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18594 verify callback function determined that a certificate was revoked.
18596 *Ralf S. Engelschall*
18598 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18599 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18600 all available ciphers including rc5, which was forgotten until now.
18601 In order to let the testing shell script know which algorithms
18602 are available, a new (up to now undocumented) command
18603 `openssl list-cipher-commands` is used.
18607 * Bugfix: s_client occasionally would sleep in select() when
18608 it should have checked SSL_pending() first.
18612 * New functions DSA_do_sign and DSA_do_verify to provide access to
18613 the raw DSA values prior to ASN.1 encoding.
18617 * Tweaks to Configure
18619 *Niels Poppe <niels@netbox.org>*
18621 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18626 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18630 * New config option to avoid instructions that are illegal on the 80386.
18631 The default code is faster, but requires at least a 486.
18635 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18636 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18637 same as SSL2_VERSION anyway.
18641 * New "-showcerts" option for s_client.
18645 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18646 application. Various cleanups and fixes.
18650 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18651 modify error routines to work internally. Add error codes and PBE init
18652 to library startup routines.
18656 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18657 packing functions to asn1 and evp. Changed function names and error
18658 codes along the way.
18662 * PKCS12 integration: and so it begins... First of several patches to
18663 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18664 objects to objects.h
18668 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18669 and display support for Thawte strong extranet extension.
18673 * Add LinuxPPC support.
18675 *Jeff Dubrule <igor@pobox.org>*
18677 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18678 bn_div_words in alpha.s.
18680 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18682 * Make sure the RSA OAEP test is skipped under -DRSAref because
18683 OAEP isn't supported when OpenSSL is built with RSAref.
18685 *Ulf Moeller <ulf@fitug.de>*
18687 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18688 so they no longer are missing under -DNOPROTO.
18690 *Soren S. Jorvang <soren@t.dk>*
18692 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18694 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18695 doesn't work when the session is reused. Coming soon!
18699 * Fix a security hole, that allows sessions to be reused in the wrong
18700 context thus bypassing client cert protection! All software that uses
18701 client certs and session caches in multiple contexts NEEDS PATCHING to
18702 allow session reuse! A fuller solution is in the works.
18704 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18706 * Some more source tree cleanups (removed obsolete files
18707 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18708 permission on "config" script to be executable) and a fix for the INSTALL
18711 *Ulf Moeller <ulf@fitug.de>*
18713 * Remove some legacy and erroneous uses of malloc, free instead of
18716 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18718 * Make rsa_oaep_test return non-zero on error.
18720 *Ulf Moeller <ulf@fitug.de>*
18722 * Add support for native Solaris shared libraries. Configure
18723 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18724 if someone would make that last step automatic.
18726 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18728 * ctx_size was not built with the right compiler during "make links". Fixed.
18732 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18733 except NULL ciphers". This means the default cipher list will no longer
18734 enable NULL ciphers. They need to be specifically enabled e.g. with
18735 the string "DEFAULT:eNULL".
18739 * Fix to RSA private encryption routines: if p < q then it would
18740 occasionally produce an invalid result. This will only happen with
18741 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18745 * Be less restrictive and allow also `perl util/perlpath.pl
18746 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18747 because this way one can also use an interpreter named `perl5` (which is
18748 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18749 installed as `perl`).
18751 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18753 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18755 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18757 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18758 advapi32.lib to Win32 build and change the pem test comparison
18759 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18760 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18761 and crypto/des/ede_cbcm_enc.c.
18765 * DES quad checksum was broken on big-endian architectures. Fixed.
18769 * Comment out two functions in bio.h that aren't implemented. Fix up the
18770 Win32 test batch file so it (might) work again. The Win32 test batch file
18771 is horrible: I feel ill....
18775 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18776 in e_os.h. Audit of header files to check ANSI and non ANSI
18777 sections: 10 functions were absent from non ANSI section and not exported
18778 from Windows DLLs. Fixed up libeay.num for new functions.
18782 * Make `openssl version` output lines consistent.
18784 *Ralf S. Engelschall*
18786 * Fix Win32 symbol export lists for BIO functions: Added
18787 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18788 to ms/libeay{16,32}.def.
18790 *Ralf S. Engelschall*
18792 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18793 fine under Unix and passes some trivial tests I've now added. But the
18794 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18795 added to make sure no one expects that this stuff really works in the
18796 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18797 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18800 *Ralf S. Engelschall*
18802 * Fix the generation of two part addresses in perl.
18804 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18806 * Add config entry for Linux on MIPS.
18808 *John Tobey <jtobey@channel1.com>*
18810 * Make links whenever Configure is run, unless we are on Windoze.
18814 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18815 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18820 * Add a useful kludge to allow package maintainers to specify compiler and
18821 other platforms details on the command line without having to patch the
18822 Configure script every time: One now can use
18823 `perl Configure <id>:<details>`,
18824 i.e. platform ids are allowed to have details appended
18825 to them (separated by colons). This is treated as there would be a static
18826 pre-configured entry in Configure's %table under key `<id>` with value
18827 `<details>` and `perl Configure <id>` is called. So, when you want to
18828 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18829 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18830 now, which overrides the FreeBSD-elf entry on-the-fly.
18832 *Ralf S. Engelschall*
18834 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18838 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18839 on the `perl Configure ...` command line. This way one can compile
18840 OpenSSL libraries with Position Independent Code (PIC) which is needed
18841 for linking it into DSOs.
18843 *Ralf S. Engelschall*
18845 * Remarkably, export ciphers were totally broken and no-one had noticed!
18850 * Cleaned up the LICENSE document: The official contact for any license
18851 questions now is the OpenSSL core team under openssl-core@openssl.org.
18852 And add a paragraph about the dual-license situation to make sure people
18853 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18854 to the OpenSSL toolkit.
18856 *Ralf S. Engelschall*
18858 * General source tree makefile cleanups: Made `making xxx in yyy...`
18859 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18860 Additionally cleaned up the `make links` target: Remove unnecessary
18861 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18862 to speed processing and no longer clutter the display with confusing
18863 stuff. Instead only the actually done links are displayed.
18865 *Ralf S. Engelschall*
18867 * Permit null encryption ciphersuites, used for authentication only. It used
18868 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18869 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18874 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18875 signed attributes when verifying signatures (this would break them),
18876 the detached data encoding was wrong and public keys obtained using
18877 X509_get_pubkey() weren't freed.
18881 * Add text documentation for the BUFFER functions. Also added a work around
18882 to a Win95 console bug. This was triggered by the password read stuff: the
18883 last character typed gets carried over to the next fread(). If you were
18884 generating a new cert request using 'req' for example then the last
18885 character of the passphrase would be CR which would then enter the first
18890 * Added the new 'Includes OpenSSL Cryptography Software' button as
18891 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18892 button and can be used by applications based on OpenSSL to show the
18893 relationship to the OpenSSL project.
18895 *Ralf S. Engelschall*
18897 * Remove confusing variables in function signatures in files
18898 ssl/ssl_lib.c and ssl/ssl.h.
18900 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18902 * Don't install bss_file.c under PREFIX/include/
18904 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18906 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18907 functions that return function pointers and has support for NT specific
18908 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18909 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18910 unsigned to signed types: this was killing the Win32 compile.
18914 * Add new certificate file to stack functions,
18915 SSL_add_dir_cert_subjects_to_stack() and
18916 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18917 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18918 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18919 This means that Apache-SSL and similar packages don't have to mess around
18920 to add as many CAs as they want to the preferred list.
18924 * Experiment with doxygen documentation. Currently only partially applied to
18926 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18927 openssl.doxy as the configuration file.
18931 * Get rid of remaining C++-style comments which strict C compilers hate.
18933 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18935 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18936 compiled in by default: it has problems with large keys.
18940 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18941 DH private keys and/or callback functions which directly correspond to
18942 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18943 is needed for applications which have to configure certificates on a
18944 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18946 For the RSA certificate situation is makes no difference, but
18947 for the DSA certificate situation this fixes the "no shared cipher"
18948 problem where the OpenSSL cipher selection procedure failed because the
18949 temporary keys were not overtaken from the context and the API provided
18950 no way to reconfigure them.
18951 The new functions now let applications reconfigure the stuff and they
18952 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18953 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18954 non-public-API function ssl_cert_instantiate() is used as a helper
18955 function and also to reduce code redundancy inside ssl_rsa.c.
18957 *Ralf S. Engelschall*
18959 * Move s_server -dcert and -dkey options out of the undocumented feature
18960 area because they are useful for the DSA situation and should be
18961 recognized by the users.
18963 *Ralf S. Engelschall*
18965 * Fix the cipher decision scheme for export ciphers: the export bits are
18966 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18967 SSL_EXP_MASK. So, the original variable has to be used instead of the
18968 already masked variable.
18970 *Richard Levitte <levitte@stacken.kth.se>*
18972 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18974 *Richard Levitte <levitte@stacken.kth.se>*
18976 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18977 from `int` to `unsigned int` because it is a length and initialized by
18978 EVP_DigestFinal() which expects an `unsigned int *`.
18980 *Richard Levitte <levitte@stacken.kth.se>*
18982 * Don't hard-code path to Perl interpreter on shebang line of Configure
18983 script. Instead use the usual Shell->Perl transition trick.
18985 *Ralf S. Engelschall*
18987 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18988 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18989 -noout -modulus` as it's already the case for `openssl rsa -noout
18990 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18991 currently the public key is printed (a decision which was already done by
18992 `openssl dsa -modulus` in the past) which serves a similar purpose.
18993 Additionally the NO_RSA no longer completely removes the whole -modulus
18994 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18997 *Ralf S. Engelschall*
18999 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19000 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19002 *Arne Ansper <arne@ats.cyber.ee>*
19004 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19005 to be added. Now both 'req' and 'ca' can use new objects defined in the
19010 * Add cool BIO that does syslog (or event log on NT).
19012 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19014 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19015 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19016 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19017 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19021 * Add preliminary config info for new extension code.
19025 * Make RSA_NO_PADDING really use no padding.
19027 *Ulf Moeller <ulf@fitug.de>*
19029 * Generate errors when private/public key check is done.
19033 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19034 for some CRL extensions and new objects added.
19038 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19039 key usage extension and fuller support for authority key id.
19043 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19044 padding method for RSA, which is recommended for new applications in PKCS
19045 #1 v2.0 (RFC 2437, October 1998).
19046 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19047 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19048 against Bleichbacher's attack on RSA.
19049 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19052 * Updates to the new SSL compression code
19054 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19056 * Fix so that the version number in the master secret, when passed
19057 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19058 (because the server will not accept higher), that the version number
19059 is 0x03,0x01, not 0x03,0x00
19061 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19063 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19064 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19065 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19069 * Support for RAW extensions where an arbitrary extension can be
19070 created by including its DER encoding. See `apps/openssl.cnf` for
19075 * Make sure latest Perl versions don't interpret some generated C array
19076 code as Perl array code in the crypto/err/err_genc.pl script.
19078 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19080 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19081 not many people have the assembler. Various Win32 compilation fixes and
19082 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19083 build instructions.
19087 * Modify configure script 'Configure' to automatically create crypto/date.h
19088 file under Win32 and also build pem.h from pem.org. New script
19089 util/mkfiles.pl to create the MINFO file on environments that can't do a
19090 'make files': perl util/mkfiles.pl >MINFO should work.
19094 * Major rework of DES function declarations, in the pursuit of correctness
19095 and purity. As a result, many evil casts evaporated, and some weirdness,
19096 too. You may find this causes warnings in your code. Zapping your evil
19097 casts will probably fix them. Mostly.
19101 * Fix for a typo in asn1.h. Bug fix to object creation script
19102 obj_dat.pl. It considered a zero in an object definition to mean
19103 "end of object": none of the objects in objects.h have any zeros
19104 so it wasn't spotted.
19106 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19108 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19109 Masking (CBCM). In the absence of test vectors, the best I have been able
19110 to do is check that the decrypt undoes the encrypt, so far. Send me test
19111 vectors if you have them.
19115 * Correct calculation of key length for export ciphers (too much space was
19116 allocated for null ciphers). This has not been tested!
19120 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19121 message is now correct (it understands "crypto" and "ssl" on its
19122 command line). There is also now an "update" option. This will update
19123 the util/ssleay.num and util/libeay.num files with any new functions.
19125 perl util/mkdef.pl crypto ssl update
19126 it will update them.
19130 * Overhauled the Perl interface:
19131 - ported BN stuff to OpenSSL's different BN library
19132 - made the perl/ source tree CVS-aware
19133 - renamed the package from SSLeay to OpenSSL (the files still contain
19134 their history because I've copied them in the repository)
19135 - removed obsolete files (the test scripts will be replaced
19136 by better Test::Harness variants in the future)
19138 *Ralf S. Engelschall*
19140 * First cut for a very conservative source tree cleanup:
19141 1. merge various obsolete readme texts into doc/ssleay.txt
19142 where we collect the old documents and readme texts.
19143 2. remove the first part of files where I'm already sure that we no
19144 longer need them because of three reasons: either they are just temporary
19145 files which were left by Eric or they are preserved original files where
19146 I've verified that the diff is also available in the CVS via "cvs diff
19147 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19148 the crypto/md/ stuff).
19150 *Ralf S. Engelschall*
19152 * More extension code. Incomplete support for subject and issuer alt
19153 name, issuer and authority key id. Change the i2v function parameters
19154 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19155 what that's for :-) Fix to ASN1 macro which messed up
19156 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19160 * Preliminary support for ENUMERATED type. This is largely copied from the
19165 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19167 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19169 * Make sure `make rehash` target really finds the `openssl` program.
19171 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19173 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19174 like to hear about it if this slows down other processors.
19178 * Add CygWin32 platform information to Configure script.
19180 *Alan Batie <batie@aahz.jf.intel.com>*
19182 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19184 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19186 * New program nseq to manipulate netscape certificate sequences
19190 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19195 * Fixes to BN code. Previously the default was to define BN_RECURSION
19196 but the BN code had some problems that would cause failures when
19197 doing certificate verification and some other functions.
19199 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19201 * Add ASN1 and PEM code to support netscape certificate sequences.
19205 * Add ASN1 and PEM code to support netscape certificate sequences.
19209 * Add several PKIX and private extended key usage OIDs.
19213 * Modify the 'ca' program to handle the new extension code. Modify
19214 openssl.cnf for new extension format, add comments.
19218 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19219 and add a sample to openssl.cnf so req -x509 now adds appropriate
19224 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19225 error code, add initial support to X509_print() and x509 application.
19229 * Takes a deep breath and start adding X509 V3 extension support code. Add
19230 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19231 stuff is currently isolated and isn't even compiled yet.
19235 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19236 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19237 Removed the versions check from X509 routines when loading extensions:
19238 this allows certain broken certificates that don't set the version
19239 properly to be processed.
19243 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19244 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19245 can still be regenerated with "make depend".
19249 * Spelling mistake in C version of CAST-128.
19251 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19253 * Changes to the error generation code. The perl script err-code.pl
19254 now reads in the old error codes and retains the old numbers, only
19255 adding new ones if necessary. It also only changes the .err files if new
19256 codes are added. The makefiles have been modified to only insert errors
19257 when needed (to avoid needlessly modifying header files). This is done
19258 by only inserting errors if the .err file is newer than the auto generated
19259 C file. To rebuild all the error codes from scratch (the old behaviour)
19260 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19261 or delete all the .err files.
19265 * CAST-128 was incorrectly implemented for short keys. The C version has
19266 been fixed, but is untested. The assembler versions are also fixed, but
19267 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19268 to regenerate it if needed.
19269 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19270 Hagino <itojun@kame.net>*
19272 * File was opened incorrectly in randfile.c.
19274 *Ulf Möller <ulf@fitug.de>*
19276 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19277 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19278 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19279 al: it's just almost always a UTCTime. Note this patch adds new error
19280 codes so do a "make errors" if there are problems.
19284 * Correct Linux 1 recognition in config.
19286 *Ulf Möller <ulf@fitug.de>*
19288 * Remove pointless MD5 hash when using DSA keys in ca.
19290 *Anonymous <nobody@replay.com>*
19292 * Generate an error if given an empty string as a cert directory. Also
19293 generate an error if handed NULL (previously returned 0 to indicate an
19294 error, but didn't set one).
19296 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19298 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19302 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19303 parameters. This was causing a warning which killed off the Win32 compile.
19307 * Remove C++ style comments from crypto/bn/bn_local.h.
19309 *Neil Costigan <neil.costigan@celocom.com>*
19311 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19312 based on a text string, looking up short and long names and finally
19313 "dot" format. The "dot" format stuff didn't work. Added new function
19314 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19315 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19316 OID is not part of the table.
19320 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19321 X509_LOOKUP_by_alias().
19325 * Sort openssl functions by name.
19329 * Get the `gendsa` command working and add it to the `list` command. Remove
19330 encryption from sample DSA keys (in case anyone is interested the password
19335 * Make *all* `*_free` functions accept a NULL pointer.
19337 *Frans Heymans <fheymans@isaserver.be>*
19339 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19342 *Anonymous <nobody@replay.com>*
19344 * s_server should send the CAfile as acceptable CAs, not its own cert.
19346 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19348 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19350 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19352 * Temp key "for export" tests were wrong in s3_srvr.c.
19354 *Anonymous <nobody@replay.com>*
19356 * Add prototype for temp key callback functions
19357 SSL_CTX_set_tmp_{rsa,dh}_callback().
19361 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19362 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19366 * X509_name_add_entry() freed the wrong thing after an error.
19368 *Arne Ansper <arne@ats.cyber.ee>*
19370 * rsa_eay.c would attempt to free a NULL context.
19372 *Arne Ansper <arne@ats.cyber.ee>*
19374 * BIO_s_socket() had a broken should_retry() on Windoze.
19376 *Arne Ansper <arne@ats.cyber.ee>*
19378 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19380 *Arne Ansper <arne@ats.cyber.ee>*
19382 * Make sure the already existing X509_STORE->depth variable is initialized
19383 in X509_STORE_new(), but document the fact that this variable is still
19384 unused in the certificate verification process.
19386 *Ralf S. Engelschall*
19388 * Fix the various library and `apps/` files to free up pkeys obtained from
19389 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19393 * Fix reference counting in X509_PUBKEY_get(). This makes
19394 demos/maurice/example2.c work, amongst others, probably.
19396 *Steve Henson and Ben Laurie*
19398 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19399 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19400 are no longer created. This way we have a single and consistent command
19401 line interface `openssl <command>`, similar to `cvs <command>`.
19403 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19405 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19406 BIT STRING wrapper always have zero unused bits.
19410 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19414 * Make the top-level INSTALL documentation easier to understand.
19418 * Makefiles updated to exit if an error occurs in a sub-directory
19419 make (including if user presses ^C) [Paul Sutton]
19421 * Make Montgomery context stuff explicit in RSA data structure.
19425 * Fix build order of pem and err to allow for generated pem.h.
19429 * Fix renumbering bug in X509_NAME_delete_entry().
19433 * Enhanced the err-ins.pl script so it makes the error library number
19434 global and can add a library name. This is needed for external ASN1 and
19435 other error libraries.
19439 * Fixed sk_insert which never worked properly.
19443 * Fix ASN1 macros so they can handle indefinite length constructed
19444 EXPLICIT tags. Some non standard certificates use these: they can now
19449 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19450 into a single doc/ssleay.txt bundle. This way the information is still
19451 preserved but no longer messes up this directory. Now it's new room for
19452 the new set of documentation files.
19454 *Ralf S. Engelschall*
19456 * SETs were incorrectly DER encoded. This was a major pain, because they
19457 shared code with SEQUENCEs, which aren't coded the same. This means that
19458 almost everything to do with SETs or SEQUENCEs has either changed name or
19459 number of arguments.
19461 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19463 * Fix test data to work with the above.
19467 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19468 was already fixed by Eric for 0.9.1 it seems.
19470 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19472 * Autodetect FreeBSD3.
19476 * Fix various bugs in Configure. This affects the following platforms:
19480 unixware-2.0-pentium
19485 * Eliminate generated files from CVS. Reorder tests to regenerate files
19486 before they are needed.
19490 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19494 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19496 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19497 changed SSLeay to OpenSSL in version strings.
19499 *Ralf S. Engelschall*
19501 * Some fixups to the top-level documents.
19505 * Fixed the nasty bug where rsaref.h was not found under compile-time
19506 because the symlink to include/ was missing.
19508 *Ralf S. Engelschall*
19510 * Incorporated the popular no-RSA/DSA-only patches
19511 which allow to compile an RSA-free SSLeay.
19513 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19515 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19516 when "ssleay" is still not found.
19518 *Ralf S. Engelschall*
19520 * Added more platforms to Configure: Cray T3E, HPUX 11,
19522 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19524 * Updated the README file.
19526 *Ralf S. Engelschall*
19528 * Added various .cvsignore files in the CVS repository subdirs
19529 to make a "cvs update" really silent.
19531 *Ralf S. Engelschall*
19533 * Recompiled the error-definition header files and added
19534 missing symbols to the Win32 linker tables.
19536 *Ralf S. Engelschall*
19538 * Cleaned up the top-level documents;
19539 o new files: CHANGES and LICENSE
19540 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19541 o merged COPYRIGHT into LICENSE
19542 o removed obsolete TODO file
19543 o renamed MICROSOFT to INSTALL.W32
19545 *Ralf S. Engelschall*
19547 * Removed dummy files from the 0.9.1b source tree:
19548 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19549 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19550 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19551 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19552 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19554 *Ralf S. Engelschall*
19556 * Added various platform portability fixes.
19560 * The Genesis of the OpenSSL rpject:
19561 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19562 Young and Tim J. Hudson created while they were working for C2Net until
19565 *The OpenSSL Project*
19567 ### Changes between 0.9.0b and 0.9.1b [not released]
19569 * Updated a few CA certificates under certs/
19573 * Changed some BIGNUM api stuff.
19577 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19578 DGUX x86, Linux Alpha, etc.
19582 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19583 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19588 * Add -strparse option to asn1pars program which parses nested
19591 *Dr Stephen Henson <shenson@bigfoot.com>*
19593 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19597 * DSA fix for "ca" program.
19601 * Added "-genkey" option to "dsaparam" program.
19605 * Added RIPE MD160 (rmd160) message digest.
19609 * Added -a (all) option to "ssleay version" command.
19613 * Added PLATFORM define which is the id given to Configure.
19617 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19621 * Extended the ASN.1 parser routines.
19625 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19629 * Added a BN_CTX to the BN library.
19633 * Fixed the weak key values in DES library
19637 * Changed API in EVP library for cipher aliases.
19641 * Added support for RC2/64bit cipher.
19645 * Converted the lhash library to the crypto/mem.c functions.
19649 * Added more recognized ASN.1 object ids.
19653 * Added more RSA padding checks for SSL/TLS.
19657 * Added BIO proxy/filter functionality.
19661 * Added extra_certs to SSL_CTX which can be used
19662 send extra CA certificates to the client in the CA cert chain sending
19663 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19667 * Now Fortezza is denied in the authentication phase because
19668 this is key exchange mechanism is not supported by SSLeay at all.
19672 * Additional PKCS1 checks.
19676 * Support the string "TLSv1" for all TLS v1 ciphers.
19680 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19681 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19685 * Fixed a few memory leaks.
19689 * Fixed various code and comment typos.
19693 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19694 bytes sent in the client random.
19696 *Edward Bishop <ebishop@spyglass.com>*
19700 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19701 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19702 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19703 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19704 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19705 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19706 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19707 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19708 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19709 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19710 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19711 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19712 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19713 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19714 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19715 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19716 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19717 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19718 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19719 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19720 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19721 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19722 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19723 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19724 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19725 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19726 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19727 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19728 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19729 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19730 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19731 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19732 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19733 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19734 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19735 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19736 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19737 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19738 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19739 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19740 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19741 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19742 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19743 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19744 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19745 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19746 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19747 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19748 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19749 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19750 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19751 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19752 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19753 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19754 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19755 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19756 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19757 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19758 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19759 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19760 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19761 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19762 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19763 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19764 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19765 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19766 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19767 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19768 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19769 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19770 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19771 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19772 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19773 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19774 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19775 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19776 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19777 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19778 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19779 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19780 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19781 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19782 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19783 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19784 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19785 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19786 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19787 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19788 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19789 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19790 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19791 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19792 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19793 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19794 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19795 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19796 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19797 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19798 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19799 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19800 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19801 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19802 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19803 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19804 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19805 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19806 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19807 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19808 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19809 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19810 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19811 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19812 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19813 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19814 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19815 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19816 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19817 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19818 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19819 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19820 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19821 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19822 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19823 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19824 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19825 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19826 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19827 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19828 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19829 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19830 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19831 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19832 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19833 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19834 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19835 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19836 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19837 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19838 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19839 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19840 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19841 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19842 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19843 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19844 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19845 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19846 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19847 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19848 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19849 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19850 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19851 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19852 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19853 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19854 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19855 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19856 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19857 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19858 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19859 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19860 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19861 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655